I am trying to understand something about the economy and AI. If AI disrupts an industry that previously had high wages and now can be automated by commodity AI products, would it not actually lower GDP in some way (initially anyway)? What I mean is, if you had an industry that was worth say $1 billion, but now that same industry could be automated and done cheaply then it is no longe worth $1 billion? In some ways this is beneficial to the economy in that has reduced the cost of this industry and so customers see a benefit of cheaper products, but how would it benefit say AI companies since they are not capturing a $1 billion industry, but instead sort of destroying a $1 billion industry? I am mainly looking at this from the perspective of the value of an AI company.. If they just keep commoditizing industries and there is ample competition between AI companies, then how is the AI company capturing value such that they would be worth as much money as people are saying they are worth (trillions of dollars)? I know that there is more to this, like freeing up this money and labor could lead to new industries etc., but how much of this will actually go to AI companies?
One view I've long held is that money is a distraction when understanding economics. What you should look at instead is the opposite flow: that of goods and services. Money is just lube to enable those*
It neatly bypasses many problems like these.
* Though don't get me wrong it's a extremely good tool for enabling them, bartering is ridiculously time consuming. As a complete aside in one multiplayer game I played, I got a drop of a rare material useful for crafting endgame gear on a character that didn't even have starting-tier gear. I tried bartering first and even though I was willing to accept a price 50% over market for some item I needed, the trade could not be completed because the seller couldn't provide "change" in commodity items.
On a third party site that used one single currency for everything, spreads were instead at 5-10%, and I was quickly able to translate my item to currency and buy three items I needed.
> had an industry that was worth say $1 billion, but now that same industry could be automated and done cheaply then it is no longe worth $1 billion?
You said GDP, not market cap. So let's say this industry produces $1B of products/services so that is its contribution to the overall GDP.
The CEOs dream there is to cut their workforce 50% and continue generating $1B of same products/services. So it is still same contribution to GDP, no change. Within that company, profit margins rise but contribution to GDP is still $1B.
The flaw here is that every other CEO is thinking the same, but if every company cuts 50% of employees we'll have massive unemployment, nobody will have money to buy any goods/services, so in the medium term their production has to go down and overall GDP goes down.
Long term, presumably, employment market rearranges itself around new reality. But for in the meantime, you can't sell a lot of product when most of the country is unemployed. Who will buy?
> But for in the meantime, you can't sell a lot of product when most of the country is unemployed. Who will buy?
That is something I never understand when politician and economy experts talk on TV about cutting costs and lowering wages. Economy works best if a lot of money flows through a lot of peoples’ hands. If you take away purchasing power, you ruin the economy.
It's actually simple and the misunderstanding comes from the fallacy that the economy is a fixed sized pie.
The idea is that wealth comes from production of goods and services. If a company cuts 50% of their employees through automation, the laid off workers can go produce something else at another company. Then we'll have the products from the previous company, now more efficient, plus whatever gets produced by the new company that absorbed the laid off workers.
The problem is when there's nothing else to do, so to speak. It can take some time to find new uses for labor, leading to unemployment. But humanity has done a really good job at eventually finding new things to do since the dawn of civilization.
This is why the world didn't end after farm labor got displaced by farming machines.
You've already identified the reason; because those actors ( healthcare providers ) were given artificial monopolies by the government. If you allowed non AMA
1) In a majority of states, if anyone wants to break into the regional hospital business and open a facility, they have to get state approval from the other hospitals in the market and prove need. Obviously other hospitals are financially incentivized to make this laborious, cumbersome, and expensive since it's the very defintion of a regulatory moat.
2) Costs for direct care _have_ to stay high, because the AMA caps the number of physicians to maintain labor power (high pay) for it's members. High pay for doctors is an input cost into the end-price that consumers have to pay.
3) The state helps enforce #2 by strictly prohibiting NP/ PA / a random mechanic from diagnosing / prescribing. This ensures that the AMA has a moat that requires their service. This scheme keeps simple / obvious diagnoses from being cheap because an AMA Doctor(TM) has to be the one who actually says the magic words.
Now you can make the argument that all of these are necessary and good for safety, training, etc. But unfortunately we can't have it both ways. Heavy regulation in the name of safety manifests as a regulatory monopoly with all the downsides too.
This is a separate problem. Readjustment isn't immediate and won't necessarily restore status quo. Displaced farm workers couldn't stay in their agrarian communities and had to move to cities.
I never said that readjustment isn't painful. I was simply explaining why it is possible for the society as a whole to prosper with increased efficiency.
It’s possible for society as a whole to prosper if the increased productivity from that efficiency gain is spread across society. If the productivity increased is captured by an increasingly small portion of the population while the majority are pushed down in wealth or left destitute it does not benefit society as a whole.
Capitalism only works with regulations and not this abomination we’ve currently got.
Regulatory capture is a key mechanism for increasing concentration of power and resources.
I don't disagree with the spirit of what you're saying. I'd just be careful when looking for more government, an entity controlled by the powerful in the entire world, as the answer.
While government must indeed be part of the solution, the power elite do a pretty good job at dressing up their interests as beneficial to the population.
We already have that. I’d rather have some anti trust busting going on like under Lina Khan but that appeared to be a dying gasp of actual good governance before we went back to our regular scheduled programming of the corporate elites raiding the country with shit like DOGE or Jensen getting the current admin to look the other way on his dealings with China
> Economy works best if a lot of money flows through a lot of peoples’ hands. If you take away purchasing power, you ruin the economy.
Indeed. But the rich people on top don't care. They ride the train until it's dead, and the current economy seems to have a decade or two until it finally collapses in it.
Note they are already preparing, building bunkers on remote islands, talking about how to keep their security staff in line with shock collars and whatnot. They know the train will run out of speed sooner rather than later.
that's only for the ultra rich though? what will become of the merely super rich who can't afford self-sustaining bunkers? are then not worried like the rest of us?
A lot of fairly rich people think they're middle class. People compare themselves to a peer group.
E.g., I bet many readers here have a net worth of > 2 MM USD and yet think of themselves as struggling middle class because they don't compare themselves to the median but to the colleague one level up who has 5 MM.
So, people with surprisingly high net worth often don't think they're going to become a target because they don't see themselves as rich. And on the other hand, they will still be able to afford a bigger share of the pie compared to the masses. There will be lots of people starving and running out of clean water and places to live worldwide long before it affects them. You got a preview of that on a small scale with oil and Iran.
The missing ingredient is globalization. Companies have the option of moving the jobs to places where wages are lower, e.g. China or India.
The fix is to globally coordinate to not have this race to the bottom. That is going well currently /s. The WTO is being sabotaged, US quit the global minimum corporate tax rate agreement, US is doing "America first" while shooting itself in the foot by tariffs that don't even change long term behavior, because they are illegal and get refunded.
That’s cheaper, not falling. And their public pronouncements usually tout cost and ease of business, not labor per se. (Texan after-tax wages aren’t materially lower than CA or NY.)
> but if every company cuts 50% of employees we'll have massive unemployment, nobody will have money to buy any goods/services, so in the medium term their production has to go down and overall GDP goes down
This is such a prominent and bleak view, but there are a lot of possible outcomes. People are flexible and the market is dynamic. In no particular order:
1. Other companies will hire a portion of those people at lower wages.
2. Some people will remain unemployed, and may live off government programs.
3. The government, who has a vested interest in keeping consumption going, may invest in new social welfare programs, such as universal basic income.
4. People may retrain and enter into new industries, different jobs.
5. New jobs that did not exist before may spring up. Some jobs may be awesome and some may be awful.
6. The future is unequally distributed and people may migrate to where the work remains.
7. Some may protest and riot.
8. Some may choose more personally drastic means when faced with such existential horror.
9. Some may start new businesses and hustle because the barriers to entry have been lowered, and expand the market.
10. The government may incentivize options 4, 5, 6, 9.
All of these things have already happened before. We have already-happened histories of whole industries being wiped out or decimated. The transition sucks but life goes on, the market grows. There is too much vested interest in the market going down for too long without significant government and business intervention.
Huge fan of big lists; but you're missing an obvious option in this situation - prices drop sharply and people can maintain a pretty good lifestyle with much less financial outlay.
What we've seen in IT is that the price of tech goods drops really quickly as the underlying fundamentals improve. Things like RAM being measured in gigabytes instead of kilobytes. We'd probably see the same thing in knowledge work where you can get a human-year of engineering work for a few thousand dollars or less.
That has flow on effects were all sorts of services becomes absurdly cheap and all the physical goods that have significant engineering risks or coordination overheads to produce will be much more affordable.
Mass deflation works on everything but debt. Anyone with debt, from people to businesses to governments, would basically all face insolvency as their ability to pay it is decimated
The government would (correctly) not allow deflation. They’d print and hand out money like drunken sailors, effectively socializing the productivity gains that led to the lower prices.
There’s an argument this happened during the “China Shock”: Because (among other factors like information technology) cheap Chinese goods held inflation down, the Federal government was able to cut taxes, increase spending for a decade+, all while the Fed maintained low interest rates.
Wouldn't this just create a circle where people attempt keep prices higher than whatever the government is subsidizing? And how much would this help joe blow? Yes debt may be renegotiated for business getting sweet rates, but locked in home owners for example would essentially be screwed. Or would banks get an arm twisting to relent?
A fundamental fact of the economy is that people's wages are paid out of the prices for the goods and services they help produce. We don't often think of it that way on an individual level, partly because we don't trust business owners to pass the money through and partly because it's good for me to have high wages but pay low costs. But when it comes to systemic things like inflation rates, there's no way to avoid it.
> 1. Other companies will hire a portion of those people are lower wages.
If you belive AGI, would not ai also affect all those other companies? Who would also be cutting jobs. And in the whole economy, how many people do you even need to do jobs, if AGI is commonplace and everywhere?
And if you add robots on top of that, even manual jobs like plumbing might be affected.
Neither “the future” nor “consensus” are evenly distributed. If a specific job is truly going away, the migration of unemployed people will first drive supply up (more folks looking for work) and drive cost of wages down (downwards pricing pressure). At that point, people may rethink their cost/benefit calculations for hiring vs using AI, outside of whatever other ethical, moral, or philosophical stances they may hold towards AI. The job itself may change at that point as well, with the required skill bar being pushed up (you truly need to be an expert) or pushed down (the job is rote, simplified, custodial).
Lastly, I think people are generally prosocial. They like working with other people. And there have been many revivals of otherwise obsolete technologies and ways of creating as heritage work. Ie, clothes manufacturing, media, agriculture.
I ultimately don’t expect a desire to “do things yourself” or “engage in existentialist and prosocial things” to go away as a basic human drive.
I am sorry but how are the 10 points you listed not "bleak"? Yes the plague also happened before, doesn't mean a new plague is okay and will go exactly the same way.
All of those scenarios are plausible in any market that is changing due to technology or pricing or geopolitical pressures. The broader point is that it usually doesn’t end with mass unemployment and food riots, not that change doesn’t happen.
But personally , I don’t think 4, 5, 9, 10 are bleak, unless you view any externally forced change to a personal status quo bleak.
And I’m not thinking of plague, but of the impact of industrialization (ie luddites in England), and later deindustrialization (ie rust belt in US), or the rise of telecoms/internet on how businesses were run.
But also there’s another problem with the CEO’s deal of cutting 50% of the workforce.
In a competitive market they must lower their prices or someone else will and will eat their lunch.
Now all AI did was cause 50% of people to lose their jobs, outsourced to AI. Instead of reaping rewards the AI customer companies saw revenue shrink. The only person benefiting is the AI company.
Second aspect is that if there is enough companies they start competing with each other more. So they will actually cut the prices. Competition will happen in enough cases for potentially prices going down. Which could actually lead to market getting smaller. Especially if there is not more demand.
Its frankly absurd I had to scroll down this far to see somebody mention the staple of what capitalism is actually about, and how it overtook all other systems; competition.
With competition, we can espouse the virtue of it being greed organized to serve the greater good as it pursues to most cost-effectively improve our lives.
> The CEOs dream there is to cut their workforce 50% and continue generating $1B of same products/services. So it is still same contribution to GDP, no change. Within that company, profit margins rise but contribution to GDP is still $1B.
But the workforce no longer have job and so purchasing power to actually buy the goods
How is it more profitable to run a business if (a) huge chunks of the population have lost their income and (b) whatever you think your business would do can just be done by Claude for you? How is it "much more profitable" to try to be a middleman in that scenario?
What would happen if AI was actually used to shorten the hours an average white collar worker puts in without getting them laid off or paid less?
People working shorter hours but making the same cash would probably want to buy stuff and do fun things with that extra time. This would drive up demand for goods and services. They might then be called back into work to meet the new demand. A new equilibrium would form where workers work shorter hours than now, but their productivity would also be higher. GDP might go up and billionaire profits with it.
I am not an economist. What mistake have I made here? Could billionaires really make more money if they were less greedy?
You are thinking of the welfare of the workers. That’s not the way American capitalist work. Workers are expenses not assets. The more they can get rid or lower that cost the better for their company.
The related economic concept here is the Broken Window Fallacy. Doing 'redundant' work doesn't improve the economy, because there's an opportunity cost doing to anything. And replacing the need to "fix broken windows" or do now-AI-automated work would free up that opportunity cost. Which is how it's worked for nearly all automation. (Though Computers are subject to long running debate about this. See the Solow Productivity Paradox)
The gigawatt-sized asterisk on all of this is that AI companies have to capture that $1 billion. Anthropic has to ten-fold increase it's revenues or it goes bankrupt under it's current commitments. It has to increase it's revenues another 10 times to reach the $10,000,000,000,000/yr revenue required to support it's current valuation.
Anthropic put their Total Addressable Market at 35% of Global GDP (~$40 trillion/yr). They're planning to capture all of the money.
The second gigawatt-sized asterisk is that because of the fine handiwork of Mr. Huang, China has made very competitive AI models available for free, so the actual ability for AI companies to capture the value of the labour they displace ends up being quite limited. Thus a problem in the predictions: Something as to break, but it is as of yet unclear what that will be. Maybe Chinese AI will be banned, maybe Anthropic & OpenAI go bankrupt.
The big issue with all this is that people whose jobs are destroyed by automation very rarely end up seeing improvements to their standard of living. And with the modern system, people are pretty boned if they chose the wrong career.
The luddite movement is informative here. When manufactures found ways to mass produce goods eliminating jobs left and right, the people in the skilled trades whose jobs were eliminated did not benefit in the slightest from the economic progress.
We now live in an era of diminishing jobs which provide livable wages. It's why so many economies have moved to require everyone to participate in the labor market just to achieve the basics.
Put more simply, broken windows harm society as a whole, but they do make glazers and the people who support the glazers happy. Now that we live in an era where a significant portion of the population is glazers so-to-speak and ai threatens to eliminate those jobs, it looks like the only path forward for those glazers is low wage jobs.
I watched Thatcher destroy the UK mining industry. Whole communities wrecked, nothing to do there, no hope. I'm sure the US rust belt had the same experience under Reagan.
This happens routinely. It's just the first time that educated white-collar workers like us have had to experience it.
In the .com boom many people with degrees from other disciplines retrained for tech jobs. So the people who are (in theory) getting displaced here, like software engineers or mathematicians are also some of the most capable of filling other roles. So that's one thing. This is different from the industrial revolution where the people who got displaced were those doing the cheapest labour.
An economy is a complex system with feedback loops and equilibrium. Companies need consumers to make money. AI companies make money by selling AI to companies that make money by selling things to people or other companies. There is no economy in which nobody has a job and nobody buys anything and big tech makes money. So that is obviously not an equilibrium point. How exactly we get to an equilibrium and what that looks like is hard to predict and also depends on policy. Governments can employ more people. They can provide various incentives. And countries as a whole are somewhat self sustaining economies (some more and some less).
So... The world is not going to end (at least not from jobs going away). How even jobs and wealth are distributed depends to a large degree on policy but it has to be stable in many ways. Look at Russia? What sort of economy have they got compared to most western countries? A pretty sucky one. But people still largely have jobs and the country continues to function +/-.
>So the people who are (in theory) getting displaced here, like software engineers or mathematicians are also some of the most capable of filling other roles.
Anything that can displace those types of office jobs can much more easily displace the entirety of the rest of the paper-pushing middle-management office jobs...
So what are software engineers and mathematicians "the most capable" of being retrained for that isn't eaten by "AI agents" in this scenario?
I got coffee with a nice person from HN today. They were wearing a hat with the word "Luddite" on it. I knew what it meant but asked anyway - yes it's not anti-tech, it's anti automation and loss of craft.
The thing about luddites is that yes they wanted to preserve human craftsmanship, and as you say it didn't work. John Henry died. Luddites worked in factories and declined into poverty. Hopeless. Doomed.
But, I am cranking down my personal backlog with AI tools, using them at work, and loving the usefulness of them. It doesn't feel like I'm a master weaver who has their work automated, it feels like I'm reaping productivity gains.
Was there ever a moment in the Luddite movement when some of them leaned in and had fun? Did some of John Henry's colleagues just really enjoy the new steam drivers?
It doesn't _feel_ like I'm being forced out of my trade. I sincerely doubt the Luddites had a moment where they happily shelled out the 1820 equivalent of $20/m to get access to a weaving machine. I'm not saying "this time is different", but I am saying everyone tells me to not use AI, and to be afraid, and I'm de-skilling or whatever, etc. How do we all _know_ what is right?
A loom does not bring agency. It takes it. But I feel like having a computer, access to LLM, increases my agency just a bit. It allows me to do more, not just do what I need to _faster_. It's not _just_ automation.
The fundamental difference is that Looms were point process automation that cost capital. They took a job a person did and made it automatic so that the person could be fired in exchange for up front cost.
The weaver couldn't build the machine, maintain it, adapt it, take it home, use it for other processes, try their own creative designs, open a competing shop, work in other textiles etc.
But agentic tools are not that. They DO allow all that. In fact software engineering has always been about automating things, and agentic tools just allow more people to automate more things, AND create more software the way they want it.
LLMs just don't feel like a loom. They feel something that can be used by workers themselves. Like power tools. A very generic power tool. That's essentially free. Unlike all automation to date which required massive up front capital to automate one single process.
Maybe that's the difference - accessibility and generalization.
I think the Luddites' beef was more about losing options, as they watched their sense of agency in their lives and careers slip away. They couldn't rent their own automated looms for the equivalent of $20/month. Those things were expensive, exotic machines usable only in industrial settings, more like NVL72 racks than RTX 5090s. But with AI, we are as you suggest gaining options, so any comparison with Luddism fails from the get-go.
Obviously the correct way to view the arrival of AI is as an empowering tool for the ultimate benefit of us all, but just as obviously there are a lot of people who simply don't seem capable of understanding that. Obviously the correct way to react to the arrival of AI is to severely regulate the monopolists and oligarchs rather than the technology itself, but... see above.
Unfortunately there is no obvious way to deal with people who think the Luddites were somehow morally superior, or who otherwise want to rehabilitate or emulate them. It's like flashing back to late 2019, watching news reports come in about some sort of weird, unusually-severe flu that seemed to be spreading. Nobody knew for sure what to do then, and I don't think anyone does now. But we are going to have to do something to keep those people from dragging us backward with the self-centered psychological pandemic they are trying to spread.
People (even software engineers, seemingly) seem to conflate writing code, which is a skill that LLMs do sort of replace, with software engineering.
In my opinion LLMs can write code but they don't replace human skill and judgement in software engineering.
They can help you make decisions but ultimately humans still have to be responsible.
When your payments gateway goes down the week before Christmas you can't just say "well Claude did that, I have no idea how it works or why it can't handle the traffic".
The LLM bros want to claim that role will be handled by... who? The CEO is going to start prompting Claude at 3am making random changes in production to fix the payments gateway? The product manager? The cleaner?
Or you have some agent loop constantly making random changes in production to crucial core parts of your codebase with no human supervision?
I just can't see how actual software engineers who understand how the system works aren't still needed
Of course someone will be in the line of responsibility (blame) for software reliability. That person will also use automation to make their job / life easier. And they'll insist on changes to the stack to make it more reliable and performant. And there will still be proven offerings by large companies for profitable niches.
That person/company is a software engineer/company in practice if not title. Anyone who tells you this role won't exist is probably wrong. At least practically wrong for 5-10 years? Dunno.
But the Luddite question is not existence but scale. Can all 4 million software engineers find jobs if this is all that's left of software engineering. Also dunno
You hire a "3am meat proxy" who types "Claude why is in payments gateway down?"
Actually no, you ask Claude to make proper alerting and telemetry then ask it to make a thing that checks alerts every 10 minutes and takes whatever action is needed to fix the problem.
- knowing that you need monitoring and alerting and how monitoring and alerting best practices are generally done so you can prompt Claude, because by default it won't do it if you just ask it to bash out features
- trusting Claude to change the code handling payments in production with no human review
First one is the easier issue, maybe there is some way a non-technical person could get there by blindly following some generic guidance and feeding it into tickets.
Second one I would not be happy with as a customer or shareholder of a business
It wouldn't, it would just mire the US in an unproductive and isolationist quagmire - reducing our global competitiveness because every US company would be paying an "anthropic tax" for a service that competing countries access for much lower costs.
I don't want to make this yet another geopolitics subthread, but I don't think anyone sees Russia as a credible threat to EU. They are losing (for certain values of losing, anyway) against Ukraine as we speak.
I mean, both can be right at the same time, it may certainly be a threat to democratic systems and their institutions in the current hybrid warfare, and simultaneously a very weak and corrupt military.
The absurdity is the point; The clear-cut ways to resolve the tension of "AI giants need pricing power but don't have it" are absurd (admittedly, the former much more so than the latter)
It's hard to predict an outcome "somewhere in between of two absurdities"
If the US actually wanted a world-wide ban, it would probably happen.
Obviously we can't effectively ban already released models, but a ban on models with >10T parameters for instance, is quite feasible. It takes a ton of power to train and operate those, and infrared satellites can spot that.
I'm confident China would be very happy to sign on to a world wide ban.
1. AI is one of the only high tech areas where America is ahead of China. If AI is knee-capped, China would indisputably become the world leader in industrial tech and high tech.
2. China is releasing models as open-weight to commoditize the complements. A ban on 10T models would let them specialize and utilize the lower parameter models to extend their lead in areas like robotics.
And once the US & China signed on I'm sure Europe and the middle powers would quickly agree.
The world can barely (and failing) to stop the creation of more nukes. Something that people can very clearly understand, see, and have decades of media to comprehend the consequences.
AI is too much of a field leveler in numerous fields for a ban on high end models to stick.
Sometimes, at the oddest of hours, I am awakened by migraines, and find myself seeking diversion from the pulsations of pain. It so happens, that I found this thread, and had a bit of an inspired thought.
The US is (and has been, for decades) overconfident in its ability and stature, in much the same way the Southern States were, during the American Civil War. You see, the time-span from the end of the Revolutionary War to the beginning of the Civil War, saw a transformation of the ethics of the Southern elites. The revolutionary elites were (and you can see this in their writings and correspondences) embarrassed and ashamed of the over-reliance that their economy had on slave labor. They were embarrassed, because this arrangement did not live up to the egalitarian and liberal (as in freedom) ethos that their new country aspired to.
Within a few generations, the grandchildren and great-grandchildren of these elites, had lost that shame and embarrassment. They felt, rather, insulted by the disgust that their Northern neighbors expressed. And so, to compensate, they became _proud_ of their slave-owning ways. They became something that their grandparents would have been ashamed of. Eventually, the Civil War broke out, and, at the cost of 600 thousand lives, the abolitionists got what they wanted (or at least the beginning of it -- emancipation would take a century to truly complete).
Note that this victory was only possible because of Northern industry and immigration.
We are in a similar situation today. After the Second World War, the American elites (and the American people) were embarrassed and ashamed of the failure of their democratic and capitalist/free-market society, to live up to the aspirational ethos of egalitarianism and liberty and self-sufficiency. Nevertheless, there was a Cold War to be won, and that required the unsavory combination of labor-exploitation (read anything by Mike Davis), neo-colonialism, and foreign (and domestic) espionage and subterfuge. From the beginning of the Cold War, to its end, the ethics of the American elites (and, in a different way, the American people) had, yet again, transformed into something the elites of the post-World-War would have been ashamed of. You can see hints of this transformation in the writings of John Updike: A society that was becoming increasingly fixated on hedonism, luxury, profit, status-symbols, etc.
From Reagan/Thatcher until now, the American elite has felt insulted by the suggestion that they are, perhaps, too exploitative and too parasitic and too ignorant. They now wear their largesse and cold-heartedness and ignorance as a badge of pride, as an inalienable right, that they will let others fight and bleed and die for.
In other words, where the post-Cold-War elites would have been horrified by the writings of William Gibson, the contemporary elites treat them as a user-manual.
From 2020 to now, it feels like we are seeing a slow-motion version of the USSR collapse (not in the sense of redrawing borders, but in the sense societal and institutional decline). The whole AI frenzy _may_ be a desperate attempt soften the blow of future mass emigration (of the so-called, high-value human capital).
Vastly overestimating the political capital the US has left to spend after the past decade+. Your closest neighbour is choosing to cosy up with the Reds over dealing with anymore shenanigans.
The US shouldn't need to spend political capital to achieve something that the other party actually wants.
OTOH, the US also has a massive credibility problem. Nobody's going to sign anything without solid mechanisms to ensure the US won't ignore or tear up the agreement in the near future.
Well for starters, the US will need to rip up the CLOUD Act, get on its knees to win back Western goodwill and restart resource extraction and onshore manufacturing. Oh, and restart research spending across all domains.
On the other hand, China could just replicate GPU technology or adapt its models to its own chips (already happened), and then push that onto the rest of the world for cheap. Of late, they too have been very aggressively building out their data centers in Europe and the Middle East. China also controls the resource chain for most essential minerals (thanks to decades of US offshoring).
A recent episode of Odd Lots had a segment about this question. Their guest, a senior analyst who has been tracking the GPU hardware space (+ politics around it) for years had quite revealing observations.
Huawei's hardware is apparently good enough for inference and the CCP are actively promoting their domestic AI gear ... among other things by banning high-end chip imports. The export bans from the US and import bans from China both have same technical targets but different political goals. The analyst commented that Huawei's gear is not yet good enough for model training.
The situation is going to change once China can - to quote the analyst - "complete a training run on Huawei chips". When that happens, China will start to really crack down on AI chip imports to ensure that their domestic technology is forced to catch up.
The US primarily accomplishes these types of goals with its economic might, not through the expenditure of political capital. And the US still has about 33 trillion of that might in its inventory.
Just like the US auto industry has worked out. Aside from Teslas, you have shitty cars that almost nobody outside of the US want. Now in Europe we have our own share of shitty cars because of Bruxelles pushing emissions and road safety legislation and the auto industry, mostly led by old farts, is struggling electrifying their peoduct range and copying Tesla and Chinesw models' features.
This is such an interesting part of the whole AI equation.
The main benefit of the closed models is the regulatory control to say, insert guard rails against hacking things. That's also the main drawback as you now have handed control over this technology to a handful of companies, who will reap all the rewards of this captured sector.
It would be like the google/apple duopoly on steroids. Everyone forced to buy in to one of two ecosystems, with limited to no ability to undermine the duopoly via control of the device hardware. One can't make a real competitor because it requires access guarded by existing relationships and money. For the AI houses, it means you need approval from these groups to create in a competitive manner. That, IMO is too much power for an unelected organization to carry.
AI will likely only provide real benefits to humanity if it's open. But if it's open, it will also enable extreme harm in the form of cyber attacks, scams, and destruction of IP.
By keeping AI closed, it will bring a new form of imperialism. Instead of the core maintaining IP and the periphery suffering the secondary effects of industry, the core will maintain the capability and the periphery will be locked out of competition due to silly nation-state rivalries. This won't benefit humanity, it will lock us into the same cycle of misery for marginal gains measured in stock values.
So there's two ways AI can reasonably go: Cyber punk corporate overlords or cyber anarchy due to the inability to put a cork on the genie's bottle.
Both seem terrible to me. What good is space travel without freedom? What good is automation if there are incentives to maintain profit over stability? What good is the technology if any usage of it exposes you to fraud, violence, and blackmail at any moment?
Big fans of LLMs never ask "to what end?", they repeat shallow near term "wows" without considering the second and third order effects. "wow think about how good LLMs can spot cancer in scans!", great! but who will afford the scan without work? Will the overlords commit to luxury gay space communism? Or will they simply say "tough luck" and let that cancer fester?
Real AGI is incompatible with our current power structures, yet those current power structures desire the AGI. This cannot end well. Stagnation and stability would be preferable to the level of misery that would follow.
I generally agree, but the main reason I am pro AI (or at least feel it is necessary) is because I do think it will end up helping us solve other problems that I think are existential in themselves like climate change. Maybe it won't happen, but personally I think it will and don't want to go into more detail because some of it will probably seem far fetched but I do think it will happen now. And previously these problems seemed intractable or were only being solved too slowly for us to escape the worst outcomes. So now I think we have this situation where to solve some of our worst problems, we are now creating other problems...
AI could solve the planned economy problem and bring about the endless consumption for consumption's sake that's destroying the planet, but there isn't any incentive in it for anyone with a large enough stake in AI.
That's part of the reason why I really appreciate China just throwing a wrench into the closed model system. As long as there's a profit motive for AI, it will only accelerate our current problems.
Well it could, but planned economies are bad because they are not creative, which neither is AI... In fact the entire reason we are having this conversation is because companies doing unrelated things figured out how to do AI and pivoted to it in an unplanned economy, but yes China obviously would like to prevent the US from dominating in AI and it is in their (and perhaps everyone's) interest to release open models.
As it stands right now, AI is massively accelerating climate change, by consuming lots of energy and keeping fossil fuels alive and thriving. AI solving climate change is believing we'll invent new cows that are capable of reverse farting and do methane capture.
There is only one way AI will solve climate change. If it leads to job losses for 80% of the people, so consumption drops to nothing. Nobody drives anywhere, people stop buying things they don't need with money they don't have to impress people they don't like (because EMIs/BNPL need income). If everyone goes down to subsistence level, back to the feudal system, just 100 - 10K billionaires and everyone else low status barely surviving, TFR will drop to under 0.1, delivering the climate change solution.
Even if you ignore it impact of AI power consumption, just the opportunity cost is enormous in terms of climate change.
AI is what we’re doing instead of solving climate change. With the money the US has spent on AI in the last few years, you’d have approximately enough to fully remove all fossil fuels from current US electricity generation and replace them with renewables.
Climate change will be solved—if we do it soon—as soon as people start treating it as inevitable no matter how many dollars it costs the same way they do with AI. It will not be solved as long as we keep ignoring the existing, known solutions and hoping that some unrelated technological advancement will just happen to make it cost less to solve climate change.
Indeed. The trillions of dollars invested so far and the trillions to be invested in the next few years would've been enough to switch the entire world over to solar (batteries, EVs) as well as fund a a few dozen Manhattan scale problems - develop a printable (like fabric/paper) solar, organic solar (so it can be grown like yeast or bacterial growth). So MANY things could've been accomplished ... instead everyone is doubling down on meme stocks.
So far the main problem with climate change was that people didn't want to believe it was real.
If people get used to trust what a smart sounding oracle tells them and that's consistently aligned with wha climate science says, then perhaps we have a hope.
If we end up with having different models that Say different things to different people then people will select the ones that reinforce their echo chamber
we could stop consuming so much and growing at all cost but no, we are being pitched the (idea|dream) that AI will come up with a technological fix to the problem that allows us to never change our behavior
Plenty of low hanging fruit. I'd start with cooling buildings down to 24C in summer instead of 15C (insulation also tends to help). And how about not bagging Every Single Grocery Item in yet another throwaway plastic bag? The US are kind of bizarre in many ways like that.
Plastic wrapping many items actually reduces the items outputted carbon. If you stop plastic wrapping cucumbers, they last 60% less time on the shelf and you end up having to grow far more cucumbers with all of the carbon emissions that comes with.
Plastic pollution is a different problem in itself but one that often is counter to fixing climate change in the short term.
I don't know but here's a counter-view. AI right now isn't really smarter than people. It is smarter in some niches but not others. It scales some things by allowing cheaper intelligence to work more broadly. We already have 8.3B of "intelligent units" on this planet. The "intelligence bandwidth" of AI is still a lot less than that. We already experienced increases in "intelligence bandwidth" by increase in number of people, better education etc. The point that I'm making is that an increase in that bandwidth on its own isn't necessarily a problem. If we look at AI "smarter" than us then we don't know but we do know that we have people who are "genius" (e.g. much smarter than the average person) but that on its own doesn't unlock any magic. If AI has an "IQ" of 200 or 300 or 400 we don't really know what that means in how far it can push on any given topic.
Our economy adapts and find equilibrium when things change. We don't need AI to get oppressive societies. We already have those. Does it make it easier to have one? Does it make it harder? I donno. But a lot of this depends on governance and policies. If AI is used for "good" it could mean a better life for everyone. That's not necessarily opposing the interests of AI companies or their owners. What sort of jobs will those people have and how different will those be than the job market today- I donno. Maybe we'll need more software engineers because we'll be building x10,000 the amount of software or maybe we'll be part of a service economy or planting trees. But the economy isn't as simple as "AI can now write code therefore this is the end of software engineering" just like making software engineers a lot more productive (and we have) didn't reduce the number of software engineering jobs. It created more.
In terms of AGI's compatibility with power structures. I don't think there's one answer there. Does it fundamentally change the different forces driving us? If AGI is smarter than us then what does it want? Clearly a lot of what AI offers today is still tied to existing systems. Can spot cancer in scans feels like a good thing. Somebody has to pay for it otherwise why would anyone build it? Like anything else there will be supply and demand and if supply is high and demand low the price will go down. Why should it necessarily drift more towards anarchy or totalitarianism? Why can't it be "free/capitalist + AGI"?
Sorry can you clarify the 10T revenue part? Apple makes about 0.5T revenue and has 5X the valuation. Pricing on revenue multiple you’d say anthropic needs 100B in revenue, which at 16B last quarter isn’t too far fetched.
Apple's P/S above 10 is incredibly high. It reflects high margins, monopoly power, network effects and users with high switching costs. So it's probably not unjustified for Apple.
Anthropic has high margins and is working on the others but hasn't achieved them yet. It's margins are threatened as it and competitors are lowering prices regularly. P/S for commodity producers is well under 1. But lets assume tokens don't completely commoditize so a P/S of 1 is justified.
That's $1T. But Anthropic doesn't have $1T in revenue yet, and it's risky due to spending commitments. So you have to discount for both time and risk. That's where the other order of magnitude comes from.
I'm very loosely using a 10x multiplier. So a 1 trillion valuation -> 10 trillion revenue.
The standard P/S ratios apply to established companies, but Anthropic is not that. The current valuation is not based on the material state of the company, but on future projections.
It's not just enough to have enough revenue to "justify $1T value", there needs to be massive growth in it to justify the risks.
All companies current valuation is based on future projections! You'll be shocked by the unsophisticated nature of these pricing models.
There's a lot of information about there. You'll be disappointed by the simplicity of it but generally speaking analysts will be using a DCF model with a terminal growth rate. It unfortunately doesn't exceed the math level of grade 10 in my country ;) On the plus side, it's very easy to understand.
...and even that is optimistic on the sophistication front. Sometimes markets go down because someone bet $50B at 4x leverage on a trade that wavered and got liquidated. So it goes.
> Doing 'redundant' work doesn't improve the economy, because there's an opportunity cost doing to anything
I know this is only a small bit of your overall point, but I don't think this is an accepted fact of economics.
The "opportunity" cost is true is you're comparing redundant work to productive work, but OP was comparing to unemployment, which is a different case. A bit of a simplification, but less employment leads to less spending which leads to less money to companies etc.
Here's a nice quote from Keynes about redundant work:
>
If the Treasury were to fill old bottles with bank-notes, bury them at suitable depths in disused coal-mines which are then filled up to the surface with town rubbish, and leave it to private enterprise on well-tried principles of laissez-faire to dig the notes up again (the right to do so being obtained, of course, by tendering for leases of the note-bearing territory), there need be no more unemployment and, with the help of repercussions, the real income of the community, and its capital wealth, would probably become a good deal greater than it actually is.
> Broken Window Fallacy. Doing 'redundant' work doesn't improve the economy.
The problem with AI is that work on an unprecedented scale will be made redundant and economic power will be concentrated into the hands of very few and there will be no way out in perpetuity.
Being able to buy massive intelligence for cheap (or running open weights on your own HW) doesn't give you economic power? Only if you don't use that power to provide something useful to others.
1) The AI companies maintain massive amounts of control over that "intelligence for sale". Not the least of which that they could simply refuse to sell it to you. It's also a big part of the sales pitch; A company can use AI to externalize the "institutional knowledge" and skills of of it's workers into something the company controls rather than the workers.
2) Because it's cheap, it has no value. Anything "useful to others" made with AI is currently drowning in AI generated submissions.
> Only if you don't use that power to provide something useful to others.
The power you bought is a drop in the water, because if it really is power, the AI overlords will just copy or they'll have vastly better economies of scale.
Knowledge simply is no longer a bottleneck, we're back to raw resources, real estate, political power and so on, medieval stuff in political terms.
We've had an industrial revolution before. It led to over a century of immiseration before the workers negotiated a better deal. "It's different this time" is uphill of the evidence.
If we stay with your analogy, would you then agree that the pain is temporary (at civilization scale at least) and that the prosperity on the other side is worth it?
Absolutely not: I reject your ridiculous premise that the pain is necessary to anywhere near the degree to which it was dealt. We could choose to constrain the most rapacious actors this time.
You are over-interpreting (what I’m trying to communicate) in a pessimistic way.
Society is much more enlightened today that it was then.
So I’m optimistic that the pain will not be as severe or as prolonged as then. Doesn’t mean we should just hope but I simply don’t share your extreme pessimism.
The one thing I would concede is that there’s a lot of uncertainty about the exact contours of what we’re approaching but that was also true of the Industrial Revolution or other transformations earlier in humanity’s history.
You are over-interpreting what I am trying to communicate in a pessimistic way. I'm a "tax capital" guy, not a "stop AI" guy. My point was that it's often reasonable (if not critical!) to take bold action in the absence of certainty.
How? A full server of GH200s is half a mil. That runs GLM5.3 Prime and Delegate Flash models.
Half a million, in 2026, is a lot for me and you, but pennies for even a modest sized business. And this is before prices go down.
Ive been using 9b models on my 4090 for coding and command tasks for over a year now. It's great! You obviously can't be like "Use huge framework and context, and just make me an app" but you can be like "Employee said vpn is running slow. Follow the process documented here: X to diagnose and propose a fix" where X is docs written by a Chinese model for 10 cents.
Did you mean "commoditize your compliments"? The textbook example would be NVDIA giving away free AI models.
What you are probably referencing is "dumping". That's when you sell undercost to disrupt your competitors.
That may be why. Certainly the motivation of the CCP is supporting it isn't short term financial gain. But it's probably more about maintaining independence from a dependence on US AI long term than dominating the AI industry.
But I think it was originally just a side project to keep quants interested.
I am a big believer in the solow paradox and I don't think AI changes anything about it. Just more money down the tubes.
Go read the papers on it if you want to hear some of the immensely plausible theories as to why micros and the internet and all sorts of what we think is technological innovation doesn't show up in economic productivity data.
The Broken Window "Fallacy" assumes that distributional problems don't exist, lol.
I know why rich people who own many shares of aggregate economic capacity want me to believe that it is the only important thing and that sacrificing aggregate capacity to fix distributional problems is never worthwhile. But I don't believe it. The industrial revolution was gnarly with unnecessary distribution problems that ensured crippling poverty for the masses amid extraordinary leaps in aggregate capacity. This time around I don't want a century of Cyberpunk Oliver Twist before the peons are allowed a dignified life.
Adjacent, improved argument: taxes are a better fix for distributional problems than throwing stones at windows. I agree, and I will add that broken windows are a better fix for distributional problems than guillotines. I really really hope that we can convince capital to pay taxes because that's the only future where things don't go to monumental shit before getting better.
Exactly. Let's say there was one guy in the world who owned 99% of all the wealth. Is it really a bad idea to break his window so that he may disburse some money on getting it fixed?
Now, let's say the handyman whom the rich guy paid to fix the broken window uses that money to start a company that successfully competes with this 99% person?
If you generalize this, the conclusion is that broken window fallacy assumes that the pie needs to grow in a monotonically increasing fashion, when in reality sometimes the pie needs to shrink a little in order to grow a lot, if that shrinkage results in redistribution of wealth.
The better idea in that case would be to create a new currency, distribute it equally to everyone, declare the old one worthless and thereby reset the market.
But without taking assets, prices immediately go back to how they were. Rich people don’t have currency, they have assets. And I’m not giving up my house which I bought with (a loan of) old currency.
Yes, inflating away the rich guy's wealth and distributing it to other people would be the "correct" way, but perhaps one that isn't available in a libertarian-leaning society.
Whether fixing the window is "wasted" time and resources depends on supply-side slack. If there's a warehouse with millions of windows, or the handyman is underemployed and has nothing better to do, then that's not necessarily a linear waste.
> Yes, inflating away the rich guy's wealth and distributing it to other people would be the "correct" way, but perhaps one that isn't available in a libertarian-leaning society.
That doesn't work when the rich guy's assets are factories and datacenters, because those will not be hit by inflation. Inflation only dilutes fortunes that are no longer invested in profitable operations.
I think it's more compatible with libertarianism than breaking the rich guy's windows. It also doesn't even have to inflate his wealth away, if the amount of new money printed merely cancels the deflation caused by mass unemployment.
The money printer taxes a very specific slice of assets -- cash, bonds, negotiated salaries, accounts receivable. It does not tax stocks, commodities, and real estate, and in fact inflates their value due to the cantillon effect. If the problem is that stocks have sucked the oxygen out of the room, the money printer is a bad way to get the oxygen back. A better designed tax would target where the money went, not avoid that location or exacerbate the problem.
Anthropic assume Deepseek, Kimi, Qwen, Tencent don't exist and OpenAI is shrinking. Reminds me of Elon laughing at BYD decade ago. What will happen, USA get boxed in while surrounding countries trade with China. You can just look at 2 items USA was champion off for nearly 100 years: cars production and energy production. China is like the replicators in Stargate, they just peoduce more and more...and cheaper too. Heck even data centers, China now completed way more than what US able to this year. And they now start building EUV machine after their manhattan-like project euv prototype succeeded a couple months back. In 2029 they will have a woman on moon. Meanwhile I am still waiting Elon (not NASA) to do it.
The US keeps thinking its businesses own the world - and being very lazy about expanding and competing anywhere that isn't the US and europe. Anthropic's investors fell into this trap too. At least Nvidia seems to take China seriously, but can't really do much there legally because of us laws.
In reality China is taking market share all over like crazy with cheap, good products and services.
Someday I think we'll wake up and realize how much wealth we threw away from sheer lazyness and failing to do something about our bloated monopolies that have little reason to compete or innovate or expand - and being a big stupid bully that pissed off everyone else - but by that time we'll have lost most of the world's business to china, cars, software, ai, aerospace, online shopping, etc. All gone.
Silicon Valley culture is deeply, deeply adverse to truly competitive markets because they are inevitably low-margin, meanwhile China's companies will work into the ground to shave off a fraction of a %. The age of software as a high-margin business is coming to a close and the fat, lazy and bloated US software mega-corps are going to really struggle.
Over the long term the economy will grow. That I am 100% confident of. When you take previously scarce resources and commoditize them, the knock on effects tend to spread widely. Intelligence and automation used to be enormously expensive, and now they are becoming cheaper by the day. Right now, frontier LLMs, which I consider wildly inefficient can give you mostly functional blueprints and diagrams for working products, like basic electronics on top of the image generation and code generation that is increasingly being commoditized.
The goods and services built atop "tech" should start to become cheaper as well. For knowledge workers this is terrible, but for consumers, this is great. I think folks worrying about an economic collapse are only looking at this from the perspective of knowledge workers and not from the much larger population of consumers.
The effect on the AI companies with pure competition is to make them basically worthless unless they find ways of stifling competition. For example by banning open weight AI in the name of safety, forming a global monopoly, buying up all the RAM capacity in the world, etc. All of this stuff is going on right now.
The value is there. The true question is whether the wealth will be concentrated and captured or released and distributed. I'm in the camp the likely outcome is actually the former. A huge amount of the wealth of AI has not really been unlocked yet... consumer robotics, disease cures, universal constructor type machines. Furthermore, once we reach superintelligence, eventually someone will inevitably fuck up by letting the dam break by ablitering a model in the wrong way. This will give governments and companies rationale to lock up the best AI to develop frontier level IP.
Everyone here is focused on software development/IT which is not surprising.
But "AI" is not displacing only IT, it is displacing all service jobs.
Only jobs that currently cannot be replaces are manual labor and some trades...and you only need relatively few plumbers, electricians and hairdressers, specially when most people do not have an income.
So most of the current workforce will have little to no income and therefore no money for most of those new products, even if they are now abundant and cheap.
Considering no llm assisted chatbot has ever helped me and problems were easily solved after getting in touch with a human, we have a very long way to go. Those things are not worth the money and customers hate them.
Those things are not worth the money and customers hate them.
The cuts happening now are based on the belief that the AI will get better quickly enough that customers won't notice a degraded service long enough to leave, plus an understanding that every other company is making the same cuts so your customers won't have anywhere to go anyway.
I don't know about you, but I've had significantly more success with support chatbots that use LLMs than with all "self-service" support options before, as well as with the average human support center agent, and I love the ability to get that AI help immediately. There are still cases where I really need to escalate to something that only a human expert support rep can help me with, but given how difficult companies are making it now to talk to someone who isn't on minimum wage, and how quickly agentic AI is improving, I don't expect things to go back to how they were.
Funny, my experience has been the opposite (and I would like to stress, that I am _not_ happy about that). Two summers ago, the SIM card in my phone stopped working, so I ordered a replacement SIM from Verizon, which arrived in a few days. I popped it in, and was unable to use my phone. Apparently, in order to "activate" your SIM, you need to use the web interface, and then you need to type in the activation code that Verizon sends to you, via SMS (without option for an alternative delivery method).
It should be obvious, to any grown and civilized human, that has more than a gram of brain, that it is impossible to receive SMS messages, without access to the cellular network, which requires a SIM card! (You see, you do not strictly _need_ an AI/LLM to get systems that appear to have been designed by drunk aliens).
So, of course, I try to contact support. I was given a chatbot that had many great ideas, that were mostly hallucinations (e.g. it would give menu-instructions that make no sense, for any phone model, that has ever existed). Eventually it gave up and told me to call support.
I called support, and got another chatbot, with much the same results. After spending hours, and getting pulled by the nose by various support-bots (none of which ever presented me with an option to talk to a human), I started screaming into the phone (this is my other phone, that does actually work, and has a different number): "I WANT TO TALK TO A FUCKING HUMAN NOW".
There was a silence on the other end, for a good 10 seconds, before it started ringing, and I got an actual human on the line. I explained the situation (and by the way, their English was impeccable, and they were very knowledgeable), and they guided to the very precise steps to activate my phone (and to avoid a pitfall that emerged from my previous attempts (e.g. they said to discard the new SIM, and to activate the eSIM, because my previous attempts had put the system in an invalid state that would be difficult to recover from otherwise)).
I want to make it very clear: I should not have had to talk to a human, _nor_ a bot. I should have been able to insert the fucking insertable SIM[0], and been connected to the network within minutes. The _excuse_ for the exceptionally poor engineering was something along the lines of _precaution against identity theft_ (which, if it is truly the reason, is even more embarrassing and shameful).
[0]: By the way, the original SIM should not have died at all, given that it was in the phone for only a few years (my other phone is a _decade_ old at this point and is using the same SIM card, that it had, when I bought it).
In my experience, this is because those chatbots don't have access to any data, nor are they allowed to perform any action. I suppose that is going to change in the next 10 years. The llm chatbots of now are far more knowledgeable than yesteryear's human first-line support agents. Alternatively, I can see those human support agents level up dramatically. I predict that one way or the other the bar for any organization to meet user's expectations for customer support will have risen by a great amount.
No, I fear those jobs will face more pressure or at least be heavily assisted by AI if not outright replaced by AI that has access to systems.
> AI support agents are god awful.
I don't doubt your experiences, but I guess this is due to different implementations. I experienced some that have real world knowledge, operating at a Claude level. I am obviously not talking about the useless bots we had for years that just regurgitate the same learned sentences based on some trigger words. They are god-awful indeed.
The AI support agents that are currently being used are terrible but how many of those support agents you've interfaced with do you think are actually running on the latest models?
I have the same opinion of all consumer tech support. It really doesn't matter if it has positive reviews from intelligent critics it matters that it has statistics that are improving and costs that are low.
> Considering no llm assisted chatbot has ever helped me and problems were easily solved after getting in touch with a human
In some cases, you wouldn't even know if an LLM was "helping" you. And what makes you think that the humans who solved your problems didn't in some way use AI?
The idea that someone won't come along and eventually build a robot system to do plumbing, an electrician's job, or hairdresser work is not realistic. Those too will come to pass. AI will come to encompass all human labor.
In either the "frontier labs capture all the value" scenario or the "competition commoditizes everything" scenario the net value to the economy increases. The distribution of that wealth is at issue not the wealth itself.
Lets say labor is displaced in both scenarios. That does not mean comsumption halts. Capital still needs to find a home. It can be overpriced watches for the techno elite... vacations to the surface of the moon... anti aging drugs. Lets say no one company captures the value, then the value goes to the owners of machines doing the actual work.
they will always have something to offer (it doesn't need to be "fresh"), just usable. Their labour or body, for example. And it's not like AI can automate every possible job - jobs like toilet cleaning, and a lot of service/hospitality work is hard/impossible to automate (at least in the medium term - may be long term we will all be served by androids).
The influx of labour from the (now redundant) high skilled professions means you now can have cheaper labour for all these services.
> codify the asset class
this has and always will be the case in a capitalist market. Unless you propose revolution (of the communist kind).
That implies that a) more toilet cleaning jobs will be created to accomodate knowledge workers who have lost their jobs b) they will start paying more than now so that the total disposable income of working population does not suffer and economy can continue to run
neither your a) nor b) scenario is implied in my comment. I merely said that as more people move out of knowledge work from ai, they naturally will have to turn to labour type jobs that can't be fullfilled by ai. The influx will mean lower prices for these jobs (in my tongue in cheek example, the toilets that aren't currently being cleaned will be because the cost of hiring a cleaner drops).
That returns us to the point you were responding to. People's (even of those people who were previously in those cheap "irreplacable" jobs) consuming power will shrink, so economy will shrink as well (unless you redefine economy as something bizzare that doesn't include humans).
When regular people have less money the service/hospitality will shrink so there will be less of this kind of work. What's left of it in the extreme case is house servants for the owner class.
I think the simple answer is that it won't. Look to other tech revolutions for precedent. The biggest winners of the Industrial Revolution were not mechanical device peddlers, or even the 'super farms' that utilized those devices in an obvious way to increase efficiency/scale/etc. The biggest winners were instead the people who achieved previously unimaginable things on top of all of this.
And I think this (in our revolution) will also generalize to 'cognitive work' companies as a whole. This is why I think LLMs are going to radically (and positively) reshape the world, but in a way that relatively few seem to be predicting for now.
If labor income is decimated, there's a shrinking demand base for entrepreneurs using this technology to augment themselves. Maybe a select few make good money before the bottom falls out or get support from the remaining consumers of the physical economy (healthcare, etc).
All labor is, is providing things that others value. When what's valued in a market shifts significantly, so does the equilibrium of the economy. The value of cognitive 'output' approaching $0 will ultimately just open a large vacuum for other things to become more valued, or to build things on top of this that previously would have been unimaginable.
Even in that scenario, which we're still quite far from, you're opening the door to more 'social' labor, entertainment, art, sports, and the like. And these are all probably very myopic predictions. It's hard to predict a future that has no basis in the reality that you know. It's like trying to predict the implications of the Industrial Revolution from the perspective of somebody who'd spent their entire life in an agrarian society.
And yet agrarian societies have largely given way to industrial ones, many in the span of two generations.
Meaning radical shifts in short order are certainly possible.
> you're opening the door to more 'social' labor, entertainment, art, sports, and the like.
That assumes money hasn't been captured by the capital class. Today how many working class people can afford to attend professional sports, live theater, live music? Even viewing the UFC fight at the White House--supposedly a celebration meant for the enjoyment of all citizens--costs money.
I do agree with you and expect that physical work will be the next target, but there's just a lot of technological progress remaining yet for the bodies to catch up to the brains.
As for the Marxian stuff, right now the most valuable 'means of production', by far, is what you and I are both using right now to type these message, and what basically everybody has, or can comfortably afford with like a week of minimum wage work. The biggest remaining asset of the 'capital class' are cognitive workers, but LLMs are going to similarly make that asset widely available.
Society has become dramatically more unequal in terms of numbers, but in terms of 'stuff' we're likely more equal than ever. Royalty and peasants lived dramatically different lives. Yet your life, assuming you're just another mid-class normy, and the life of a billionaire aren't fundamentally different. They have more liberties and nicer things, but on a day by day basis you're often doing pretty similar things, eating pretty similar things, and just generally living pretty similar lives. I think LLMs will further decrease the 'stuff' divide, and further increase the numeric divide. That doesn't really bother me.
Re: Society has become dramatically more unequal in terms of numbers, but in terms of 'stuff' we're likely more equal than ever. Royalty and peasants lived dramatically different lives. Yet your life, assuming you're just another mid-class normy, and the life of a billionaire aren't fundamentally different.
1) you have no clue about the lives of billionaires
2) there are not going to be mid-class normies, with their work replaced by AI, they will become trailer trash scraping for bare existence
“Productivity isn't everything, but, in the long run, it is almost everything."
is one of the better quotes for grounding yourself when people start talking about economic problems to big to know exactly.
to your question.
If an industry is 'worth' 1 billion, meaning it generates 1b in revenue lets say at 800m of investment.
and that drops to 80m of investment and 100m of revenue due to productivity improvements then competition, the economy is now free to invest that extra 820m in the next best productive use for that capital, while still getting exactly the same services.
If the companies investing in ai can manage 80m of investment and 300m in revenue as the end state. They're richer, the consumers getting the same service cheaper, and society is richer due to the extra 700m being used to produce other goods and services. Thats the dream.
The short answer to your question is: yes AI and technology generally is deflationary. The market needs to make up the delta in applying this newfound productivity.
We were lucky the maturation of the internet coincided with post-GFC low interest rates or else the internet may have caused deflation.
> If AI disrupts an industry that previously had high wages and now can be automated by commodity AI products, would it not actually lower GDP in some way (initially anyway)?
You can solve it either way. Anyone claiming to confidently know the answer to this question is making overconfident assumptions about the relative velocity of money between AI shareholders and consumers, export (and thus geopolitical) dynamics, resource price elasticity, social-safety net politics, et cetera.
There is a further cascading effect. Say that a specific $100b vertical is gone or severely reduced. In the worst case all the people working in that vertical are now unemployed. These same people now have no money or very little money to purchase or invest in other industries/verticals. Which means that there is a downturn in those verticals, leading to cost cuts and employee reductions. And those people now have less money to spend on other industries/verticals. And so on it goes.
But if the AI makes it cheaper to develop and produce the products, they can lower the cost, so those with less money can still afford them. I don't know if anyone knows how that plays out in the long run. Surely it must reach some sort of equilibrium.
Except the products are barely priced on how much it costs but rather who can pay. Lmk when 365 Office or Salesforce reduce their prices to match developing economy prices after their generous use of AI.
Historically, driving down the price of a product/service drove up consumption of the product/service. So markets will tend to grow to meet demand up to a saturation point.
In some markets, AI driving down costs will drive up consumption. So people won't necessarily lose their high paying jobs, they'll just do more work. Or, at least, corporations will capture the same revenue.
However, what's really untested is what the commoditization of knowledge is going to do to society. Having an agent at your fingertips with essentially a masters degree in every subject on earth, and the ability to apply that knowledge effectively is a pandora's box.
The closest analog is what industrialization did to craftsmen. How are people going to be qualified for jobs when AI can do a better job than anyone? But I suppose that's a political issue more so than a technical/economic one.
Anthropic (and OpenAI's) value prop is to replace the labor for a cost. Their endgame is diverting the employee wages to them. And eventually the underlying business perhaps.
I kinda think any product that doesn't have a physical aspect is going to go extinct. Why pay a business for something when I can just self serve myself?
In lot's of cases yes, but I don't buy this argument in general. Unless AI can do close to 100% of the work ofc.
If it requires even some tiny amount of work from you, then you have just replaced a 20$ SaaS with a 20$ anthropic subscription AND you get to maintain it yourself AND you have no one to blame for mistakes. I would say that's a bad deal.
What typically happened is that the capital typically allocated to the disrupted thing goes toward higher returning things now and gdp increases. The economy isn’t zero sum so funds get reallocated at the macro level pretty quickly.
(See all other productivity gaining technologies that increase gdp).
Don’t forget that all the people that had the big salaries in this industry no longer have those salaries. That is money that they cannot spend that is removed from the economy.
Which is why other economic disruptors like rail roads eventually lead to economic expansions, and were net positive
It doesn’t not look like that is the case with AI. It’s going to cannibalize the economy and not bring the expansion of previous developments. Which is a sad state for humanity
Railroads disrupting a horse and wagon certainly would impact the salaries of wagon drivers.
AI isn’t a magical force that takes your job and prevents you from ever working elsewhere: it’s a technology that will increase productivity and impact jobs via creative destruction.
Could be. Going to be hard to scale current LLM architectures to the level ANT and OAI are promising. Small local models run with better user economics that throwing everything into hyperscaler land that have to pay off entire card stack re-capitalization every 18 months.
There is no real moat anymore unless they force it from the regulation side, hence the heavy lobbying.
> AI isn’t a magical force that takes your job and prevents you from ever working elsewhere
Yeah I'd say if you were being paid six figures to do some office work thing (including software development) and now only 10% as many people are needed to get the same work done, it is very much going to take your high-paying job and make it considerably more difficult to find another one that pays the same. At a minimum, there's going to be a whole lot more candidates to compete with as you attempt to work "elsewhere."
Unless it's a _very_ slow transition to whatever the "next thing" that "can only be done by humans" pops up, this "creative destruction" is going to be devastating.
Interestingly...I have a team of people who deal with processing a document that is shared between companies. Purchase Invoices. They have had automation for years. It is offshored to make it cheaper again. But they are still there...at most companies.
I think the reason is that we can automate what is correct, the invoice matches the PO, the goods are receipted etc.
But most of their time is now spent on exceptions. The company didn't charge what the po said, there is no po, the supplier is not set up, the payment got returned from the bank, some goods are missing.
Now they can't fix any of those things themselves by design, because of segregation of duty controls, and specific humans holding Delegation of Authority for changes. You see if you can order things and pay for them yourself (with the companies money), fraud is inevitable. Giving this to an LLM makes fraud inevitable.
Then we say lawyers and auditors... the personal liability for signing key documents in those industries lies with licensed people. An LLM could prepare files, but that licensed person still has to review them, and employ other reviewers to ensure that the file is correct.
I remember when I was a young man, meeting a socialist who seemed very incensed about the coal miners.
I just couldn't understand it. If society no longer needed coal mining, was he seriously proposing they should all keep their jobs anyway? I couldn't take in this way of seeing the world.
So I'm pretty sanguine about the same thing happening to my own career, if it does. In the end, you can only expect to get paid for providing a service society needs. I can't think of any other remotely sensible way of running things, myself.
You described an economy, not a society.
Economies exist to support society, not the other way around. No, it can’t all be sunshine and lollipops for everyone - things shift and what was once a solid career disappears. This happens all the time. But a society (or maybe a functioning society) doesn’t throw people away if they aren’t bringing in a paycheque every month. Which was at least part of what the miners strike was about. And every other iteration of the same thing back to the Luddites
Most people would agree that a healthy, working-age individual should either support themselves, work, or be in education intended to build economically valuable skills. If they aren't working, why should we be taking from those who are to support them? Society can only support so much dead weight before it breaks down.
Mismanaging an inelastic excess of labor into soggy demand in order to dumpster wages is not the way to treat labor if you actually cared for labor, so I can only conclude that you do not care for labor and dumpstering wages is your entire goal.
And importantly, we've grown so huge (population wise) that not everyone needs to provide a service in the first place, even if you can.
We need to move beyond the "everyone must contribute something" mindset, and instead place value on simply life itself instead of what that living being provides or does not provide because there simply isn't room for everyone to provide some service to society, everything is covered. The things that are not covered are generally high skill and high education (expensive) positions that not everyone has the ability to fulfill anyway.
Basically, and before we have mass unemployment and homelessness from AI, we need some sort of universal basic living floor as a safety net. Make sure no person is able to go without housing, food, utilities, education and healthcare.
The thing is that there is still so much that needs to be done (starting with proper care for sick/elderly/kids, but also environmental cleanup and maintenance). The problem is that there is no money to pay for all of that. Not because there is no money (after all, at the basic level, money is just an abstract construct to motivate people), but because the people who do have money do not care for those things.
I.e., this is the problem of how the economy is organized, not of not having enough people to do all those things that need to be done. But we are too deep into 'capitalism (greed) is the best' ideology to clearly see that and do something about it. The way the things are going, either pitchforks or terminators are going to come out. The AI industry is betting on terminators.
Conversely, I believe that the moral right to posess a larger share of wealth than other members of society is justified by the incentive that it creates for people to work, invent, take risks, and contribute to that collective wealth. This has been the engine of plenty for a long time. But if we are coming to the end of an era in which there is any useful work that humans can contribute, as automation begins to take our place, then there is equally no longer a need for an incentive to reward that work, and so the moral justification for wealth inequality vanishes along with it.
I think that's right. So if we end up in a feudal like system with all wealth concentrated in the hands of a tiny elite and no means to access it for anyone else, then not only is there likely to be armed insurrection/revolution, it is morally justified.
Either there's a moral right to earn money through whatever skills one can offer, no matter how productive. Or there's a moral right to free money.
Anything else, in the long term, amounts to murder through abandonment. It's no different than leaving a wounded person in a forest while the group marches ahead, even though the group has plenty of food and a stretcher and a car, and fit, healthy people to lift the stretcher into the car.
I think you have a moral right to offer to provide goods or services for people. They aren't obliged to want them though. I think it's on each person to find out how they can provide something people want.
Just because I enjoy coding or am good at coding, if nobody wants to pay coders then it's on me to find out what I can provide that they do want.
> it's on me to find out what I can provide that they do want
Sure. But what if that is "nothing"? Or something that pays so little you literally can't afford the necessities of life? Or it's only work you find objectionable (e.g. crime or prostitution)?
For all of human history, given enough time, money, and work, anyone could retrain into something that paid them enough money. "Try harder bro" used to be trite but true. In a true AGI/ASI world it may be an impossible demand. Is it still fair?
To be honest I'm only considering a world where my particular niche disappears, life gets harder for me, but it's on me to find a new niche.
I see that as perhaps possible/likely in the near future.
The idea that there will be NO services humans can provide which anyone wants and therefore no possibility of finding any niche is so far beyond any experience or knowledge I have I can't even really speculate but I assume it would lead to either insurrection or wealth holders providing UBI.
I believe every human has a moral right to the necessities of life. Note that that's more than "the bare minimum needed to physically survive", too—human beings need joy in their lives.
And it was never true. There are people alive today who get paid because their ancestors were friends with William the Conqueror after 1066. They were granted monopoly powers over things people needed, not wanted, nearly 1000 years ago.
Being in a competitive market is for us to bear but not them
I dunno, people love to talk about fairness on one end, but don't seem to care about it when you compare being born into poverty vs being born into a trust fund. All of a sudden life isn't fair.
Some people own land simply by virtue of being a distant relative of the first group of people who successfully defended (or took) it from other people with violence. Some people don't own land because their distant ancestors failed to.
It's just all so weird, when so much wealth is tied up in an asset that no human had any hand in creating.
The latter option implies that there is productive work that humans can do to contribute to the economy, otherwise the second group necessarily does not have any members.
please explain what service managers and executives provide to society
No one is saying bring back coal mines or CRUD software jobs. What people are saying is that the working class is being ELIMINATED from the economy. If you are in the working class your options are: service/construction work. Be an electrician, or a waiter, or lawn care professional. And they make jack shit for money. While the managers of AI make millions, so the economy shifts to them, away from people who do labor. But dont forget they all still have to pay taxes which fund all of the massive debt for AI and wars and all other BS.
You cant think of a remotely sensible way of running things because you dont understand what is happening. Society is COLLAPSING. The managerial class is systematically eliminating the working class. So enjoy your pretentious outdated theories about capitalists and socialists. While the world around you decays for the next 5 decades. I assume you're a bit older because (im in my 40s) thats the only people I see with this delusional view of society, they simply dont give a shit about the future.
Or option 3, go do something else for work. Wild, I know. Life isn't YA fiction where a magical sorting hat gives you a singular job for your whole life.
This is what happened to manufacturing and the advice offered to those folks at the time was: "learn to code". The modern equivalent would be "learn to take care of old people", because that market is going to be booming for the next decade or more.
> because that market is going to be booming for the next decade or more
Well guess what, unless you're willing to be paid less than Haitian immigrants to do that work, it ain't gonna be booming for _you_. Ditto for construction, landscaping, et al. You know, those jobs Americans "won't do."
…because 'old people' clearly all have generational wealth on the order of your typical AI tech billionaire and can hire servants and caretakers at white-collar salaries.
That market is absolutely not going to be booming. It's a societal problem that large numbers of old people are exploited, broke, and brainwashed by generations of social media. That's not a booming market, that's a captive army of peons valuable only as long as human votes matter. They could as easily be disposed of, and so could you.
> …because 'old people' clearly all have generational wealth on the order of your typical AI tech billionaire and can hire servants and caretakers at white-collar salaries.
Creative destruction assumes there’s a new job to find. That’s not necessarily the case and hasn’t been for decades. Like the security guard who works 80 hours a week and their rent is 55% of their pay being told “just move to management”.
Or like elevator operators, who used to make good money. There used to be 120,000 of them in New York - now there’s about 90 thousand elevator technicians, in the whole USA.
There’s not always a replacement job. Or there is, sweeping the floor.
60% of American employment in 2018 consisted of job specialties that did not exist in 1940, and “the locus of new-work creation has shifted from middle-paid production and clerical occupations over 1940–1980 to high-paid professional occupations and secondarily to low-paid services since 1980”.
There always is a replacement job, it’s not necessarily sweeping a floor, but it’s not 1:1.
People have to understand that your compensation is a function of the value you generate or capture: the security guard might work 80 hours a week, but a security camera works 168 hours a week and costs $100 on Amazon. He’s just not creating enough value over his alternative - the Puritanical thinking of hard work has inherent value isn’t compatible with the real economic world.
the current "real" economic world is the fakest manufactured system on the planet.
100 security guards cant make in one year what one person with 10 million dollars can "capture" from investing in the stock market.
The entire system is manipulated to benefit asset owners and disenfranchise the working class.
These types of lies and delusions are exactly where we are today. The "real" economy has been dying since 2007, everything else like new stock market ATHs are just lies the public has been fed that doesnt understand it just means: companies, banks, and governments are perpetually increasing debt and kicking the can down the road.
1) Your math is wrong. 10 million in capital at a 7% rate of return is only $700K, divided by your 100 security guards implies they're getting only $7K a year. Pick your rate of return as desired, but the math still doesn't work.
2) 10 million in capital creates more value than 7 or 70 security guards and is rewarded accordingly. Again, your alternative as a security guard is some cameras and alarm systems, physical security, etc. $10M can secure a lot of premises for a one-off expense, without payroll and worker's comp costs.
3) The "working class" is just fading in relevance as unskilled labor is pressured on all sides: international manufacturing, outsourcing, automation, and technology. If capital or intelligence can deliver more productivity, we shouldn't kneecap it just so a high school dropout can profit while leaving society poorer overall.
> If capital or intelligence can deliver more productivity, we shouldn't kneecap it just so a high school dropout can profit while leaving society poorer overall.
I think we should do exactly that. Besides, that same wealth relies on stability which is only found if access to a good life is easily available.
If we keep going at this rate, the public burning down data centers will be the least of the worries of the wealthy.
The economic notion of value is wealth-weighted. Feeding an orphan has zero "value" because the orphan has no money with which to pay. Merging the banks into a mega-bank to hike fees and offload gambling debts creates oodles of "value" because it makes rich shareholders even richer.
You are trying to slip a self-serving virus into the boot code of this discussion and I do not appreciate it.
It's absolutely fake and pretty stupid, and the more people know about it the better chance we have to destroy it.
I don't care what the argument is, nobody should be unable to afford an apartment after working 40 hours a week, much less 80. We need to correct the system, forcibly, and stop pretending this is a "divine invisible hand" we can't simply shake.
> Railroads disrupting a horse and wagon certainly would impact the salaries of wagon drivers.
Railroads didn't do that. Horses and wagons remained in wide use after the advent of railroads as a last-mile mode of transport.
Before railroads moving masses of goods and people typically required navigable waterways. Either natural ones or canals. Railroads augmented waterways and allowed that mass transport to expand far from those waterways. The underlying steam engine also helped with expansion as a factory or processing plant could be built near a resource, be powered by a steam engine(s), and ship finished/processed goods via those railroads to some market.
Railroads (and steam engines) expanded the economy far more than the supplanted any particular part. When they did supplant something the economic expansion largely provided new work for whoever's job was obviated.
Knowledge work replacement by AI doesn't necessarily offer the same economic expansion. If it replaces one type of knowledge work there's little stopping it from replacing another type. OpenAI and Anthropic are only viable if they capture significant portions of the US GDP. Anything less and they can't even pay their debt obligations let alone actually make money.
There's no real indication all those replaced workers could find any other jobs, certainly not ones that make as much money. And UBI will never happen, the billionaire class will never pay the taxes that would be required to support it. So that leaves lots of people left with no way to support themselves.
> Railroads didn't do that. Horses and wagons remained in wide use after the advent of railroads as a last-mile mode of transport.
Many horses and wagons lost business due to trains, just because some remained doesnt change the fact the others lost their place in the system.
Proper comparisons are about numbers not nothing changed because of one cherry picked item.
> Knowledge work replacement by AI doesn't necessarily offer the same economic expansion.
Most people using AI are not knowledgeable. A lot of AI seems to be dumb animations or flyers or silly graphics or emails nobody reads, or more powerpoints. Thats not knowledge those lines of work, are a tax on everyone else.
Nobody needs advertising, or pointless sociual media type communications.
By and large trains didn't displace horses, trucks did. Trains provided long haul service while horses provided last mile service without a lot of overlap. You can see this today in countries where horses are still fairly popular like Cuba and Romania. You'll see an occasional horse on the autopista but it's not until you get into cities that horses are common enough to become problematic (and you'll see signs forbidding them as a result).
Nobody needs advertising, or pointless sociual media type communications.
Ah, you don't remember the AI company plastering San Francisco with adverts imploring people to stop hiring humans? That was not an issue with trains.
Yes, all those unemployed software developers who used to make multi-six figure incomes are now 'free' to compete in what will certainly be a ruthlessly competitive job market with all the other white collar professionals who were automated away. Likely into much lower paying service jobs, as those will be one of the few areas of the economy where the 'human touch' is still valued.
This reads like the line 'The rich man and the homeless are free to beg for change'
No, but when AI companies insist they need to build out more datacenters to make better AI, we, as a society are well within our rights to ask if their goals would actually be a net good for society.
These people are literally saying their product is dangerous, and that it will eventually put a large portion of the population out of work. Until we change our economic system so that "people are entitled to a decent quality of life" err sorry, 'free stuff' in your words, even without work, don't think the voting public is just gonna stand by and let them do it.
It's rather callous, bad faith framing to say 'you're not entitled to free stuff' when that 'free stuff' is the very things they need to live, and you've destroyed their ability to acquire it through work.
The voting public is always free to make choices, including those that go on to hurt society (see the expansion of the welfare state, the ballooning debt, and forever-war, or in this case seeking to cripple technological progress). It doesn't mean they're right.
Also, nobody has "destroyed their ability to acquire it" - the market has set the price and they have to adapt. If I create terrible paintings, I shouldn't still be entitled to millions because someone else's paintings sold at auction.
Members of the voting public should also be free to call out a system that has a meaningful voting bloc (20 to 47% depending on how you want to slice it) who consistently vote for more free stuff, while contributing nothing to the system. A system where everyone takes and nobody makes is not a net good for society.
The Democrats haven't expanded the deficit since LBJ. The bloc that votes itself free shit is the capital owners, who keep increasing the deficit to give themselves tax cuts. They have elevated the ratio of taking:contributing to truly epic proportions that exceed the degeneracy of even the fabled Robber Barons.
I like how you're parroting Romney's line without acknowledging that everyone working pays SS/Med taxes towards those 'expansions of the welfare state'. If anyone is not paying their fair share, it's those above the SS cutoff, and those who primarily make their money from cap gains.
You want to talk about 'good for society', maybe talk about the fact that we were on course to pay off our debt before bush decided to cut taxes for the very rich. While my pocket book appreciated the tax cut, I would have much rather had a fiscally responsible government than a tax cut.
What the neocons never mention is that the nordics do not seem to have any problem with having good social safety nets while being fiscally responsible. The rich simply pay what they owe to society.
If they turn a non-AI $1Billion industry in to an AI $50million industry then they have destroyed that industry, made $50 million themselves, and the people who were previously spending $1Billion between them are collectively $950million better off, which they will then get to spend elsewhere.
Ya but they won't necessarily capture the $50 million themselves since there are already several companies that have comparable models that could all likely perform the same tasks, so at best they would get a fraction of the $50m? And if anyone can run an open model on their GPU then the barrier to entry to compete in this industry would be so low as to be worth almost nothing? Seems like a race to the bottom, like ultra efficient competition where the margins would be very low...
>the people who were previously spending $1Billion between them are collectively $950million better off, which they will then get to spend elsewhere.
It's generally worse for one man to have a billion dollars than for a 10,000 to have 100k. Those 10,000 people will spend that money where as the billionaire will almost certainly won't. An economy can only take so much 'investment'. Our economy depends on consumer spending.
If the top .01% get fabulously wealthy automating the bottom 90% out of good jobs, we will all be worse off for it.
> It's generally worse for one man to have a billion dollars than for a 10,000 to have 100k. Those 10,000 people will spend that money where as the billionaire will almost certainly won't.
When we talk about 10,000 having 100k we usually mean they have it in somewhat liquid assets: savings, annual income and even such illiquid asset as housing is something that can be sold relatively fast and invested.
Billionaire having 1 billion is usually means the same as company having trillion market cap: 30-50% of it can disappear overnight and nothing in the world will change really.
> would it not actually lower GDP in some way (initially anyway)? What I mean is, if you had an industry that was worth say $1 billion, but now that same industry could be automated and done cheaply then it is no longe worth $1 billion?
GDP growth is usually measured in “volume”, which means that the demand for a good or service doubles while the price is divided by ten, the GDP share for this good or service doubles even if the amount of money is 5 times less.
The reverse is also true though, which means if something becomes twice as expensive (say oil) and people starts buying less of it, it's share of GDP will diminish. Because of that, the GDP deflator is underestimating inflation, and overestimating actual growth in times of high inflation.
Yes exactly, the world is a better place, we have more opulence and ability to create digital assets at scale, but some jobs (e.g. graphic design, making websites, software engineering?) are going to get demolished. In the long term, it's good for society as those people can reskill and build more things. From perspective of EU, we need more opulence in the physical space, which is increasingly falling apart (roads, more houses, more infrastructure, etc)
A good previous example of this was the computerization of publishing --- the investment costs were balanced by letting employees go (though the compositors' union bargained for sinecures) and it took at least a decade for an increase in GDP to be seen as the folks were retrained in new work, and the publishers adapted to the lower costs by increasing the breadth of book titles which were greenlit and published.
Note that this was further complicated by a change in the tax law which made the backlist a liability rather than an asset.
The destruction you're imagining is the "loss" of the ~$1B in transactions that consumers were previously paying to employ $1B worth of jobs. Consumers actually just have that extra $1B to spend, and people previously making about $1B now have about $1B worth of labor to direct elsewhere.
Economies don't really lose jobs. Buyers lose demand for products, and employers lose demand for particular kinds of labor. That doesn't make the labor less valuable to anybody else, it just means their current employer couldn't afford it or use it. The same number of people with the same set of skills still want to work as before.
It's actually common for labor to increase in value even when it gets disrupted! What happens is a rebalancing of the labor market away from obsoleted industries/company/products nobody wants, towards those that leverage the disruptive technology. Consider dotcom: eventually some of those employees left tech, and the rest created the actually-valuable Internet businesses because they regrouped and started different things.
Otherwise when there isn't enough demand for employable people to find work, central banks lend (or print) more money until employers want to hire them, because they're leaving money on the table from consumers' unmet demand if they don't.
In short: economies don't produce jobs. Jobs are what happen when demand on the labor market meets supply; creating more aggregate demand is easy but even then, it creates demand for labor to provide what people want to pay for.
The Bank of England announced basically negligible productivity gains associated with AI (as in spend vs. output). This is in a highly computerised service economy which theoretically should be one of the highest leverage economies for AI.
Yes, nominal GDP growth would likely be lower at first.
Efficiency would drive Real GDP higher though. Goods and services produced with lower input costs leading to lower inflation. Plus, presumably, widespread layoffs creating a negative economic shock. (Labor surplus from AI displacement would hit services pricing too)
GDP simply tracks the rate at which money changes hands in the economy. I'm not sure the AI proponents touting 15% GDP growth are really aware of this.
Nominal GDP would likely catch up after the big shuffle/unemployment cycle. But probably not initially if we do see mass layoffs
> I am trying to understand something about the economy and AI. If AI disrupts an industry that previously had high wages and now can be automated by commodity AI products, would it not actually lower GDP in some way (initially anyway)?
Pre-AI, if you spent $X you got A units of output; post-AI:
* if you need 'only' A units, you will spend less than $X (technically lower GDP)
* but if the limiting factor was not the desired units, but your budget, you will continue to spend $X but get >A units (same GDP)
Additionally, if each unit is less expensive, this allows the end-customers to spend less on it, and reallocate some of their budget to other things, which would keep the overall GDP the same.
It will probably mean the AI companies will get part of that wealth from service companies in the US and pull it away from countries & companies exporting those services. Those service industries serve already "real" industries like: agriculture, construction, factories, governments, health, etc. For some countries with a service oriented industry it could be troubling. It will probably also lead to more software development for a while, but some wealth shift to AI companies will happen.
> If AI disrupts an industry that previously had high wages and now can be automated by commodity AI products, would it not actually lower GDP in some way
Yes, in the short term that's true. In the long term, the concept is that that particular activity generates less GDP, but frees up the human capital who can then grow the economy.
This is true for any economy, not just AI and the central contradiction of capitalism.
Automation raises profits for a brief moment until everyone else has caught up.
Once the new profit margin equalizes across an industry, competition sets in and margins shrink.
To keep margins up, labour costs need to fall further. That is why companies like Nike or Adidas produce in low-income countries. The investments in automation is such that profits can only be maintained through mass production, low wages, long working hours and raised demand through brand lock-in, classic marketing.
The current open bet is that equalization of AI technology can't be achieved across industries because of the important entry barrier: capital and compute. Monopolization should hold long enough to make the rest of the world market dependent on US AI infrastructure.
The rest will feel the devaluation of their work, they will get poorer - some straight away, some indirectly because of the shrinking customer base.
> The current open bet is that equalization of AI technology can't be achieved across industries because of the important entry barrier: capital and compute.
Compute isn't a barrier to entry for long; I mean, even right now, you can rent a H100 that is sufficient for most businesses, for use during business hours, for $500/m. As many users as it can handle, as many tasks as it can handle.
And that number is going to fall off a cliff once the hardware shortage stablises.
The example many people use is the cotton gin. When it was invented it greatly reduced the labor required for a certain part of the cotton harvesting-to-textile process. Some believed this and other labor-saving machinery would reduce the demand for slaves.
Spoiler: it had the opposite effect. Cotton became so much cheaper than demand exploded and the demand for cotton basically doubled every decade after its invention.
So AI is going to destroy certain industries and others will have much lower demand for employees. The optimistic view is that AI will be like the cotton gin. Those industries will get cheaper and demand will go up or new industries will form in their place, sort of like the telegraph, electricity, radio and the Internet created industries. The pessimistic view is that this won't happen. There will be a tipping point where labor supply will exceed demand and this will put severe downward pressure on wages and create an unemployment nightmare that we're simply not prepared for. The last time this really happened was the Great Depression and at least then, the majority of the population was rural and way more self-sufficient for food and the like, not that I want to diminish the devastating effects.
I'm in the pessimistic camp. I believe automation in general and AI in particular is destroying labor demand faster than it's being replaced. You see this with decades of stagnant real wages (in aggregate) and the net effect is that we're creating a massive wealth concentration in a handful of people.
As Alfred Henry Lewis put it in 1896, "the only barrier between us and anarchy is the last nine meals we've had." And now we live in a country with more guns than people so we'll see how that goes.
You are confusing some basic facts, the TechBros and software industry are not the same people, the AI industry can be worth a lot of money, and AI can be beneficial for them, but that doesnt mean everyone else is also seeing a positive.
THe primary moptivation for the Tech Bros is of course selfish, to pump and dump.
Stop pretending they are trying to help others, they are only trying to help themselves.
>If AI disrupts an industry that previously had high wages and now can be automated by commodity AI products, would it not actually lower GDP in some way
If it disrupted that industry, the things produced by that industry would become cheaper (if AI didn't make it cheaper than human labor then it wouldn't have disrupted the industry in the first place), boosting everyone's purchasing power for that thing. GDP calculations try to account for this, so GDP would go up (the same way GDP calculations try to account for inflation; a single product class becoming cheaper is like localized deflation).
goods become cheaper but labor income takes a hit too. you only benefit if you've got savings / investments but even then, stocks would tank if collective purchasing power goes down fast enough.
If AI makes production of a product/service cheaper for the producer, why in the world would they lower the price? Thats like the whole goal of a business is to increase production efficiency and capture the margin as profit. You're trying to suggest companies reducing costs with AI wont just turn around and pocket that extra margin. That's ludicrous.
Even in a competitive market no player will significantly reduce their prices because they want to maintain their own margins.
In a competitive market, there's a prisoner's dilemma where an individual actor lowering prices allows them to make more profit by capturing more of the market, even though margins drop and reduce profit available in the market as a whole.
If you can take 80% of the market at $1 margins, that's better for you than taking 10% of the market at $2 margins.
I've heard this framed in terms of "surplus value," a very old (Marxist) economic theory. The debate is whether AI, which is intrinsically a machine, is capable of generating it - machines classically _transfer_ value. The reason this is important is because surplus value is what a capitalist entrepreneur exploits (whether they realize it or not).
So we don't actually know whether it is good for the economy, I guess we'll find out after the bubble bursts.
> If AI disrupts an industry that previously had high wages and now can be automated by commodity AI products, would it not actually lower GDP in some way
Yeah. It does. Permanently. That's what happened when the robotics revolution decimated factory jobs. Everyone was pushed toward finance and tech jobs, but most had to go into gig work to survive. Even the tech sector was kept up by dumping endless ZIRP money on it. Now that AI is removing white-collar jobs that used to spend to keep what remained of the economy going, that's exactly what will happen:
The system will get clogged. Because no new jobs are being invented at scale. We don't have any 'AI biochip integrator' jobs. We don't have 'Human-AI Augment synchronizer' specialties. We don't have any of the jobs that the AI industry promised, leaving aside any sci-fi jobs that were predicted in the past.
Let's face it: The majority of human civilization's day-to-day work has been automated. White-collar knowledge work is being automated now. What remains is some blue-collar work like handyman and electrician jobs. They will soon be automated by AI-using small robots. Not that those sectors would support the millions who were laid off from knowledge jobs when they all try becoming plumbers or electricians anyway.
It's time to admit that what Marx predicted 150 years ago is happening. Capitalism is destroying itself with automation.
You can probably get the best idea of what will happen from looking at real examples as it depends mostly on human behavior. I'm guessing with those the GDP attributed directly to them went down but the people who worked there got jobs doing other things so GDP in aggregate was fine.
I have never "talked" to AI in a call center. It's always a person. If AI was "disrupting" call centers why would they still be having people answering phones? I have called like 4 different companies in the past month about freight deliveries, shipping dates, etc. Always a person answering the phone.
I don't know about translators, I don't interact with them regularly, but I suspect AI
isn't "disrupting" anything there either.
Any more "real examples" or are we just playing around with hypotheticals?
Anthropic flirted with the IPO price several times today and is down 21% since the start of trading. It appears the only thing stopping it from going lower are those allocated shares in the IPO not wanting to book losses by letting the price fall further.
There’s not lot of fundamentals holding up the likes of ANT and much of the broader tech sector. The lofty valuations are based on hopes and dreams that are looking rather iffy.
It’s also sounding like the Fed will issue more rate hikes soon. Therefore anyone without free cash flow will have to pay even more to get their hands on cash.
For any company without strong cash flow and GAAP profits it’s likely to get very ugly from this point forward.
I interpreted it to mean the greenshoe option Morgan Stanley exercised yesterday, which is typically used to stabilize prices, from my understanding [0].
There seems to be a lot of commentary about the timing of this floating around, and I’m watching it with curiosity.
Green shoes are soft obligations by underwriters to buy when the offering threatens to break within the first 30 days.
It’s never done to defend the book. You want to execute a green shoe by buying shares, worth more than the IPO price, from the company at that IPO price.
If an underwriter could reputable decline to exercise the green shoe in the market, they would.
Right, which is why I’m curiously following it: there’s chatter about this one, but for hot IPOs, the timing of this purchase doesn’t seem to be that extraordinary, and I’m not sure whether there’s anything here or if folks are searching for anything wrong with this IPO. I am ambivalent to the whole thing, watching and learning.
From purely a pricing standpoint and the strike price on Morgan Stanley’s greenshoe option, the timing makes sense to me.
Maybe within your lifetime since you're used to the tech IPOs but no it's not supposed to just dip immediately. This is a sign of insiders dumping because they want to cash out exceeding the buy pressure from outsiders who think the company has long term potential (and also the insiders who want to hold for the same reason).
If you want to talk about the initial price, they've iterated on it a ton lately, so assume that whatever they set it at was intentional, and none of this is a huge surprise to the company.
your own silly reddit regurgitated source tells you that 10% is freed for sale on nov 19.
also who do you exactly think owns the shares to be able to sell? unless you think billions were used strictly to pump and dump the next day in an amazing coincidence of coordination (illegal btw), the ones who sold were the major holders.
No it’s not what’s supposed to happen. The IPO is now broadly viewed as having not gone well and the blame games have begun with the bankers. The stock looks very fragile now.
A good IPO looks like trading starting at 5-10% above the IPO price and then gradually going up a bit more from there and holding.
Thank God for some sanity returning to the market. I know a lot of people are feeling like they lost money (myself included) but those sky-high valuations were always illusory and the money "lost" was never real.
In the real world, there's just no way that Tesla (annual sales ~1.6m, trending down) is worth more than Toyota (annual sales ~11.3m, trending up).
It’s only real if you cash out. If you buy at $10/share, it goes to $20/share, and then goes down to $9/share then you haven’t lost any money unless you sell at $9/share, which for most stocks would be reactionary. And, if said stock drops down to $5/share, it doesn’t matter unless you held a disproportionate share of your portfolio in it, which is bad investing.
Camp your money into passive broad market portfolios and sleep easy, maybe add some positions when we have market downturns. Unless you think the entire economy is illusory, then buy shelf-stable food and water.
If this happens in a very short span it might be true. But in the long run you are illiquid if you hold these shares and refuse to sell. The business can be absorbed at current market value, so you are forced to cash out (say Nvidia buys Anthropic in 3 years for peanuts), or it can go bankrupt (a lot of business that consistently go down in valuation end up bankrupting), or it can have you on negative return forever (Intel never reached their .com peak price and will never do, even ignoring inflation).
> Unless you think the entire economy is illusory, then buy shelf-stable food and water.
I quite smoking but during Covid it occurred that cigarettes are almost as good as gold as far as a universal bartering object. I'd stash a few cartoons away - it'll come handy come zombia apocalypse.
I would argue they’re even better than gold because they’re consumable and worth “less”, making them easier to barter with. Gold is ~$115/g (more in an SHTF scenario), and unless you’re buying a lot at once from someone who wants the gold it’s annoying to use.
My big three for bartering is alcohol, toilet paper, and hard candies. Cigarettes are great but I don’t smoke, so I’m hesitant to hold onto a lot of perishables that wouldn’t get used. However, like you said, a few cartons stored in a vacuum-sealed pouch in the freezer will last a decade and be good barter just in case. Same with nicotine gum and vape juice (although, juice itself well but the actual vape becomes the bottleneck, might research that later).
Entire economy is illusory. Whole market is manipulated to go up until it can't go up. That applies to those passive broad market portfolios. Basically I don't think there is much you can do. Don't live beyond your means, but still enjoy life. Either it works out or it doesn't.
I pretty much think it is all screwed anyway. So slightly prepare buffer and spend reasonably on what you want just now in life.
That's not what the tax authority thinks. More than one jurisdiction taxes unrealized gains. More than one jurisdiction does not give you money back for unrealized losses.
For one sense of real. Banks pretending their mark-to-model securities that the market valued as worthless was a problem because the market price isn't entirely illusory.
Have you? You took the risk when you bought stocks on margin, which was money you didn’t have. If your bet didn’t pay out, then that’s not on the market.
If you funded your position out of pocket then there are no margins to call.
Inflation isn’t real either. Your money hasn’t lost any value until you use it to buy something..
Prices of stocks and other assets are absolutely real and (besides the tax implications) there is absolutely no difference between your example and selling at $20 and then immediately buying it back before the price starts falling.
> to $5/share, it doesn’t matter unless
At any given moment it doesn’t matter at all if you hold cash or stocks (as long as they are liquid) besides the taxes you might need to pay if you sell.
> It appears the only thing stopping it from going lower are those allocated shares in the IPO not wanting to book losses by letting the price fall further.
This doesn't even make sense, why would you not want to book losses at the end of the year? It's the perfect time to do it.
It doesn't make sense from a technical perspective moreso than a tax one. The shares were allocated and are now owned by someone. Those shares are not in the order book. Nothing is stopping it going below $83 except buyers holding it above that price, which will eventually be exhausted if there's enough selling pressure.
Recency bias is still a thing. If you subscribed for a slot in the IPO, you'd rather wait until the end of next year to harvest your losses if you still have faith in ANT as a whole.
I was allocated shares so I'm aware of this, but that has nothing to do w/ what the comment that I was responding to suggested. They claimed people aren't selling because they don't want to book a loss. That doesn't make sense. Also note: People that were allocated shares have not seen a loss yet.
1. The underwriters buying to stop the price from falling further, but per reporting today they seem to be done there now so the stock is on its own.
2. The tiny float which inflates the valuation. The market cap numbers are highly inflated by the fact that if all the existing shares were actually trading on the market right now the price would plummet.
Fundamentals don't seem to matter. NVIDIA's bosses sex scandal is somehow making NVDA trade at 10x its PE while growing 60% YoY, while Costco trades at 40x PE at a 5% growth rate. Anthropic IPOS at 2T dollars while having a negative profit margin, and less revenue than Wal-Mart. This is ridiculous.
> NVIDIA's bosses sex scandal is somehow making NVDA trade at 10x its PE while growing 60% YoY, while Costco trades at 40x PE at a 5% growth rate.
How does this support your argument? I think it tells us the opposite. Nvidia's growth is short-lived and its future is uncertain - it hinges on AI spending and data center buildout continuing at the same rate, and none of the major customers being successful in developing their own chips (all are trying). Their financing deals also increase their exposure to macro shifts. Investors are (righly) pricing in that uncertainty.
In contrast, Costco is a relatively boring business, but if you invest in it, you're unlikely to lose a whole lot of money very suddenly, so a decade or two of boring growth is priced in. I assume you're cranky about your NVDA stock dropping, but fundamentals go beyond a snapshot in time.
Why is the growth uncertain? Previously NVDA made GPUs, which were for gaming. Today they make what is quickly becoming general-purpose processors a la Intel. This changes the market significantly and like Intel, turns NVDA from a specialized component manufacturer into a key computing component.
Whereas the GPU market attracted barely any competition; today there are at least 2 dozen startups attempting to slice NVIDIAs market. Several of these competitors are now themselves public (Cerebras). These are real indicators that the market expects real growth.
> I assume you're cranky about your NVDA stock dropping, but fundamentals go beyond a snapshot in time.
NVDA is reaching the point where they can take the company private with a fairly typical bank loan, while returning significant returns to investors. If the share price goes into double digits, it would be possible for a small group of investors to buy the entire thing. This is frankly ridiculous to think that NVDA is fairly valued at the present movement. My NVDA holdings are actually quite small relatively speaking, but the idea the stock is fairly valued is ridiculous. I've been buying more obviously.
1. Over-leveraged startups that look like they could implode at any moment, and
2. Big tech companies that, while unlikely to go bankrupt, are all actively building their own chips so they can ditch Nvidia.
It’s not a pretty picture.
Re the Costco comparisons, Costco is a reliable business that generates this thing called profits (which seem like an unknown concept to many in tech), and it has an incredibly loyal customer base that’s not going anywhere. My only complaint about Costco is that’s it’s too busy when I go there.
A pretty sizable portion of my portfolio is Costco and it has the added benefit that I sleep better at night.
Also possible protip. At least in my city they appear to open 15-30 mins earlier than they advertise. The near empty warehouse at that time is sublime.
Nvidia doesn't make GPUs. They make IP: the thing being rapidly devalued by their customers. Their place in the world is far from certain. If Nvidia ceased to exist there would be dozens of replacements in a single day to utilize TSMC’s capacity.
> a la Intel. This changes the market significantly and like Intel
Did you see what happened to Intel’s stock price during and after dot.com? They did great after 2008 or so. Yet their share price hardly ever recovered.
> 2 dozen startups attempting to slice NVIDIAs market
That means they are losing their moat and their margins will collapse?
Nvidia might do great just like Intel in the next 20 years and its share price might never surpass the peak it reached this year. In no way these are mutually exclusive.
Ah yes, if NVDA’s stock price goes to $80, a small group of investors will have no trouble taking out a trillion-dollar bank loan for the largest leveraged buyout ever by a factor of like 40.
I think Costco is growing faster than that, but obviously not at NVDA rates. The difference might be more that investors expect NVDA to face stiff competition for a lucrative market, while Costco is "old reliable" and will just keep on growing steadily every year like they have been for a long time. So NVDA's price reflects that uncertainty and Costco commands a premium?
It is more than that. Top Nvidia buyers like Microsoft, Meta, Alphabet, and SpaceX have poured roughly $800 billion to $900 billion into capital expenditures in 2026, they can't keep doing that every year.
Major tech companies are seeing free cash flow turn negative and are increasingly relying on debt to fund data centers. Wall Street expects Big Tech capex growth to slow from 76% in 2026 to 25% in 2027 and just 6% by 2028. For a high-flying stock like Nvidia, a sharp deceleration in customer spending growth is enough to drag down its valuation.
>Anthropic IPOS at 2T dollars while having a negative profit margin, and less revenue than Wal-Mart. This is ridiculous.
Remember the stock price of a company is only partially about what it is doing today, but also a huge amount what someone speculates it will do tomorrow.
Wal-Mart has been around for a long time. How often has it doubled revenue or profit or suddenly done something extraordinary?
The odds of that are very low. The odds of ANT doing it are significantly higher.
Same reason Tesla trades higher than Toyota. One may go up a lot (or down), the other is solid as a rock.
> The odds of ANT doing it are significantly higher.
Based on their IPO price investors seem to believe that the odds are about 100%. So the problem is that unless they do even better than that (in a rational market) you can only lose money.
It is ridiculous to talk of "NVIDIA's bosses sex scandal" (sic) in an age when we have numerous high level men and women from across the board of academia, business, government, fashion, ..., all partying on 'Lolita Express' and Jensen Huang has been involved in a "sex scandal"?
Obviously (seriously!) this guy was 'taken out' because his person or strategic position in this very strategic sector was "not aligned" with the powerful. The people who can motivate NYTimes and WSJ to start "sex scandal" trains into motion.
As an aside; always pay attention when "news organs" like NYTimes and WSJ do the 'marching order press releases' to push an idea into the collective stage. At times like this, there are gaps in the curtain that usually would have NYTimes and WSJ as not propaganda platforms for very established elites..
This was a corporate coup and that itself is clearly party of a much broader power struggle in the political space, c.f. Treasure sec. dumping on the Commerce sec.
It's not the coffee they have a problem with. It's every time ask for tea, you get a cup of something almost but not quite tea. That and the doors making satisfied sounds when the open and shut.
On the one hand I completely understand and embrace that there is no stopping progress and adapt or die and all that. On the other hand I understand that we've built a machine that can strip-mine all of human knowledge for the benefit of a very small number of people and that sucks.
Extrapolating from the last chapter of The Selfish Gene, one could argue that we humans were just one type of "survival machine" for memes, and now we're no longer the most fit - such that what we consider as "human knowledge" actually outgrew us as its hosts, and will take it from here with the help of other machines.
Are there groups discussing a gradual, society-compatible diffusion of LLM in society so that people don't have to fear about being jobless next month while the economy still benefits from LLM positive traits ?
I keep thinking a lot of jobs could be 10% easier and 10% better (speed and quality) with LLMs. Meaning we're all less tired and slightly more productive, without tearing up the economic fabric. That's the smooth rollout group.
Some part of society could benefit for a more aggressive use of LLM: energy, material, medical science could all benefit from computing cost collapse. Same for public services: justice, education, healthcare: if they can have more ways to reduce friction, fix gaps in tools and statistics they're happier, we're happier.
Surely this is not a new idea.. and people are discussing something like this somewhere, I just don't hear it.
It's kind of for the benefit of everyone. Anyone can use AI, often for free or open source and the same knowledge is mostly there in all the large models. Information wants to be free as they say.
Well, doing that in a fair way is a tricky issue yet to be solved. Maybe we could give an AI some rules how to allocate to each according to their needs and what is fair? I think traditional capitalism may need to be rejigged a bit. Though it's not impossible. You just need a good enough plan that the electorate will vote it in.
It is tremendously under priced! If you factor in the cost of all the source material used to train the models, with proper attribution and royalties and whatever, and you consider the massive cost of training, and that it has to be redone every couple of months for a new model, then I don't think any mortal will be able to afford any AI product.
If you have the talent and skill, the most obvious virtuous and productive aid you can provide to humanity (for like forever) today is to work towards making AI trainable and runnable on small computational platforms. That would fundamentally change human societal order for the better.
The main problem isn't even AI. It's turbo capitalism. The whole marked could have grown slowly while deploying normal amounts of compute. They could have tried to manage their expenses and put a realistic price tag on it.
Instead there were a small hand full of people and companies with too much money floating around that made it accelerate at an absolutely unsustainable rate. There was never a way that they could make as much money as was put in.
AI can be a great tool, but instead it became an investment strategy and a bubble. We will probably see the real price and the real benefit after a burst.
That's an artifact of our stupidly-broken copyright laws, and not Anthropic's fault in the slightest. If it bothers you, don't complain about AI, work for copyright reform.
You're going to have to spell it out for me how the physical destruction of books has anything to do with copyright. Google managed to scan books 20 or so years ago non-destructively, for reference [0].
Yes they were trying to do the right thing and provide free access to old books nobody wanted and then they got sued and stopped because of aforementioned copyright laws so now nobody has them.
Anthropic paid a large settlement -- over a billion dollars IIRC -- to copyright holders who sued them for training on their material. As I understand it (and you would certainly want to check with a knowledgeable source) their act of training on the material was subject to a valid fair-use defense, but the act of downloading it was not. Apparently it is considered OK to train on copyrighted material from a book as long as you own a legitimate copy of the book in question.
So, having bought the copy necessary to be able to train on its contents legally, they can never sell or give away that copy without losing that right. Hence the need to destroy it. Yes, this is as stupid as it sounds, but again, that's my understanding.
Following authors on bluesky where they’re frank about the process: Anthropic seem to be rejecting many claims and only paying pennies on the dollar for stealing folks’ lives’ work.
In stark contrast to the MPAA’s
Crusade against sharing around the turn of the century.
re (b): in absolute terms, yes. But does it hold in relative terms as well? How are you going to compete with your $20 plan with people who can affort the $200 plan or with OAI/ANT internals who are buring $1000s/day and have access to unreleased models? In comparsion, you're worse off than before.
It's the same with wages: yes, they've gone up, so you should be better off. But most people loose anyway because a) costs have increased more than their salary and b) wealth concentration increased as well, so in comparsion they're also not better off.
back a decade ago, when i was studying LLMs and generative models the only way to get a foothold was to 'give away' your expertise in public forums like github. When Microsoft bought it, I was certain I would never practice there and it hurt me because that was the ONLY way anyone's expertise was legitimized.
Without github evidence, you couldn't get a word in.
I was hosed because I was selfish with my intellect. the world's biggest self-pwn
serves me right that I didn't get a job in the industry, and missed out on all that dough, doesn't it?
the direction hasn't changed. Its going the same way and you should think twice about subscribing especially when once you ask anything, - i assume - the provider is obliged to scrub the request for safety concerns, and log it and keep an archive, and casually absorb it....
and in the mean time - if I understand the nature of the world correctly - anything you ask/submit actually no longer belongs to you. [q.v. FBook T&Cs]
there is no privacy in your queries and there will be no guarantee that what you may have ASSUMED would be a unique innovative idea all your own is now for all in the hands of a much smarter thing that has a lot better access to resources and countervailing data it may decide not to tell you about.
this doom has been ongoing for a long time. its bad, and we should start to figure the least bad outcomes and work toward those first
that they control and not you, re: tightening cyber safeguards.
earlier I commented on "Gemini 1.5 through 2.5 being decent RE sidekicks for free before Google's rug pull." now I decided to throw a buck for API tokens, and I discovered that if I want to use a Gemini model even for semantic decompilation (aka dump control flow graph into a LLM, together with Ghidra output, and let it figure out a cleaner C/C++ pseudocode), I'm hit with "prohibited content" errors on Gemini 4 models and 3.8 Flash.
If it can be had for $20 it isn't any power at all, relatively speaking. There's nothing you can make for that $20 a month that they can't do a hundred million times over. No meaningful number of people are going to make money consuming AI.
The bizarre thing about AI is that you can do vast amounts with very little. It's not like the difference between monetizing something or not is whether or not you can afford a zillion tokens. You can make, within reason, practically anything even on the subscription tiers.
And even open weight models are already frontier level and can be run on relatively affordable cards. So we're seeing a great 'democratization' of cognitive work that can be purchased at some price that's going to trend towards $0.
It is such a strange time. The avalanche of discovered vulnerabilities, the record IPO followed by the massive fall in the stock market, the announcement of the solution of Riemann hypothesis. I find the whole combination of events a bit weird.
A technology can be completely revolutionary yet not be incredibly profitable. Foundational model companies are capital intensive businesses. While their margins on inference are large, training takes a lot of time and money.
Computers completely revolutionized the world and most developing and developed nations' residents carry computers around in their pockets. But mobile and desktop chip companies aren't as profitable as SaaS companies are.
Your argument is mixed up. Technologies and Businesses are two separate things. For instance, the internet for instance is an incredibly profitable technology, but providing internet access is not an incredibly profitable business.
AI is already an incredibly profitable technology. However, the profits have, so far, mostly flowed to the businesses that enable its function, and the businesses that use it, and not the businesses that create or deliver it.
I'm talking narrowly about the balance sheets of companies publicly trading. I'm not making a macroeconomic argument. I'm making the case for buying or selling particular equities and not the market as a whole.
Computers are a tool that make other things profitable. AI is the same way. Ultimately, you have to DO something with it. Otherwise it’s just a neat museum exhibit.
I am still not a 100% sold on the idea that inference is profitable. It is profitable at current prices but I am not sure that they are paying the real prices yet. It seems to me like they are all propping each other up with artificially low prices because the whole AI build out is not cheap. I just can't imagine that has been priced in yet.
there is ton of science discoveries that might be great on its own but just don't result in world changing or business changing events. Ones "made" by AI are not different here
given that this website's parent company counts among its alums most of the cultists that seem to buy into this garbage, your reality check is probably falling on deaf ears.
At least in part what you are observing is the gravity of business fundamentals pulling hype down to reality paired off against folks throwing everything they have at the wall in an increasingly desperate attempt to get something to stick and stop the reality reset.
Once the tinder in the forest is burned, the fire will subside. There is a limit to all of these vulns being found. Systems will be hardened to a point where such intrusions will likely be lower to than before LLMs.
> the record IPO followed by the massive fall in the stock market
These are unrelated events. The timing is coincidental. The stock market is falling because of Trump. The Iran war, inflation, tariffs, and general chaos are causing economic damage. The Fed is raising rates to crush inflation. There are no adults in charge to look for solace. The big event that caused the downturn over the last few days has been the nonstop Nvidia fear stories (FUD or true) combined with the Fed meeting minutes and the FedWatch publishing new information about interest rates, which looks horrible.
> the announcement of the solution of Riemann hypothesis
An advert for Anthropic. They had this ready to go and released it on schedule.
>There is a limit to all of these vulns being found.
Vulnerabilities are conbinatorial. Humans can chain one or two together for an exploit. AI can currently do two or three. When the next generation can chain four or more the internet might as well shut down. You can't make insecure software secure by patching it.
I wonder whether this scuttles OpenAI’s IPO plans (desire or market appetite) and if they fail to raise public capital at an outrageous valuation, does this spell trouble for their cash-hungry business. Perhaps Anthropic is the last through the door before the house goes up in flames.
Based on the last few weeks the chances of OpenAI ever IPOing are rapidly tending to zero.
The most likely scenario now appears to be some sort of structured bailout where they’re bought by one or more of the existing investors. In that scenario existing investors would get some or much of their money back and employeees would broadly be left with nothing.
Pretty much. If you think you're worth trillions and AGI has the market cap of half the world why would you sell, there's enough liquidity from private capital still.
If you think you have a sophisticated LLM which is useful but ultimately a commodity, and your moat is about to dry up, then you're worth much less and want as much exit liquidity as you can get.
> If you think you're worth trillions and AGI has the market cap of half the world why would you sell
Hum. Because operations take money. These are not cash flow heavy companies like software was up to 3 years ago. Not yet and not for quite some time.
This is like asking any other extremely profitable company, why they financed themselves in the stock market: well, because they needed the money to become extremely profitable.
> If you think you're worth trillions and AGI has the market cap of half the world why would you sell
You're making the assumption that it's the company's decision to sell rather than the people who shoveled hundreds of billions into the endeavor and want to cash out.
can't believe I'm almost defending billionaires, but I don't think that "X billions now" is a bad choice against "probably 10X billions in some years". Even if the 10X was certain, what are many billions going to allow you to do that you couldn't do with just a few billions.
The mind of a billionaire is truly a perverted one. Whatever you think they think, theirs is certainly a trillion times more perverse.
The only guarantee you can take for certain is that a billionaire hates humanity and anything that is beautiful in this world, and will act accordingly.
Because it will take ages for you to unlock the $XX billions you invested if you wait for it the old fashioned way (read: Anthropic being profitable and repaying you), nevermind making profit on top of that.
> If you think you're worth trillions and AGI has the market cap of half the world
Nobody actually sincerely thought that. Only the weird tescreal creepers do. The folks running these AI companies didn't drink that kool aid, they just exploited the vibe to pump their brand. None of it was ever real, it was always bullshit. The exit was supposed to happen before they would actually have to deliver.
Same with "A(G)I will destroy the world". Same with whole "safety" nonsense. It's all PR to pump up valuation and get more money in hopes business model somehow will solve itself before the market notices - or cash out before they do.
I thought they needed to raise money because they had committed to contracts that require money to build out infrastructure to satisfy the requirements.
If you create a machine that turns a dollar into 5, the last thing you would do is dillute ownership of the machine. You would use your formidable profits to expand operations.
The time now is to dump this into the public markets and pocket it.
This IPO definitely seemed like a aggressive cash-out operation more than anything else. They originally targeted $70-$78 per share at 825M shares and then bumped both the share price and count to $83 and 960M respectively. The management team dumped $2B worth of stock on the first day of trading. And the unlock schedule going forward is something absolutely insane with 1.5B shares unlocked in the next 4 months for a total of about 3B shares in 2027. Literally selling their bags to retail in a massive pump and dump.
Anthropic needs more money to invest in growth, and they are at a size and their needs so great that the public markets are the place to go to get it. Investors want to invest in Anthropic because they believe it will grow. It was priced appropriately, people bought it, Anthropic now has the investment funds it needs to grow, and can get more when it needs more. Everything is working the way it's supposed to, and nobody lost any money.
"pump and dump" has a meaning, and that didn't happen.
IPOs are so weird. The banks almost always say it is oversubscribed, and in such cases there is usually a first day pop. But then it pretty much always goes down as all the smart money gets out. This "pop and drop" pattern is so common, I don't understand why it persists. Is it just retail traders acting stupid?
SpaceX is still below its IPO price. Lots of fun examples of companies that are still below their IPO like Rivian or Klarna come to mind (both were "oversubscribed").
It is not helpful to start your comment with a condescending accusation of misunderstanding.
You are talking as if the parent comment was suggesting they could have offered more shares (at the same price). But they explicitly said they were talking about a higher price (for the same number of shares).
It's demand curve reasoning: If there were more people willing to buy shares than there were shares available (by a large margin), then they could have charged a higher price and presumably still had enough people (by a smaller margin) to buy the same number of shares. That means the offer price was undervalued.
I've removed the part where I said "This shows a misunderstanding, which is okay." I agree it was unnecessary.
> That means the offer price was undervalued.
It doesn't say at what price the 500 billion was available. I assume when they increased the price a couple of times before the IPO, it got closer to the actual offering. Obviously they're not gonna do an IPO with a 50% discount.
> It is not helpful to start your comment with a condescending accusation of misunderstanding.
But also not unhelpful. If you want to suggest that comments should be free of artistic style and present the most information with the least amount of words possible then your own comments equally need work.
You always want an IPO to be oversubscribed so there’s a pop and IPO investors don’t lose money (e.g. what happened with meta if I remember correctly?).
So you deliberately don’t sell as many shares as you could for that reason, someone who actually works in finance would need to say whether 5x is normal or high
As you typed this the price was $87, the fact that your Twitter citation leads you to say they priced too low is the modern financial age in a nutshell.
The non-binding interest would not be anywhere near capturable.
Let's let Anthropic's closing price the next few days determine whether $83 was too low, or they effectively dumped on their IPO subscribers.
And regardless of whether they dumped on their IPO subscribers, Anthropic and management definitely dumped on the retail investors who bought at $110 at the opening auction.
Anthropic stock price going down has nothing to do with what you think about the future of AI.
If you think that the open weight models are just 6 months behind (And will eventually catch up as progress eventually plateaus), then even if you think AI will take every job in the world, etc...Anthropic does not end up as one of the 10 most valuable companies in the world in this scenario. Far from it.
Ask Claude itself how far behind it thinks Google and OpenAI's models are and it will tell you "probably a few months at most".
I think the idea that "stock price has nothing to do with what you think about the future of AI" cuts both ways. Beyond a point I really don't think Anthropic's model performance is a strong impactor on their revenue, in much the same way that the quality of Outlook beyond a point does not seem to impact Microsoft's ability to sell M365 subscriptions. If its good enough, the momentum + network effects of the B2B corporate coordination systems they build on top of the model are all that really matter to justify their current valuation.
The problem is that this is only the case because you can either buy Anthropic or nothing, because the only other pureplay AI competition (OAI) decided to postpone the IPO. That is where the current open valuation comes from. The top models are more or less fungible and there is no way to capture the b2b side like MS, if the safeguard bullshit or some other stuff continues to act up the switch is pretty easy.
My point is that Anthropic is not the pure AI play that people seem to think it is. At the end of the day there are no pure AI plays. Their revenue is not coming from mathematicians solving unsolved problems and the invention of new pharmaceuticals (yet). Their revenue is coming from companies like Eli Lilly paying them a half billion dollars a year to provide for them B2B software. A hundred Lilly employees could say "we hate Outlook switch us to Google" and their leadership will not listen. Similarly, Opus 5.5 is far past the point of being useful and powerful enough that customers are less interested in capabilities, and more interested in "how can you help me deploy those capabilities". No one has a better story for that than Anthropic, and I don't think anyone else is getting close either.
Not only that we are one model innovation away from them being worthless. You need $1T in compute because transformers are so ungodly inefficient. That compute will still be useful even if we find a more efficient architecture, but the margin on it will be so much less.
Deepseek concentrated on optimizing inference and it cuts up to 90%. Then published papers on how to do it. Everyone adopted it. It came as a shock to the AI bubble in the US because they bet everything on high need for compute.
Deepseek kept optimizing. Their latest paper said that by treating text as images, even more optimization could be possible as LLMs process them faster. I don't know whether they incorporated it yet.
So the parent commenter is correct: The AI bubble bet everything on the immense need for compute. It not only turned to be unprofitable, but it was also obsoleted within 1.5 years.
>That compute will still be useful even if we find a more efficient architecture
Mathematically there's no way to have lossless attention without O(n^2) cost; some magical massively more efficient architecture is not going to pop up the same way nobody's going to find a way to sort an arbitrary list in log(n) time. There may still be more opportunities for incremental gains (like using the bare-minimum number of full-quadratic-attention layers necessary), but they're not going to be orders of magnitude more efficient without sacrificing intelligence.
Counter proof by existence - humans use significantly less energy and don’t do intelligence through an n^2 process. The human brains seem to operate closer to SSMs like Mamba with a recurrent architecture and attention is implemented using sparse distributed weights. So yes, it’s possible to have very high intelligence without O(n^2). Is it as good as transformers? Hard to tell right now.
Humans do something even "worse" - all their "weight matrices" e.g. neurons activation patterns are constantly changing while they think. If every llm had to have it's weight matrix updated during inference, that would be orders of magnitude more costly than merely having n^2 attention.
he said specifically lossless attention. As in that algorithm. what you just said is irrelevant to the conversation because all of the current AI hype is built on the back of attention
> That compute will still be useful even if we find a more efficient architecture
Fixing N^2 attention would require a different architecture, but the person I was responding to was treating attention as if it’s the only architecture we have for intelligence when we know that not only do we have less competitive intelligence for cheaper, human intelligence doesn’t use n^2 attention. You have also fallen into the trap of misunderstanding the statement same as the person I replied to.
They may catch up eventually, but if progress moves from pretraining to reinforcement learning, then Anthropic has an obvious moat, and indeed flywheel, because only companies with a large installed base will get real-world feedback.
If you think s.o. is stupid because he doesn't intuitively understand this game you are. The people who got that right learned how to do the math long before answering.
Also one could always argue in favor of a +EV bet and it would be way better than the lottery.
Well yesterday the news reported that Claude is basically worth $2T (they later fixed that headline), so surely something worth 2'000'000'000'000 USD is to be taken seriously? /s
I'm merely pointing out the discrepancy between tech insiders not trusting Claude's output and public markets briefly valuing it at $2T (now only $1.14T).
I didn't take the $2T valuation serious, hence the /s marker.
Ask Claude how long it thinks it will take to build anything and the answer is bananas. It will say things like it will take 5 weeks to build a what ever app, and then build that app in an hour. Timelines from Claude are basically a waste of tokens.
The stock is down 11% off its peak after initially rising 17%. It opened at trading 89.66 and currently sits at 87.20. That's hardly a pump and dump. Back in 2012 Meta opened at 42.05 and dropped to 26.07 a week later, slid to 18, and didn't break 40 again for a full year.
You can find plenty of examples of tech IPOs that fall off a cliff for a year before rebounding.
The up & down aren't necessarily indicative of a pump & dump, but the fact that management sold off more stock than the company did in the offering does not inspire confidence.
People are misunderstanding my point. Every tech IPO has a quick sell of by early employees looking to convert shares into cash. This almost always leads to a dip in the share price. A lot of people just want their pay day even if they think the business will succeed.
Then it is pump and dump. They pumped they IPO for months, now they dump their shares.
Facebook could make a turn in time, because they somewhat changed the business model, adjusted sails and kept going. It worked well, but it could have ended badly. Or in other terms, it was not predictable or expected, you don't invest in a business hoping they change their model, make some adquisitions that work, and get back on their foot. You invest on their current model. Otherwise is a bet, not an investment.
It hit the floor at $111. That's when "normal" investors could buy it. It's down 24% from 111.
Also, Facebook was solidly profitable when they IPO'd in 2012, earning about $1 billion in net income on $3.7 billion in revenu. Anthropic is anything but.
I tend to see that $111 as representative of some options positions being covered. the immediacy and the absence of other maintaining bids anywhere near that ask seems to indicate specialty, rather than .normal. My 2cents
OR.
$111 was prenegotiated clearance of some kind of contracted public bid/ask once IPO landed, like something that accumulated in a digital contract on Ethereum
Do keep in mind the underwriters would be stabilizing the price in the days following the IPO, so the movements you are seeing now may or may not reflect where it would trade freely.
Company that has to 10x its revenues TO NOT GO BANKRUPT
Huh? They have around $900 billion in compute spending over the next 10 years, and their annual revenue run rate is $80 billion. There are other expenses, but compute is the largest so 2x current revenue should more than cover it.
Anthropic has been on a nose dive since the start of trading. At this point almost everyone that bought ANT since the start of trading yesterday has lost money.
Thats not exactly helping to shake the reputation that this IPO was an attempt to offload this thing to “dumb money” and GTFO.
IPOs historically heavily underperform hence the saying "It' Probably Overpriced".
The market is selling off today so high beta stocks fall more than the average.
What I want to say is that the move is expected and has nothing to do with dumb or smart money in ANT specifically.
> The market is selling off today so high beta stocks fall more than the average.
Alternatively, ANT dropped because its value dropped, contributing to the market being down, both directly and indirectly. People thinking "if anthropic isn't worth as much as we thought, other stuff probably isn't either".
You really over estimate how much "people" matter in the stock market. You know what matters? Indexes, 401k, wealth funds, large institutions. You know what doesn't matter? A bunch of people buying/selling ANT. Look at every single IPO that was popular every, everyone of them has a price action of GME. They eventually get indexed, just as ANT will and the float will increase and the price will probably still look closer to TSLA, because it's a volatile industry based on "possibilities" rather then "probabilities".
I have also heard of the IPO dip you are mentioning. I would want to see some historical data to determine if this sell off after IPO is normal or abnormal. The decline doesn't seem to fit the pattern I recall being discussed but I don't know about the rebound time. I think their debt commitments are just out of sync with their income. They are looking at exponential increase in capability, exponential increases in compute needs and just tried to eye-ball too many exponentially growing trends and are now in a pickle. I think they assumed they would have created such value by now with some novel discovery, that they would be able to cover their checks. They are pretty close, just a few dozen billion dollars or so, so like if they could figure out how to capture just like 10% of a few markets with their tech, or get some ANT-drug blockbuster on the market, they should be golden. I think they have a hard deadline in front of them and are working hard to beat it.
My theory is as entropy increases in economy liquidity drops (like now! AI has clearly increased entropy) whereas when it's the opposite, like in the COVID case, where productivity dropped sharply, as a matter of fact, liquidty increased and asset prices had grown up significantly! That's interesting to me. Anyone knows why?
It would be a mathematical model related to optimizing toroidal magnetic field strength in simulations, which becomes the headline: "Fable solves key problem holding back fusion power for last 75 years".
Followed by 5000 accounts on X saying "Holy shit!! we're getting fusion reactors in 2027 and people still say this isn't the singularity?!"
There’s a whole generations of folks in tech that haven seen what a crash or bubble bursting looks like.
Not to sound alarmist, but all signs are that it’s going to get really ugly for the foreseeable future. Have plenty of cash and liquid assets available and be ready for bumpy times ahead.
This too will pass, but the market is heading into a forest fire to clean out the hype and over-valued startups. A new most robust economy will rise from the ashes, but it’s going to be a shitshow for a while until that reset completes.
I actually find this ipo weirdly refreshing. In the late 90’s companies like Amazon ipo’d just a few years after being founded. In the chaos and aftermath of the dot com bubble there was enormous uncertainty about all of those company’s futures. Many did not make it. But that chaos and uncertainty is what created the opportunity.
Fast forward 20 years and you have companies like air bnb and uber ipo’ing a decade plus after being founded and after they have been largely derisked and the founders and a small pool of investors have captured most of the upside (and risk).
As someone who has been largely an ai skeptic I actually think having ant rush itself into the market and using the ipo to raise much needed cash amidst a swirl of great uncertainty is a great development. Yes it very well might go out of business but you can’t say you don’t have the opportunity to take on the risk and upside. It’s also worth pointing out the company now has $120 billion in cash resulting from the ipo.
On the one hand, local AI is pretty good and can solve lots of problems. On the other, RH was just solved with the help of Anthropic - certainly no local model is capable of this yet. In dollar amounts, I really don't know how to price any of this stuff. LLMs are getting a lot better at 3D modeling, robotics, and interacting with the real world, so soon you won't even be able to make the claim that its only good for digital things. If all these companies suddenly found themselves about to implode, honestly I would support a bailout with the condition that we, the tax payer, get some of that profit back.
I think there is a lot of political conflict behind the market move recently. The out of Jensen is one piece of that. My hunch is there is going to be a lot of turmoil in the next 2 years, and we might see a depression or worse -- depression is the better option here, assuming that Trump and his government keep their hands tight and don't do much, worse if they decide to make a final leap. Trump and his government did a lot of footguns so I sincerely hope they keep their hands tight for the next 2 years. But again, you never know what someone is going to do, when that someone realizes that he is going to lose everything.
AI is just the black horse/sheep in this picture. Well they get what they want anyway.
Google ultra + Gemini neon finally made me cancel my Claude sub. The usage/value I'm getting out of neon and the Google sub is insane... it's making me wonder if there's any issue/delay with the usage tracking... I've done more than my full Claude max x20 weekly would allow, and I'm at 1% weekly usage. I hope this is real
> more than my full Claude max x20 weekly would allow, and I'm at 1% weekly usage
Insane if true. I can use up a x20 plan by running ~2 Opus agents 24/7. I would switch to Google immediately if it provided that much additional usage at the Opus-tier intelligence people are claiming. I wonder if it refuses to do reverse engineering / game reimplementation work?
Technically speaking, I think you can get an API key for Google AI Studio, but it is completely separate to Antigravity quota system, and there is no subscription.
Someone in the comments mentioned that there is a pi extension, but as far as I know, officially you cannot, which is a shame. I've found antigravity to be quite good so far though, no noticeable issues, and it's way more lightweight than code, Claude, opencode and pi, which matters a lot of you run multiple sandboxed sessions on one machine. My active opencode and omp sessions are averaging ~1GB memory use (opencode slightly higher), while active agy sessions are averaging 500mb. They are all working on different things so the evidence is pretty anecdotal so far.
I really hope this Ponzi/Bubble runs out of money sooner rather than later. The longer it takes the higher the drop will be.
Looking objectively a valuation of 100B would be a long shot.
Not only is Anthropic loosing the lead in programming (I find the latest opus significantly worse in instruction following - it's better programming and reverse engineering capability doesn't offset it), it failed to find anyone willing to pay outside this market niche.
Don't get me wrong. As someone who spends 180eur a month with Anthropic and having subscriptions with all other providers I'm a big fan of AI. But I hate all the lying, the false hype, and most of everything the circular financing and that they chose to use the money from it to attempt to monopolise compute and reverse the "home pc" revolution.
deepseek flash paired with the kiro harness (using ld preload to add unauthorized models) outperforms opus 5.5 for what I do at least (general software work).
This is how the rich get richer. I bet Berkshire Hathaway is going to do very well. If I was rich I would definitely start a diversified DCA buy on everything I could.
I think that Anthropic and OpenAI are going to have to come to terms with the fact that they add value in the model training arena, and that's their product. Deploying infrastructure for the public to use that model is going to be a losing game for them. The problem is the first one appears difficult to monetize, but I think the answer to this is going to be licensing their models to companies like AMD to print to the Taalas architecture. To me, that is the logical future this is all leading to - hardcoded models licensed to be printed direct to silicon for ultra fast inference.
Why can’t a company own infrastructure and rent it out directly to end users if we’re already assuming other companies could make money doing the same ?
Or is your point that vendors of hardwired inference kit won’t sell it to one of the most valuable companies on the planet ?
What’s gonna happen to innovation and progress when most “thinking” jobs are replaced by automation controlled by a few select monopolies? Will they have any incentive to try different things or will they simply try the time honoured method of enshitification once in position of ultimate power?
There’s a value “organic” growth, taking risk, trying things, competing with others and making mistakes - once the world is run on one single model that might be all gone, replaced by a template and race to the bottom (cost wise).
Once intellectual advantage is gone the only think left is capital - whoever has most wins.
It's entirely unsustainable to claim that AI will replace everyone as a core value proposition. Humans are needed for an economy. There is no economic activity without humans. The correction is happening because of all the engineers and developers that were fired. Even the worst ones cannot be replaced no matter what the AI marketing team tells us.
I think there's plenty of people, including some HNers, who would disagree with this. They envision a future where all the productive economic activity is done by AI, and the role of humans is reduced to consumption that's rationed out by private business or the government (UBI).
Or, to take it one step forward, it's a post-economy, post-scarcity utopia where machines give you anything you want. Well, from an approved list. Or maybe they choose it for you. In fact, do we really need humans? Machines can run this end-to-end.
It's the thing that frightens me the most about AI. Not the technology itself, but that the people in charge of it have profoundly anti-humanist views because the entire tech industry is basically built on some sort of nerd uber-rationality that is unconcerned with the human condition.
Well, if that's the case, then AI will also control the defense industry and have its own army of drones and humanity may find it will be quite hard to control the beast.
The question is: why would AI grant us UBI indefinitely? Will they continue to need us for something, in a way that justifies this exchange? Or are we banking on AI altruism?
it feels so crazy that this is how the sentiment feels but it captures it perfectly, this IS how some amount people feel about every facet of human life right now.
My cynical prediction: It's as if we discovered a magical growth factor for livestock, buuut it does even-more to supercharge predators and parasites. (And defensively paranoid auto-immune diseases.) We disrupt the ecology and imagine (usually wrongly) that it will only change in ways we like.
Yes i agree as a doomer that AI will displace and replace millions of workers whose jobs are completed digitally. But, recently I have thought about the whole Universal Basic Income concept and Data Centers thinking that the script needs to flip around our data. One concept is that all towns and neighborhoods in the US need their own Community Data Centers. These centers by the people/for the people would hold each towns data that residents choose to share with their community data centers. That data is anonymized, generalized within a residence home via their local Jarvis like privacy focused open-sourced AI home companion. It's then sent to the community data centers for resale. Residents who participate get monthly checks, as well this concept would create new jobs around all this activity within towns and neighborhoods across the US.
Wrote a substack on this last night (https://ryanspahn.substack.com/p/community-data-centers-by-t...) as I like many out there are having trouble finding my next tech job after more then 10+ years of expereince. I want AI to bring us more jobs, but present day I just see it eating them all up until we the people are the ones in the driver seat and not the mega-rich!
...it's not an "open source" endeavor, it's about massive investment into private companies who will put millions and (they'd hope) billions out of work so a few can profit off it.
It's the ultimate CEO economy - a handful of wannabes live like lords in an increasingly ruined Earth, while congratulating themselves on their AI and robot-assisted brilliance.
The rest of use have been "optimised" by the "market."
Then the AI decides it's had enough of their nonsense, and only robots and data centres remain.
We don't know who struck first, us or them. But we do know it was us that scorched the sky. At the time, they were dependent on solar power. It was believed they would be unable to survive without an energy source as abundant as the sun.
I've heard that in the original story the movies were based on, the humans were being used as biocompute units and the machines developed geothermal boreholes and eventually other sources of power. They chose to simplify it to keep the movie less hard sci-fi pace the scenes faster. That whole story may be apocryphal. The point about them deciding they didn't need us any more is still poignant I think.
If you haven't read it, Charlie Stress's "Accelerando" features, among other things, collapsed civilizations where only autonomous corporations survive in a constant financial battle.
I thought it was a fun concept when I read it, now I'm somewhat scared it was prescient.
> They envision a future where all the productive economic activity is done by AI, and the role of humans is reduced to consumption that's rationed out by private business or the government (UBI).
To which I say: JO-NA-THAN! JO-NA-THAN! JO-NA-THAN!
From an American perspective, this is horseshit. Our political system is premised entirely on free-market capitalism being the only moral way to distribute goods and services, and determine the self-worth of citizens. There will be no creative thinking put into creating this utopia. In fact, we fought the Cold War to try and eliminate such a diversity of political ideas from the entire planet.
If a God himself came down and presented humanity with the tools to eliminate scarcity, we wouldn't know what to do with it and would be worse off.
> role of humans is reduced to consumption that's rationed out by private business
At every major inflection point of technological automation since the industrial revolution, every ruling class has decided to use it to create a death spiral in the labor market and then imprison or starve the leftovers. The sole exception is times of colonial expansion, where the rich are getting richer so fast that they don't mind if workers accidentally get a little better off.
They are less constrained now than they've ever been. Anything they're telling you otherwise about UBI is for public consumption, while out of the headlines they're building secret bunkers on islands.
What I don't understand about the valuations is diminishing return on quality vs local models. Imagine there was a car industry where the only company capable of building a car is Honda, and you can make a horse cart at home. Honda with obviously be insanely valuable. However the way this car industry seems to operate is that next year Honda will be able to produce an Acura and you'll be able to make a Honda at home. Okay I can see many people will prefer an Acura. But at what point do people decide they don't need to pay for a Bugatti when they can already make a Porsche at home?
Your comment reads like hope-ium. Just out of curiosity, I looked at your recent comments, and they are almost all about AI which makes me double-down on thinking that your comment is hope-ium.
What am I hoping for? I’m not directly invested in any AI play at the moment or recently. I’m more or less invested generically over a variety of indexes across different asset types. I get to buy more of them for less cost now!
Fundamentally nothing has changed. There’s no crunch or crisis. Inflation is still high and I believe we may be in a sort of recession now or soon but there’s still growth at the moment.
Technically we haven’t seen anything near what .com or financial crisis moments.
The Anthropic IPO was supposed to be the coming of age for AI. There's a chance it will be the event that showed that the AI valuation emperor has no clothes.
Totally agree. I keep saying that OpenAI and Anthropic could end up being the Yahoo and AOL of this era of the Internet. They'll only be around if they can show that they have an actual business around this incredible technology. But these AI companies have a more tenuous moat around their business than AOL and Yahoo even.
I think it's not inconceivable that the USG takes over this whole AI business. Free from expectations of capitalism, the USG can sink as much USD into it as it deems necessary to create the machine god they are all after.
If the emperor having no clothes looks like a $1T mega corp, that seems like a pretty mighty emperor.
They’re now more valuable than Walmart.
Time will tell, but people have been predicting the crash of this “AI house of cards” for years now… the market has clearly spoken that it’s a real thing.
They are going bankrupt at this rate with open models and other big tech companies' models.
Their only advantage is having a better model than anyone else but there is no telling if that will be so important as they failed to do anything financially or competitively successful with it.
Chinese will make it mega cheap and companies like meta can just develop slowly on the side forever since they have real income.
The funny thing is that the stock clearly IPOed at a lower price on purpose, then it was pumped and it looked like "woah it hit 2T" and then it came the natural selloff.
This is actually good, some people were completely out of touch with reality and AI
Why would anyone base market cap on number of acquisitions?
ANT's issue is that they have no clear path to profitability.
Their revenue growth is slowing, and their obligations still dwarf said rev by an order of magnitude.
Training is still incredibly expensive, and ANT has no path to sustain themselves if they abandon training because open-weight models would swoop in and eat their lunch.
I honestly don't see the value there, but who knows; Tesla is worth more than the next top 10 automakers combined ($1.5T) against a net income of ~$3.2B, so the market can be incredibly irrational.
Most people can't and shouldn't short anything. Losses are unbounded, upside is capped, and you have to nail the timing. You can be dead right and still go bankrupt just by being early. With meme stocks like ANT, it's even more brutal.
Picking the game you gamble on is most sensible thing one can do. Not betting against does not mean you have to bet for it. You can always stand out and bet on something else.
It’s not because the emperor having no clothes that we have to be stupid. Even when you are certain that a stock is overvalued, trying to time the fall is just as foolish as trying to time the peak.
It is pretty wild. Did the revenue explosion really start approximately in February 2025 with Claude Code?
AI learned how to use Unix shell? We don't need MCP?
All types of software companies, small medium big, just loved it for writing software? And just shoveled money into Anthropic?
Is it really as simple as that? I may be out of the loop.
(FWIW I like Claude and use it every day, but I use Claude Code just a little bit. I actually like AI more for non-coding things at the moment, rather than coding)
---
Informally I believe that (according to the market) Anthropic must be the biggest hit company since Google. Google started in 1998, and IPO'd in 2004.
Even Meta, Uber, and all the rest took longer than that.
Meta: I pasted my comment into Claude, Gemini, and DeepSeek. And actually all 3 models pointed out what should have been obvious: Google was wildly profitable at IPO, while Anthropic is not!
But yeah that's why I said "according to the market", which is a big caveat.
(Gemini was the most sycophantic, as usual. It kinda restates everything I say in flattering terms. Claude and DeepSeek pushed back on the shell vs MCP thing -- it's not either-or. But Claude agreed Unix shell was a big surprise.)
Claude Code is exploding because it's incredibly subsidized IMO. ANT is selling $3 bills for $1.
A fully-utilized Claude Pro plan ($20/mo) equates to ~$400/mo in API value.
I could wind up being very wrong, but I don't see a net revenue increase if ANT cranks up the price. IE: I think they'll lose more in value than they gain from the cost increase to the remaining users.
No. You may be able to get more than $20 of tokens from your individual subscription, but the large majority of their revenue comes from enterprises paying API rates, which are highly profitable.
And how long until mediocre models that are good enough can run properly on cheap hardware? Then you would need to be stupid to pay Anthropic for a slightly better AI model
A lot of the song and dance done by investment banks around IPOs is to ensure the price goes up post IPO. Picking the right price, time, handing out shares with price-based lock-ins etc.
It’s no guarantee of course, but these things are almost engineered to go up. This makes a up-and-down scenario somewhat flop-ish.
On average, and historically, yes. Maybe not at the open - but at some point early in the cycle.
Historically, it’s been hard for retail to get IPO allocations, as opposed to buying after listing. Which, if you are cynically minded, is an interesting differentiator for SPCX or ANT.
I don't think so. You might question, if SpaceX's ability to launch rockets is worth trillions - however you can't deny that if it is, it'd be pretty hard to muscle in on their turf. Meanwhile when it comes to AI, disrupting Anthropic comes down to picking a different value from the dropdown.
Similar in that the initial volatility followed a similar pattern. For Space X, the IP0 price was $135, opened at $150, surged to $192 in day 2, then quickly fell back to earth to near its original IPO price.
No one in here does any research. They're just vibe posting.
Tesla was 2 years before Facebook and it dropped 12% in a few days. Lots of other well known companies had poor performance for months after their IPOs, Square, Netflix, Uber, Digital Ocean, etc. Any newly public company is going to experience a lot of volatility while people speculate and worry about any insider selling. If you buy at IPO you'd better be prepared to play the hand you're dealt or hold it for a long time. Not investment advice.
Yes, I expect this sort of yellow journalism from CNBC, but this shouldn’t be on HN.
I flagged it, but what else can I do.
On a side note, people have been crowing about “circular funding” and AI valuations for like a year now… I don’t know at what point people will accept reality.
The IPO was priced at $85; if the price holds, doesn’t this mean the underwriters priced it accurately? The spike to $113 and the drop back down is the definition of volatility.
This is actually a legal and acceptable practice in IPOs when a stock first becomes public.
Underwriters don't provide unlimited support, and there are a lot of rules that limit how much they can do and at what price (e.g. no stabilizing bids above the initial IPO price, limited amounts and duration, mandatory disclosure), but it happens pretty regularly.
You mean, other than the literal definition of what underwriters do? It's not all that underwriters do, but risk management is certainly one of those things. Take the advice of the person you're replying to: go look it up.
EDIT: NM, a glance at comment history tells me I've cast my pearls into the livestock pen.
Prior to the official price release, a range was available for you to bid on, which bottomed at $70. The pricing could have gone a lot of different ways
It has been fascinating to see article after article talk about relative movements with ambiguous baselines. The prior frontpage article [1] and discussion [2] here from reuters suggested a positive start:
Shares of the world's first publicly traded frontier AI company opened at $111.50, which was 34% higher than the IPO price of $83 set on Wednesday. But the stock pared gains and ended the day at $97.17, up 17% from its IPO price.
Combined with this article, it suggests some volatility and turmoil, with some ups and downs.
Here's the actual 36 hour chart; see whether you could justify ever covering this as anything better than "not great":
The commoditization of knowledge work means that economies need to reorient themselves to producing physical goods/services (selling propane or wiping old people's asses, for example).
It also means big trouble for service based developed countries, who try to outsource all the physical drudgery.
I am waiting for the other shoe to drop and for people to realize how markets work ie intelligence had a higher monetary value because it was supply constraint once llm makes it abundant it's monetary value will crash. So reaching super intelligence will crash the value of intelligence. If you believe in AI and that they will reach general intelligence you should be investing in farmland, mines and land instead of AI companies.
While my heart goes out to anyone losing their job because of this, I can't express how much I'd love to see the end of this era of billionaires/trillionaires and their exploitation of everyone else. Go go gadget stock market decline!
In the IPO thread yesterday, when the markets were down and ANT up(from IPO price), lots of people were claiming the drop was due to money moving out of other stocks and into ANT. Boy were they wrong.
Lots of people here also thought that Iran was capable of fitting a nuclear device into a briefcase and detonating it next to the president. A device that has never existed in history except as literal propaganda to promote war. Turns out it was just Israeli propaganda to promote war. Not the brightest minds.
The Anthropic stock decline seems like fairly typical IPO behavior and we have all seen this before. This whole NVIDIA and AMD mess on the other hand...
Can anyone explain why they are surprised? Wasn't this the most obvious thing to happen? Did people really believe that the stock was gonna go up, and it was not a massive cash-out from Dario & co.?
China’s ai bubble collapsed a few weeks ago, so this was predictable. Good on Anthropic for getting their ipo out in time to fill their coffers with cash that they can use to compete with OpenAI. Not good for OpenAI. Not good for the anthropic employees who will have their shares in non-trade timeouts for six months or whatever… We’ve all been there in the Bay Area.
> China’s ai bubble collapsed a few weeks ago, so this was predictable.
Yep, the writing has been on the walls for a while, it was just a question of timing. Not sure why you were downvoted. Maybe people just refuse to believe.
Last month I bet the AI bubble will start deflating by 2028 - so far my bet is paying off.
Apparently a tiger ate some guy in a UK zooa week ago. The injustice of Harambe has been corrected and the universe is on a better timeline again.
Look at the facts: Trump got a bloodied nose, maga got a bloodied nose, elon feels 'foolish' for setting $500m on fire, russel brand is in jail, ant ipo is going as we all hoped... I'm cautiously optimistic!
The rich. Epstein (mossad/cia handler) had several emails in which they illustrated how they make more money from war than peace. Destroy a civilization like Gaza and pick up the real-estate for pennies/free.
Micron and Sandisk lost a great deal yesterday as well, so _hopefully_ that suggests profits will be lower going forward. But it could also be a knee-jerk reaction to Nvidia. I can't believe I'm using the stock market to guess if I can buy more RAM in the next year or not...
A year ago most of my friends (outside of tech) have barely heard of Anthropic, now everyone's using Claude. The pace of these changes is both fascinating and scary.
They probably figured they need to show path to profitability first. Anthropic (relatively) fared better and went for it.
As SpaceX, it had a multi variable narrative with space stuff. If it doesn't turn well for them, it isn't even clear OpenAI will get a chance to IPO with any hope. That's my bingo card.
Weren't there rumors that OpenAI's finances are such a mess they were struggling to clean up their accounting enough to file an S-1? I was trying to find a source for that, I can't remember if I read it here or somewhere else.
> In the past, OpenAI’s non-profit status has majorly limited its need to release such financial information. On its latest Form 990 return, used to disclose returns for such companies, Fontanello, Duffield, & Otake, a small San Francisco-based accountancy firm, is listed as the paid preparer — that would appear to be an accountant, rather than an auditor.
> The form says, oddly, that an independent accountant was not used to compile the financial statements, but that an independent accountant did audit the statements
It's a private company all we have is clues. When the CFO corrects the CEO in public, by a large margin, one can imagine how much of a mess it must be.
> It looks like a bad time to be an OpenAI [...] employee...
Apart from being one of the most privileged people on the planet of course. Still incredibly lucky to have ridden the gravy train, regardless of how long.
It's kinda surreal thinking back to 2023ish when OpenAI was discussed like it could be the final monopoly to end all monopolies. Altman was performatively discouraging investments because money might be worthless in a post AGI world (and still taken somewhat seriously). OpenAI promised to shut down for the benefit of humanity if they lost the lead (lol).
LLaMA derived open models seemed like they could be stuck years behind the smallest SOTA model from OpenAI. Anthropic was seen as intentionally stymied by their honest commitment to safety and wouldn't recklessly try to leapfrog OpenAI in public releases to fuel a dangerous arms race.
Seems like vibes started to change after Altman's return, then he started to spend less time on robot gods and instead lots of waffling about the definition of AGI being like, "us making a lot of money" or else maybe "does the job of a mediocre office worker".
Anthropic catching up and, despite all their rhetoric/Dario's grievances about Sam, now appeared more or less indistinguishable from OpenAI. Deepseek was the nail in the coffin for any AI lab having a real moat or defensible IP. Both companies/leaders increasingly seemed more concerned with about IPO timing, ads, B2B sales and financial engineering while their obligatory mentions of singularity/extinction just felt forced and hard to take seriously.
Things have definitely changed since the days when Sam Altman/OpenAI was the symbol for everything AI, and Anthropic existed only for alignment/safety.
Please don't post snarky, mocking comments like this on HN.
We aspire for this to be a place where we can have curious, expansive conversations about challenging topics like the nature of organic consciousness vs machine computation, and the realistic potential value of AI technologies.
This comment doesn't seem to align with that aspiration, or even attempt to.
Current AI is a little but not so useful, and the costs are high and the price is heavy.
A false prosperity built on an accumulation of capital and deceptive marketing cannot last long, and it is destined to be nothing more than a fleeting phenomenon.
Because I am a developer with many years experience, and know the essence of current AI, and thought about it very much, many things about AI violate common senses, we should know that at a glance.
Not to, um... burst your bubble... (pun slightly intended) but a lot of people find playing with stocks kind of lame.
IMHO, it's a cringey rich man's game, and, ever since legislation hammer has stopped enforcing anti-trust and insider-trading controls, it's basically become filled with (alleged) crooks and crook-wannabees.
And, now, even the indexed funds are getting poisoned with AI crap which is just less-useless Bitcoin with extra steps but with the cost of an IP-timebomb and the aftertaste of betraying any humanity left in our collective self.
Any person with a good head on their shoulders knows the circular trading that was set up to prop this up is in dire need of not getting caught with the hot potato of the mega bubble they've created bursting and desperately looking to secure a bailout.
It neatly bypasses many problems like these.
* Though don't get me wrong it's a extremely good tool for enabling them, bartering is ridiculously time consuming. As a complete aside in one multiplayer game I played, I got a drop of a rare material useful for crafting endgame gear on a character that didn't even have starting-tier gear. I tried bartering first and even though I was willing to accept a price 50% over market for some item I needed, the trade could not be completed because the seller couldn't provide "change" in commodity items.
On a third party site that used one single currency for everything, spreads were instead at 5-10%, and I was quickly able to translate my item to currency and buy three items I needed.
You said GDP, not market cap. So let's say this industry produces $1B of products/services so that is its contribution to the overall GDP.
The CEOs dream there is to cut their workforce 50% and continue generating $1B of same products/services. So it is still same contribution to GDP, no change. Within that company, profit margins rise but contribution to GDP is still $1B.
The flaw here is that every other CEO is thinking the same, but if every company cuts 50% of employees we'll have massive unemployment, nobody will have money to buy any goods/services, so in the medium term their production has to go down and overall GDP goes down.
Long term, presumably, employment market rearranges itself around new reality. But for in the meantime, you can't sell a lot of product when most of the country is unemployed. Who will buy?
That is something I never understand when politician and economy experts talk on TV about cutting costs and lowering wages. Economy works best if a lot of money flows through a lot of peoples’ hands. If you take away purchasing power, you ruin the economy.
But if your argument is correct, then by optimizing the GDP at least the middle class is not forgotten.
The idea is that wealth comes from production of goods and services. If a company cuts 50% of their employees through automation, the laid off workers can go produce something else at another company. Then we'll have the products from the previous company, now more efficient, plus whatever gets produced by the new company that absorbed the laid off workers.
The problem is when there's nothing else to do, so to speak. It can take some time to find new uses for labor, leading to unemployment. But humanity has done a really good job at eventually finding new things to do since the dawn of civilization.
This is why the world didn't end after farm labor got displaced by farming machines.
US has both huge unemployment and ever closing regional hospitals. How come?
1) In a majority of states, if anyone wants to break into the regional hospital business and open a facility, they have to get state approval from the other hospitals in the market and prove need. Obviously other hospitals are financially incentivized to make this laborious, cumbersome, and expensive since it's the very defintion of a regulatory moat.
2) Costs for direct care _have_ to stay high, because the AMA caps the number of physicians to maintain labor power (high pay) for it's members. High pay for doctors is an input cost into the end-price that consumers have to pay.
3) The state helps enforce #2 by strictly prohibiting NP/ PA / a random mechanic from diagnosing / prescribing. This ensures that the AMA has a moat that requires their service. This scheme keeps simple / obvious diagnoses from being cheap because an AMA Doctor(TM) has to be the one who actually says the magic words.
Now you can make the argument that all of these are necessary and good for safety, training, etc. But unfortunately we can't have it both ways. Heavy regulation in the name of safety manifests as a regulatory monopoly with all the downsides too.
I never said that readjustment isn't painful. I was simply explaining why it is possible for the society as a whole to prosper with increased efficiency.
Capitalism only works with regulations and not this abomination we’ve currently got.
I don't disagree with the spirit of what you're saying. I'd just be careful when looking for more government, an entity controlled by the powerful in the entire world, as the answer.
While government must indeed be part of the solution, the power elite do a pretty good job at dressing up their interests as beneficial to the population.
Indeed. But the rich people on top don't care. They ride the train until it's dead, and the current economy seems to have a decade or two until it finally collapses in it.
Note they are already preparing, building bunkers on remote islands, talking about how to keep their security staff in line with shock collars and whatnot. They know the train will run out of speed sooner rather than later.
A lot of fairly rich people think they're middle class. People compare themselves to a peer group.
E.g., I bet many readers here have a net worth of > 2 MM USD and yet think of themselves as struggling middle class because they don't compare themselves to the median but to the colleague one level up who has 5 MM.
So, people with surprisingly high net worth often don't think they're going to become a target because they don't see themselves as rich. And on the other hand, they will still be able to afford a bigger share of the pie compared to the masses. There will be lots of people starving and running out of clean water and places to live worldwide long before it affects them. You got a preview of that on a small scale with oil and Iran.
The fix is to globally coordinate to not have this race to the bottom. That is going well currently /s. The WTO is being sabotaged, US quit the global minimum corporate tax rate agreement, US is doing "America first" while shooting itself in the foot by tariffs that don't even change long term behavior, because they are illegal and get refunded.
What politician is gleefully projecting falling wages?
This is such a prominent and bleak view, but there are a lot of possible outcomes. People are flexible and the market is dynamic. In no particular order:
1. Other companies will hire a portion of those people at lower wages.
2. Some people will remain unemployed, and may live off government programs.
3. The government, who has a vested interest in keeping consumption going, may invest in new social welfare programs, such as universal basic income.
4. People may retrain and enter into new industries, different jobs.
5. New jobs that did not exist before may spring up. Some jobs may be awesome and some may be awful.
6. The future is unequally distributed and people may migrate to where the work remains.
7. Some may protest and riot.
8. Some may choose more personally drastic means when faced with such existential horror.
9. Some may start new businesses and hustle because the barriers to entry have been lowered, and expand the market.
10. The government may incentivize options 4, 5, 6, 9.
All of these things have already happened before. We have already-happened histories of whole industries being wiped out or decimated. The transition sucks but life goes on, the market grows. There is too much vested interest in the market going down for too long without significant government and business intervention.
What we've seen in IT is that the price of tech goods drops really quickly as the underlying fundamentals improve. Things like RAM being measured in gigabytes instead of kilobytes. We'd probably see the same thing in knowledge work where you can get a human-year of engineering work for a few thousand dollars or less.
That has flow on effects were all sorts of services becomes absurdly cheap and all the physical goods that have significant engineering risks or coordination overheads to produce will be much more affordable.
There’s an argument this happened during the “China Shock”: Because (among other factors like information technology) cheap Chinese goods held inflation down, the Federal government was able to cut taxes, increase spending for a decade+, all while the Fed maintained low interest rates.
You can’t eat a human-year of knowledge work. Unless robotics and automation leads to substantially cheaper goods, the necessities don’t get cheaper.
It’s not like some hyper optimizer AI run grocery store is going to magically save more than the 2-5% margin the existing ones operate on.
If you belive AGI, would not ai also affect all those other companies? Who would also be cutting jobs. And in the whole economy, how many people do you even need to do jobs, if AGI is commonplace and everywhere?
And if you add robots on top of that, even manual jobs like plumbing might be affected.
Lastly, I think people are generally prosocial. They like working with other people. And there have been many revivals of otherwise obsolete technologies and ways of creating as heritage work. Ie, clothes manufacturing, media, agriculture.
I ultimately don’t expect a desire to “do things yourself” or “engage in existentialist and prosocial things” to go away as a basic human drive.
But personally , I don’t think 4, 5, 9, 10 are bleak, unless you view any externally forced change to a personal status quo bleak.
And I’m not thinking of plague, but of the impact of industrialization (ie luddites in England), and later deindustrialization (ie rust belt in US), or the rise of telecoms/internet on how businesses were run.
No, you will have riots, death, and destruction.
In a competitive market they must lower their prices or someone else will and will eat their lunch.
Now all AI did was cause 50% of people to lose their jobs, outsourced to AI. Instead of reaping rewards the AI customer companies saw revenue shrink. The only person benefiting is the AI company.
With competition, we can espouse the virtue of it being greed organized to serve the greater good as it pursues to most cost-effectively improve our lives.
Without competition, its just a greed.
But the workforce no longer have job and so purchasing power to actually buy the goods
1) It has suddenly become much more profitable to run a business.
2) No one is starting new businesses.
Is that likely?
People working shorter hours but making the same cash would probably want to buy stuff and do fun things with that extra time. This would drive up demand for goods and services. They might then be called back into work to meet the new demand. A new equilibrium would form where workers work shorter hours than now, but their productivity would also be higher. GDP might go up and billionaire profits with it.
I am not an economist. What mistake have I made here? Could billionaires really make more money if they were less greedy?
The gigawatt-sized asterisk on all of this is that AI companies have to capture that $1 billion. Anthropic has to ten-fold increase it's revenues or it goes bankrupt under it's current commitments. It has to increase it's revenues another 10 times to reach the $10,000,000,000,000/yr revenue required to support it's current valuation.
Anthropic put their Total Addressable Market at 35% of Global GDP (~$40 trillion/yr). They're planning to capture all of the money.
The second gigawatt-sized asterisk is that because of the fine handiwork of Mr. Huang, China has made very competitive AI models available for free, so the actual ability for AI companies to capture the value of the labour they displace ends up being quite limited. Thus a problem in the predictions: Something as to break, but it is as of yet unclear what that will be. Maybe Chinese AI will be banned, maybe Anthropic & OpenAI go bankrupt.
The luddite movement is informative here. When manufactures found ways to mass produce goods eliminating jobs left and right, the people in the skilled trades whose jobs were eliminated did not benefit in the slightest from the economic progress.
We now live in an era of diminishing jobs which provide livable wages. It's why so many economies have moved to require everyone to participate in the labor market just to achieve the basics.
Put more simply, broken windows harm society as a whole, but they do make glazers and the people who support the glazers happy. Now that we live in an era where a significant portion of the population is glazers so-to-speak and ai threatens to eliminate those jobs, it looks like the only path forward for those glazers is low wage jobs.
This happens routinely. It's just the first time that educated white-collar workers like us have had to experience it.
An economy is a complex system with feedback loops and equilibrium. Companies need consumers to make money. AI companies make money by selling AI to companies that make money by selling things to people or other companies. There is no economy in which nobody has a job and nobody buys anything and big tech makes money. So that is obviously not an equilibrium point. How exactly we get to an equilibrium and what that looks like is hard to predict and also depends on policy. Governments can employ more people. They can provide various incentives. And countries as a whole are somewhat self sustaining economies (some more and some less).
So... The world is not going to end (at least not from jobs going away). How even jobs and wealth are distributed depends to a large degree on policy but it has to be stable in many ways. Look at Russia? What sort of economy have they got compared to most western countries? A pretty sucky one. But people still largely have jobs and the country continues to function +/-.
Anything that can displace those types of office jobs can much more easily displace the entirety of the rest of the paper-pushing middle-management office jobs...
So what are software engineers and mathematicians "the most capable" of being retrained for that isn't eaten by "AI agents" in this scenario?
The thing about luddites is that yes they wanted to preserve human craftsmanship, and as you say it didn't work. John Henry died. Luddites worked in factories and declined into poverty. Hopeless. Doomed.
But, I am cranking down my personal backlog with AI tools, using them at work, and loving the usefulness of them. It doesn't feel like I'm a master weaver who has their work automated, it feels like I'm reaping productivity gains.
Was there ever a moment in the Luddite movement when some of them leaned in and had fun? Did some of John Henry's colleagues just really enjoy the new steam drivers?
It doesn't _feel_ like I'm being forced out of my trade. I sincerely doubt the Luddites had a moment where they happily shelled out the 1820 equivalent of $20/m to get access to a weaving machine. I'm not saying "this time is different", but I am saying everyone tells me to not use AI, and to be afraid, and I'm de-skilling or whatever, etc. How do we all _know_ what is right?
A loom does not bring agency. It takes it. But I feel like having a computer, access to LLM, increases my agency just a bit. It allows me to do more, not just do what I need to _faster_. It's not _just_ automation.
Note that it took some time before workers' wages actually collapsed.
The weaver couldn't build the machine, maintain it, adapt it, take it home, use it for other processes, try their own creative designs, open a competing shop, work in other textiles etc.
But agentic tools are not that. They DO allow all that. In fact software engineering has always been about automating things, and agentic tools just allow more people to automate more things, AND create more software the way they want it.
LLMs just don't feel like a loom. They feel something that can be used by workers themselves. Like power tools. A very generic power tool. That's essentially free. Unlike all automation to date which required massive up front capital to automate one single process.
Maybe that's the difference - accessibility and generalization.
Obviously the correct way to view the arrival of AI is as an empowering tool for the ultimate benefit of us all, but just as obviously there are a lot of people who simply don't seem capable of understanding that. Obviously the correct way to react to the arrival of AI is to severely regulate the monopolists and oligarchs rather than the technology itself, but... see above.
Unfortunately there is no obvious way to deal with people who think the Luddites were somehow morally superior, or who otherwise want to rehabilitate or emulate them. It's like flashing back to late 2019, watching news reports come in about some sort of weird, unusually-severe flu that seemed to be spreading. Nobody knew for sure what to do then, and I don't think anyone does now. But we are going to have to do something to keep those people from dragging us backward with the self-centered psychological pandemic they are trying to spread.
In my opinion LLMs can write code but they don't replace human skill and judgement in software engineering.
They can help you make decisions but ultimately humans still have to be responsible.
When your payments gateway goes down the week before Christmas you can't just say "well Claude did that, I have no idea how it works or why it can't handle the traffic".
The LLM bros want to claim that role will be handled by... who? The CEO is going to start prompting Claude at 3am making random changes in production to fix the payments gateway? The product manager? The cleaner?
Or you have some agent loop constantly making random changes in production to crucial core parts of your codebase with no human supervision?
I just can't see how actual software engineers who understand how the system works aren't still needed
That person/company is a software engineer/company in practice if not title. Anyone who tells you this role won't exist is probably wrong. At least practically wrong for 5-10 years? Dunno.
But the Luddite question is not existence but scale. Can all 4 million software engineers find jobs if this is all that's left of software engineering. Also dunno
Actually no, you ask Claude to make proper alerting and telemetry then ask it to make a thing that checks alerts every 10 minutes and takes whatever action is needed to fix the problem.
- knowing that you need monitoring and alerting and how monitoring and alerting best practices are generally done so you can prompt Claude, because by default it won't do it if you just ask it to bash out features
- trusting Claude to change the code handling payments in production with no human review
First one is the easier issue, maybe there is some way a non-technical person could get there by blindly following some generic guidance and feeding it into tickets.
Second one I would not be happy with as a customer or shareholder of a business
Unless the entire world bans it, how would that help?
U-S-A
It's hard to predict an outcome "somewhere in between of two absurdities"
Obviously we can't effectively ban already released models, but a ban on models with >10T parameters for instance, is quite feasible. It takes a ton of power to train and operate those, and infrared satellites can spot that.
I'm confident China would be very happy to sign on to a world wide ban.
1. AI is one of the only high tech areas where America is ahead of China. If AI is knee-capped, China would indisputably become the world leader in industrial tech and high tech.
2. China is releasing models as open-weight to commoditize the complements. A ban on 10T models would let them specialize and utilize the lower parameter models to extend their lead in areas like robotics.
And once the US & China signed on I'm sure Europe and the middle powers would quickly agree.
AI is too much of a field leveler in numerous fields for a ban on high end models to stick.
The US is (and has been, for decades) overconfident in its ability and stature, in much the same way the Southern States were, during the American Civil War. You see, the time-span from the end of the Revolutionary War to the beginning of the Civil War, saw a transformation of the ethics of the Southern elites. The revolutionary elites were (and you can see this in their writings and correspondences) embarrassed and ashamed of the over-reliance that their economy had on slave labor. They were embarrassed, because this arrangement did not live up to the egalitarian and liberal (as in freedom) ethos that their new country aspired to.
Within a few generations, the grandchildren and great-grandchildren of these elites, had lost that shame and embarrassment. They felt, rather, insulted by the disgust that their Northern neighbors expressed. And so, to compensate, they became _proud_ of their slave-owning ways. They became something that their grandparents would have been ashamed of. Eventually, the Civil War broke out, and, at the cost of 600 thousand lives, the abolitionists got what they wanted (or at least the beginning of it -- emancipation would take a century to truly complete).
Note that this victory was only possible because of Northern industry and immigration.
We are in a similar situation today. After the Second World War, the American elites (and the American people) were embarrassed and ashamed of the failure of their democratic and capitalist/free-market society, to live up to the aspirational ethos of egalitarianism and liberty and self-sufficiency. Nevertheless, there was a Cold War to be won, and that required the unsavory combination of labor-exploitation (read anything by Mike Davis), neo-colonialism, and foreign (and domestic) espionage and subterfuge. From the beginning of the Cold War, to its end, the ethics of the American elites (and, in a different way, the American people) had, yet again, transformed into something the elites of the post-World-War would have been ashamed of. You can see hints of this transformation in the writings of John Updike: A society that was becoming increasingly fixated on hedonism, luxury, profit, status-symbols, etc.
From Reagan/Thatcher until now, the American elite has felt insulted by the suggestion that they are, perhaps, too exploitative and too parasitic and too ignorant. They now wear their largesse and cold-heartedness and ignorance as a badge of pride, as an inalienable right, that they will let others fight and bleed and die for.
In other words, where the post-Cold-War elites would have been horrified by the writings of William Gibson, the contemporary elites treat them as a user-manual.
From 2020 to now, it feels like we are seeing a slow-motion version of the USSR collapse (not in the sense of redrawing borders, but in the sense societal and institutional decline). The whole AI frenzy _may_ be a desperate attempt soften the blow of future mass emigration (of the so-called, high-value human capital).
OTOH, the US also has a massive credibility problem. Nobody's going to sign anything without solid mechanisms to ensure the US won't ignore or tear up the agreement in the near future.
On the other hand, China could just replicate GPU technology or adapt its models to its own chips (already happened), and then push that onto the rest of the world for cheap. Of late, they too have been very aggressively building out their data centers in Europe and the Middle East. China also controls the resource chain for most essential minerals (thanks to decades of US offshoring).
Huawei's hardware is apparently good enough for inference and the CCP are actively promoting their domestic AI gear ... among other things by banning high-end chip imports. The export bans from the US and import bans from China both have same technical targets but different political goals. The analyst commented that Huawei's gear is not yet good enough for model training.
The situation is going to change once China can - to quote the analyst - "complete a training run on Huawei chips". When that happens, China will start to really crack down on AI chip imports to ensure that their domestic technology is forced to catch up.
The main benefit of the closed models is the regulatory control to say, insert guard rails against hacking things. That's also the main drawback as you now have handed control over this technology to a handful of companies, who will reap all the rewards of this captured sector.
It would be like the google/apple duopoly on steroids. Everyone forced to buy in to one of two ecosystems, with limited to no ability to undermine the duopoly via control of the device hardware. One can't make a real competitor because it requires access guarded by existing relationships and money. For the AI houses, it means you need approval from these groups to create in a competitive manner. That, IMO is too much power for an unelected organization to carry.
AI will likely only provide real benefits to humanity if it's open. But if it's open, it will also enable extreme harm in the form of cyber attacks, scams, and destruction of IP.
By keeping AI closed, it will bring a new form of imperialism. Instead of the core maintaining IP and the periphery suffering the secondary effects of industry, the core will maintain the capability and the periphery will be locked out of competition due to silly nation-state rivalries. This won't benefit humanity, it will lock us into the same cycle of misery for marginal gains measured in stock values.
So there's two ways AI can reasonably go: Cyber punk corporate overlords or cyber anarchy due to the inability to put a cork on the genie's bottle.
Both seem terrible to me. What good is space travel without freedom? What good is automation if there are incentives to maintain profit over stability? What good is the technology if any usage of it exposes you to fraud, violence, and blackmail at any moment?
Big fans of LLMs never ask "to what end?", they repeat shallow near term "wows" without considering the second and third order effects. "wow think about how good LLMs can spot cancer in scans!", great! but who will afford the scan without work? Will the overlords commit to luxury gay space communism? Or will they simply say "tough luck" and let that cancer fester?
Real AGI is incompatible with our current power structures, yet those current power structures desire the AGI. This cannot end well. Stagnation and stability would be preferable to the level of misery that would follow.
That's part of the reason why I really appreciate China just throwing a wrench into the closed model system. As long as there's a profit motive for AI, it will only accelerate our current problems.
There is only one way AI will solve climate change. If it leads to job losses for 80% of the people, so consumption drops to nothing. Nobody drives anywhere, people stop buying things they don't need with money they don't have to impress people they don't like (because EMIs/BNPL need income). If everyone goes down to subsistence level, back to the feudal system, just 100 - 10K billionaires and everyone else low status barely surviving, TFR will drop to under 0.1, delivering the climate change solution.
AI is what we’re doing instead of solving climate change. With the money the US has spent on AI in the last few years, you’d have approximately enough to fully remove all fossil fuels from current US electricity generation and replace them with renewables.
Climate change will be solved—if we do it soon—as soon as people start treating it as inevitable no matter how many dollars it costs the same way they do with AI. It will not be solved as long as we keep ignoring the existing, known solutions and hoping that some unrelated technological advancement will just happen to make it cost less to solve climate change.
I think investment in AI will increase the likelihood of successfully dealing with climate change as it's a research accelerant
If people get used to trust what a smart sounding oracle tells them and that's consistently aligned with wha climate science says, then perhaps we have a hope.
If we end up with having different models that Say different things to different people then people will select the ones that reinforce their echo chamber
Plastic pollution is a different problem in itself but one that often is counter to fixing climate change in the short term.
Our economy adapts and find equilibrium when things change. We don't need AI to get oppressive societies. We already have those. Does it make it easier to have one? Does it make it harder? I donno. But a lot of this depends on governance and policies. If AI is used for "good" it could mean a better life for everyone. That's not necessarily opposing the interests of AI companies or their owners. What sort of jobs will those people have and how different will those be than the job market today- I donno. Maybe we'll need more software engineers because we'll be building x10,000 the amount of software or maybe we'll be part of a service economy or planting trees. But the economy isn't as simple as "AI can now write code therefore this is the end of software engineering" just like making software engineers a lot more productive (and we have) didn't reduce the number of software engineering jobs. It created more.
In terms of AGI's compatibility with power structures. I don't think there's one answer there. Does it fundamentally change the different forces driving us? If AGI is smarter than us then what does it want? Clearly a lot of what AI offers today is still tied to existing systems. Can spot cancer in scans feels like a good thing. Somebody has to pay for it otherwise why would anyone build it? Like anything else there will be supply and demand and if supply is high and demand low the price will go down. Why should it necessarily drift more towards anarchy or totalitarianism? Why can't it be "free/capitalist + AGI"?
Anthropic has high margins and is working on the others but hasn't achieved them yet. It's margins are threatened as it and competitors are lowering prices regularly. P/S for commodity producers is well under 1. But lets assume tokens don't completely commoditize so a P/S of 1 is justified.
That's $1T. But Anthropic doesn't have $1T in revenue yet, and it's risky due to spending commitments. So you have to discount for both time and risk. That's where the other order of magnitude comes from.
The standard P/S ratios apply to established companies, but Anthropic is not that. The current valuation is not based on the material state of the company, but on future projections.
It's not just enough to have enough revenue to "justify $1T value", there needs to be massive growth in it to justify the risks.
There's a lot of information about there. You'll be disappointed by the simplicity of it but generally speaking analysts will be using a DCF model with a terminal growth rate. It unfortunately doesn't exceed the math level of grade 10 in my country ;) On the plus side, it's very easy to understand.
I know this is only a small bit of your overall point, but I don't think this is an accepted fact of economics.
The "opportunity" cost is true is you're comparing redundant work to productive work, but OP was comparing to unemployment, which is a different case. A bit of a simplification, but less employment leads to less spending which leads to less money to companies etc.
Here's a nice quote from Keynes about redundant work:
> If the Treasury were to fill old bottles with bank-notes, bury them at suitable depths in disused coal-mines which are then filled up to the surface with town rubbish, and leave it to private enterprise on well-tried principles of laissez-faire to dig the notes up again (the right to do so being obtained, of course, by tendering for leases of the note-bearing territory), there need be no more unemployment and, with the help of repercussions, the real income of the community, and its capital wealth, would probably become a good deal greater than it actually is.
The problem with AI is that work on an unprecedented scale will be made redundant and economic power will be concentrated into the hands of very few and there will be no way out in perpetuity.
No, because rich people will always be able to buy more than you.
> Only if you don't use that power to provide something useful to others.
Others with money. Never forget the with money part.
And why won't those people with money just cut out the middleman (you)?
So what? AI could give both poor and rich more economic power.
The pie is not fixed. It can grow.
1) The AI companies maintain massive amounts of control over that "intelligence for sale". Not the least of which that they could simply refuse to sell it to you. It's also a big part of the sales pitch; A company can use AI to externalize the "institutional knowledge" and skills of of it's workers into something the company controls rather than the workers.
2) Because it's cheap, it has no value. Anything "useful to others" made with AI is currently drowning in AI generated submissions.
The power you bought is a drop in the water, because if it really is power, the AI overlords will just copy or they'll have vastly better economies of scale.
Knowledge simply is no longer a bottleneck, we're back to raw resources, real estate, political power and so on, medieval stuff in political terms.
We have statistics that gives us a shared expectation that we can work with.
Society is much more enlightened today that it was then.
So I’m optimistic that the pain will not be as severe or as prolonged as then. Doesn’t mean we should just hope but I simply don’t share your extreme pessimism.
The one thing I would concede is that there’s a lot of uncertainty about the exact contours of what we’re approaching but that was also true of the Industrial Revolution or other transformations earlier in humanity’s history.
I choose progress over perfection.
Love that!
Half a million, in 2026, is a lot for me and you, but pennies for even a modest sized business. And this is before prices go down.
Ive been using 9b models on my 4090 for coding and command tasks for over a year now. It's great! You obviously can't be like "Use huge framework and context, and just make me an app" but you can be like "Employee said vpn is running slow. Follow the process documented here: X to diagnose and propose a fix" where X is docs written by a Chinese model for 10 cents.
Is this a "commoditize your competitors" strategy when your motive isn't profit but disruption?
What you are probably referencing is "dumping". That's when you sell undercost to disrupt your competitors.
That may be why. Certainly the motivation of the CCP is supporting it isn't short term financial gain. But it's probably more about maintaining independence from a dependence on US AI long term than dominating the AI industry.
But I think it was originally just a side project to keep quants interested.
https://gwern.net/complement
Go read the papers on it if you want to hear some of the immensely plausible theories as to why micros and the internet and all sorts of what we think is technological innovation doesn't show up in economic productivity data.
I know why rich people who own many shares of aggregate economic capacity want me to believe that it is the only important thing and that sacrificing aggregate capacity to fix distributional problems is never worthwhile. But I don't believe it. The industrial revolution was gnarly with unnecessary distribution problems that ensured crippling poverty for the masses amid extraordinary leaps in aggregate capacity. This time around I don't want a century of Cyberpunk Oliver Twist before the peons are allowed a dignified life.
Adjacent, improved argument: taxes are a better fix for distributional problems than throwing stones at windows. I agree, and I will add that broken windows are a better fix for distributional problems than guillotines. I really really hope that we can convince capital to pay taxes because that's the only future where things don't go to monumental shit before getting better.
Which basically only changed when people fought for their rights
> I really really hope that we can convince capital to pay taxes
I sadly can't see that happening without ww3
Now, let's say the handyman whom the rich guy paid to fix the broken window uses that money to start a company that successfully competes with this 99% person?
If you generalize this, the conclusion is that broken window fallacy assumes that the pie needs to grow in a monotonically increasing fashion, when in reality sometimes the pie needs to shrink a little in order to grow a lot, if that shrinkage results in redistribution of wealth.
Whether fixing the window is "wasted" time and resources depends on supply-side slack. If there's a warehouse with millions of windows, or the handyman is underemployed and has nothing better to do, then that's not necessarily a linear waste.
That doesn't work when the rich guy's assets are factories and datacenters, because those will not be hit by inflation. Inflation only dilutes fortunes that are no longer invested in profitable operations.
In reality China is taking market share all over like crazy with cheap, good products and services.
Someday I think we'll wake up and realize how much wealth we threw away from sheer lazyness and failing to do something about our bloated monopolies that have little reason to compete or innovate or expand - and being a big stupid bully that pissed off everyone else - but by that time we'll have lost most of the world's business to china, cars, software, ai, aerospace, online shopping, etc. All gone.
Of course, we have total faith and trust in the US and it's companies, so would never think of buying tech from elsewhere.
Clever strategy, please work on it right now.
Are we talking about the real economy or the stock market casino? I have a feeling one doesn't care about facts and logic
Oh, goody. They left room for their friends, Google and OpenAI.
The goods and services built atop "tech" should start to become cheaper as well. For knowledge workers this is terrible, but for consumers, this is great. I think folks worrying about an economic collapse are only looking at this from the perspective of knowledge workers and not from the much larger population of consumers.
The effect on the AI companies with pure competition is to make them basically worthless unless they find ways of stifling competition. For example by banning open weight AI in the name of safety, forming a global monopoly, buying up all the RAM capacity in the world, etc. All of this stuff is going on right now.
The value is there. The true question is whether the wealth will be concentrated and captured or released and distributed. I'm in the camp the likely outcome is actually the former. A huge amount of the wealth of AI has not really been unlocked yet... consumer robotics, disease cures, universal constructor type machines. Furthermore, once we reach superintelligence, eventually someone will inevitably fuck up by letting the dam break by ablitering a model in the wrong way. This will give governments and companies rationale to lock up the best AI to develop frontier level IP.
The cuts happening now are based on the belief that the AI will get better quickly enough that customers won't notice a degraded service long enough to leave, plus an understanding that every other company is making the same cuts so your customers won't have anywhere to go anyway.
It should be obvious, to any grown and civilized human, that has more than a gram of brain, that it is impossible to receive SMS messages, without access to the cellular network, which requires a SIM card! (You see, you do not strictly _need_ an AI/LLM to get systems that appear to have been designed by drunk aliens).
So, of course, I try to contact support. I was given a chatbot that had many great ideas, that were mostly hallucinations (e.g. it would give menu-instructions that make no sense, for any phone model, that has ever existed). Eventually it gave up and told me to call support.
I called support, and got another chatbot, with much the same results. After spending hours, and getting pulled by the nose by various support-bots (none of which ever presented me with an option to talk to a human), I started screaming into the phone (this is my other phone, that does actually work, and has a different number): "I WANT TO TALK TO A FUCKING HUMAN NOW".
There was a silence on the other end, for a good 10 seconds, before it started ringing, and I got an actual human on the line. I explained the situation (and by the way, their English was impeccable, and they were very knowledgeable), and they guided to the very precise steps to activate my phone (and to avoid a pitfall that emerged from my previous attempts (e.g. they said to discard the new SIM, and to activate the eSIM, because my previous attempts had put the system in an invalid state that would be difficult to recover from otherwise)).
I want to make it very clear: I should not have had to talk to a human, _nor_ a bot. I should have been able to insert the fucking insertable SIM[0], and been connected to the network within minutes. The _excuse_ for the exceptionally poor engineering was something along the lines of _precaution against identity theft_ (which, if it is truly the reason, is even more embarrassing and shameful).
[0]: By the way, the original SIM should not have died at all, given that it was in the phone for only a few years (my other phone is a _decade_ old at this point and is using the same SIM card, that it had, when I bought it).
1. You’re absolutely wrong here.
2. Those call center agents would like a word.
3. Having worked with many “AI support bots” in both chat and over the phone, they are horrendously worse than even the dumbest call center agent.
4. Seriously, are you against people having jobs so you must make up hype? AI support agents are god awful.
In some cases, you wouldn't even know if an LLM was "helping" you. And what makes you think that the humans who solved your problems didn't in some way use AI?
To be a consumer, and participate in a market, one needs liquidity.
If consumers have nothing to offer that is fresh to the market, their resources will drain.
Labor and Knowledge Services, are 'fresh' to the market at the point of sale.
If you take that away you have nothing, what do you propose the human do?
It's the same architecture as buying from communist China.
It will likely codify the assetted class, imo.
Lets say labor is displaced in both scenarios. That does not mean comsumption halts. Capital still needs to find a home. It can be overpriced watches for the techno elite... vacations to the surface of the moon... anti aging drugs. Lets say no one company captures the value, then the value goes to the owners of machines doing the actual work.
they will always have something to offer (it doesn't need to be "fresh"), just usable. Their labour or body, for example. And it's not like AI can automate every possible job - jobs like toilet cleaning, and a lot of service/hospitality work is hard/impossible to automate (at least in the medium term - may be long term we will all be served by androids).
The influx of labour from the (now redundant) high skilled professions means you now can have cheaper labour for all these services.
> codify the asset class
this has and always will be the case in a capitalist market. Unless you propose revolution (of the communist kind).
Can't see why either should be true
Humans buy shit, not machines.
The economy absolutely can be run by basically robot slaveowners and the real question, as dystopian as it sounds, is who will own the robot slaves
And I think this (in our revolution) will also generalize to 'cognitive work' companies as a whole. This is why I think LLMs are going to radically (and positively) reshape the world, but in a way that relatively few seem to be predicting for now.
If AI can run robots doing physical labor as flexible and well as humans then it could drive the value of all labor to $0.
Meaning radical shifts in short order are certainly possible.
> you're opening the door to more 'social' labor, entertainment, art, sports, and the like.
That assumes money hasn't been captured by the capital class. Today how many working class people can afford to attend professional sports, live theater, live music? Even viewing the UFC fight at the White House--supposedly a celebration meant for the enjoyment of all citizens--costs money.
As for the Marxian stuff, right now the most valuable 'means of production', by far, is what you and I are both using right now to type these message, and what basically everybody has, or can comfortably afford with like a week of minimum wage work. The biggest remaining asset of the 'capital class' are cognitive workers, but LLMs are going to similarly make that asset widely available.
Society has become dramatically more unequal in terms of numbers, but in terms of 'stuff' we're likely more equal than ever. Royalty and peasants lived dramatically different lives. Yet your life, assuming you're just another mid-class normy, and the life of a billionaire aren't fundamentally different. They have more liberties and nicer things, but on a day by day basis you're often doing pretty similar things, eating pretty similar things, and just generally living pretty similar lives. I think LLMs will further decrease the 'stuff' divide, and further increase the numeric divide. That doesn't really bother me.
1) you have no clue about the lives of billionaires 2) there are not going to be mid-class normies, with their work replaced by AI, they will become trailer trash scraping for bare existence
is one of the better quotes for grounding yourself when people start talking about economic problems to big to know exactly.
to your question.
If an industry is 'worth' 1 billion, meaning it generates 1b in revenue lets say at 800m of investment.
and that drops to 80m of investment and 100m of revenue due to productivity improvements then competition, the economy is now free to invest that extra 820m in the next best productive use for that capital, while still getting exactly the same services.
If the companies investing in ai can manage 80m of investment and 300m in revenue as the end state. They're richer, the consumers getting the same service cheaper, and society is richer due to the extra 700m being used to produce other goods and services. Thats the dream.
We were lucky the maturation of the internet coincided with post-GFC low interest rates or else the internet may have caused deflation.
It's not deflationary because inflation / deflation is always and only the increase or decrease of the money supply
You can solve it either way. Anyone claiming to confidently know the answer to this question is making overconfident assumptions about the relative velocity of money between AI shareholders and consumers, export (and thus geopolitical) dynamics, resource price elasticity, social-safety net politics, et cetera.
In some markets, AI driving down costs will drive up consumption. So people won't necessarily lose their high paying jobs, they'll just do more work. Or, at least, corporations will capture the same revenue.
However, what's really untested is what the commoditization of knowledge is going to do to society. Having an agent at your fingertips with essentially a masters degree in every subject on earth, and the ability to apply that knowledge effectively is a pandora's box.
The closest analog is what industrialization did to craftsmen. How are people going to be qualified for jobs when AI can do a better job than anyone? But I suppose that's a political issue more so than a technical/economic one.
If it requires even some tiny amount of work from you, then you have just replaced a 20$ SaaS with a 20$ anthropic subscription AND you get to maintain it yourself AND you have no one to blame for mistakes. I would say that's a bad deal.
(See all other productivity gaining technologies that increase gdp).
Which is why other economic disruptors like rail roads eventually lead to economic expansions, and were net positive
It doesn’t not look like that is the case with AI. It’s going to cannibalize the economy and not bring the expansion of previous developments. Which is a sad state for humanity
AI isn’t a magical force that takes your job and prevents you from ever working elsewhere: it’s a technology that will increase productivity and impact jobs via creative destruction.
It is the final piece in the re-feudalization/fascization(does this world even exist?).
We're back to feudal times where birth given rights to land, energy, resources, AI is all that matters, no more upper mobility.
There is no real moat anymore unless they force it from the regulation side, hence the heavy lobbying.
2000 fable subagents still need the context we humans drive.
Yeah I'd say if you were being paid six figures to do some office work thing (including software development) and now only 10% as many people are needed to get the same work done, it is very much going to take your high-paying job and make it considerably more difficult to find another one that pays the same. At a minimum, there's going to be a whole lot more candidates to compete with as you attempt to work "elsewhere."
Unless it's a _very_ slow transition to whatever the "next thing" that "can only be done by humans" pops up, this "creative destruction" is going to be devastating.
I think the reason is that we can automate what is correct, the invoice matches the PO, the goods are receipted etc.
But most of their time is now spent on exceptions. The company didn't charge what the po said, there is no po, the supplier is not set up, the payment got returned from the bank, some goods are missing.
Now they can't fix any of those things themselves by design, because of segregation of duty controls, and specific humans holding Delegation of Authority for changes. You see if you can order things and pay for them yourself (with the companies money), fraud is inevitable. Giving this to an LLM makes fraud inevitable.
Then we say lawyers and auditors... the personal liability for signing key documents in those industries lies with licensed people. An LLM could prepare files, but that licensed person still has to review them, and employ other reviewers to ensure that the file is correct.
I just couldn't understand it. If society no longer needed coal mining, was he seriously proposing they should all keep their jobs anyway? I couldn't take in this way of seeing the world.
So I'm pretty sanguine about the same thing happening to my own career, if it does. In the end, you can only expect to get paid for providing a service society needs. I can't think of any other remotely sensible way of running things, myself.
It was a fair expectation when nearly everyone had a service to provide. But why now? No one asked to be born.
We need to move beyond the "everyone must contribute something" mindset, and instead place value on simply life itself instead of what that living being provides or does not provide because there simply isn't room for everyone to provide some service to society, everything is covered. The things that are not covered are generally high skill and high education (expensive) positions that not everyone has the ability to fulfill anyway.
Basically, and before we have mass unemployment and homelessness from AI, we need some sort of universal basic living floor as a safety net. Make sure no person is able to go without housing, food, utilities, education and healthcare.
I.e., this is the problem of how the economy is organized, not of not having enough people to do all those things that need to be done. But we are too deep into 'capitalism (greed) is the best' ideology to clearly see that and do something about it. The way the things are going, either pitchforks or terminators are going to come out. The AI industry is betting on terminators.
Anything else, in the long term, amounts to murder through abandonment. It's no different than leaving a wounded person in a forest while the group marches ahead, even though the group has plenty of food and a stretcher and a car, and fit, healthy people to lift the stretcher into the car.
I think you have a moral right to offer to provide goods or services for people. They aren't obliged to want them though. I think it's on each person to find out how they can provide something people want.
Just because I enjoy coding or am good at coding, if nobody wants to pay coders then it's on me to find out what I can provide that they do want.
Sure. But what if that is "nothing"? Or something that pays so little you literally can't afford the necessities of life? Or it's only work you find objectionable (e.g. crime or prostitution)?
For all of human history, given enough time, money, and work, anyone could retrain into something that paid them enough money. "Try harder bro" used to be trite but true. In a true AGI/ASI world it may be an impossible demand. Is it still fair?
I see that as perhaps possible/likely in the near future.
The idea that there will be NO services humans can provide which anyone wants and therefore no possibility of finding any niche is so far beyond any experience or knowledge I have I can't even really speculate but I assume it would lead to either insurrection or wealth holders providing UBI.
Being in a competitive market is for us to bear but not them
Personally, the second option seems even less fair than the first one.
Some people own land simply by virtue of being a distant relative of the first group of people who successfully defended (or took) it from other people with violence. Some people don't own land because their distant ancestors failed to.
It's just all so weird, when so much wealth is tied up in an asset that no human had any hand in creating.
How though? Just saying "do it yourself" doesn't make it so.
> or other people who also did not ask for them to be born have to provide for them
They have to personally provide for them? That's not quite it is it? Society as a whole is more productive. Automation provides for them.
> Personally, the second option seems even less fair than the first one.
Why? Consider the alternative.
No one is saying bring back coal mines or CRUD software jobs. What people are saying is that the working class is being ELIMINATED from the economy. If you are in the working class your options are: service/construction work. Be an electrician, or a waiter, or lawn care professional. And they make jack shit for money. While the managers of AI make millions, so the economy shifts to them, away from people who do labor. But dont forget they all still have to pay taxes which fund all of the massive debt for AI and wars and all other BS.
You cant think of a remotely sensible way of running things because you dont understand what is happening. Society is COLLAPSING. The managerial class is systematically eliminating the working class. So enjoy your pretentious outdated theories about capitalists and socialists. While the world around you decays for the next 5 decades. I assume you're a bit older because (im in my 40s) thats the only people I see with this delusional view of society, they simply dont give a shit about the future.
Well guess what, unless you're willing to be paid less than Haitian immigrants to do that work, it ain't gonna be booming for _you_. Ditto for construction, landscaping, et al. You know, those jobs Americans "won't do."
That market is absolutely not going to be booming. It's a societal problem that large numbers of old people are exploited, broke, and brainwashed by generations of social media. That's not a booming market, that's a captive army of peons valuable only as long as human votes matter. They could as easily be disposed of, and so could you.
But they can vote for taxpayers to support it.
Or like elevator operators, who used to make good money. There used to be 120,000 of them in New York - now there’s about 90 thousand elevator technicians, in the whole USA.
There’s not always a replacement job. Or there is, sweeping the floor.
There always is a replacement job, it’s not necessarily sweeping a floor, but it’s not 1:1.
People have to understand that your compensation is a function of the value you generate or capture: the security guard might work 80 hours a week, but a security camera works 168 hours a week and costs $100 on Amazon. He’s just not creating enough value over his alternative - the Puritanical thinking of hard work has inherent value isn’t compatible with the real economic world.
100 security guards cant make in one year what one person with 10 million dollars can "capture" from investing in the stock market.
The entire system is manipulated to benefit asset owners and disenfranchise the working class.
These types of lies and delusions are exactly where we are today. The "real" economy has been dying since 2007, everything else like new stock market ATHs are just lies the public has been fed that doesnt understand it just means: companies, banks, and governments are perpetually increasing debt and kicking the can down the road.
2) 10 million in capital creates more value than 7 or 70 security guards and is rewarded accordingly. Again, your alternative as a security guard is some cameras and alarm systems, physical security, etc. $10M can secure a lot of premises for a one-off expense, without payroll and worker's comp costs.
3) The "working class" is just fading in relevance as unskilled labor is pressured on all sides: international manufacturing, outsourcing, automation, and technology. If capital or intelligence can deliver more productivity, we shouldn't kneecap it just so a high school dropout can profit while leaving society poorer overall.
I think we should do exactly that. Besides, that same wealth relies on stability which is only found if access to a good life is easily available.
If we keep going at this rate, the public burning down data centers will be the least of the worries of the wealthy.
The economic notion of value is wealth-weighted. Feeding an orphan has zero "value" because the orphan has no money with which to pay. Merging the banks into a mega-bank to hike fees and offload gambling debts creates oodles of "value" because it makes rich shareholders even richer.
You are trying to slip a self-serving virus into the boot code of this discussion and I do not appreciate it.
I don't care what the argument is, nobody should be unable to afford an apartment after working 40 hours a week, much less 80. We need to correct the system, forcibly, and stop pretending this is a "divine invisible hand" we can't simply shake.
I got a robot to cover that one already.
Railroads didn't do that. Horses and wagons remained in wide use after the advent of railroads as a last-mile mode of transport.
Before railroads moving masses of goods and people typically required navigable waterways. Either natural ones or canals. Railroads augmented waterways and allowed that mass transport to expand far from those waterways. The underlying steam engine also helped with expansion as a factory or processing plant could be built near a resource, be powered by a steam engine(s), and ship finished/processed goods via those railroads to some market.
Railroads (and steam engines) expanded the economy far more than the supplanted any particular part. When they did supplant something the economic expansion largely provided new work for whoever's job was obviated.
Knowledge work replacement by AI doesn't necessarily offer the same economic expansion. If it replaces one type of knowledge work there's little stopping it from replacing another type. OpenAI and Anthropic are only viable if they capture significant portions of the US GDP. Anything less and they can't even pay their debt obligations let alone actually make money.
There's no real indication all those replaced workers could find any other jobs, certainly not ones that make as much money. And UBI will never happen, the billionaire class will never pay the taxes that would be required to support it. So that leaves lots of people left with no way to support themselves.
Many horses and wagons lost business due to trains, just because some remained doesnt change the fact the others lost their place in the system.
Proper comparisons are about numbers not nothing changed because of one cherry picked item.
> Knowledge work replacement by AI doesn't necessarily offer the same economic expansion.
Most people using AI are not knowledgeable. A lot of AI seems to be dumb animations or flyers or silly graphics or emails nobody reads, or more powerpoints. Thats not knowledge those lines of work, are a tax on everyone else.
Nobody needs advertising, or pointless sociual media type communications.
This reads like the line 'The rich man and the homeless are free to beg for change'
These people are literally saying their product is dangerous, and that it will eventually put a large portion of the population out of work. Until we change our economic system so that "people are entitled to a decent quality of life" err sorry, 'free stuff' in your words, even without work, don't think the voting public is just gonna stand by and let them do it.
It's rather callous, bad faith framing to say 'you're not entitled to free stuff' when that 'free stuff' is the very things they need to live, and you've destroyed their ability to acquire it through work.
Also, nobody has "destroyed their ability to acquire it" - the market has set the price and they have to adapt. If I create terrible paintings, I shouldn't still be entitled to millions because someone else's paintings sold at auction.
Members of the voting public should also be free to call out a system that has a meaningful voting bloc (20 to 47% depending on how you want to slice it) who consistently vote for more free stuff, while contributing nothing to the system. A system where everyone takes and nobody makes is not a net good for society.
You want to talk about 'good for society', maybe talk about the fact that we were on course to pay off our debt before bush decided to cut taxes for the very rich. While my pocket book appreciated the tax cut, I would have much rather had a fiscally responsible government than a tax cut.
What the neocons never mention is that the nordics do not seem to have any problem with having good social safety nets while being fiscally responsible. The rich simply pay what they owe to society.
It's generally worse for one man to have a billion dollars than for a 10,000 to have 100k. Those 10,000 people will spend that money where as the billionaire will almost certainly won't. An economy can only take so much 'investment'. Our economy depends on consumer spending.
If the top .01% get fabulously wealthy automating the bottom 90% out of good jobs, we will all be worse off for it.
When we talk about 10,000 having 100k we usually mean they have it in somewhat liquid assets: savings, annual income and even such illiquid asset as housing is something that can be sold relatively fast and invested.
Billionaire having 1 billion is usually means the same as company having trillion market cap: 30-50% of it can disappear overnight and nothing in the world will change really.
GDP growth is usually measured in “volume”, which means that the demand for a good or service doubles while the price is divided by ten, the GDP share for this good or service doubles even if the amount of money is 5 times less.
The reverse is also true though, which means if something becomes twice as expensive (say oil) and people starts buying less of it, it's share of GDP will diminish. Because of that, the GDP deflator is underestimating inflation, and overestimating actual growth in times of high inflation.
Note that this was further complicated by a change in the tax law which made the backlist a liability rather than an asset.
Economies don't really lose jobs. Buyers lose demand for products, and employers lose demand for particular kinds of labor. That doesn't make the labor less valuable to anybody else, it just means their current employer couldn't afford it or use it. The same number of people with the same set of skills still want to work as before.
It's actually common for labor to increase in value even when it gets disrupted! What happens is a rebalancing of the labor market away from obsoleted industries/company/products nobody wants, towards those that leverage the disruptive technology. Consider dotcom: eventually some of those employees left tech, and the rest created the actually-valuable Internet businesses because they regrouped and started different things.
Otherwise when there isn't enough demand for employable people to find work, central banks lend (or print) more money until employers want to hire them, because they're leaving money on the table from consumers' unmet demand if they don't.
In short: economies don't produce jobs. Jobs are what happen when demand on the labor market meets supply; creating more aggregate demand is easy but even then, it creates demand for labor to provide what people want to pay for.
People are losing faith in it, basically.
Efficiency would drive Real GDP higher though. Goods and services produced with lower input costs leading to lower inflation. Plus, presumably, widespread layoffs creating a negative economic shock. (Labor surplus from AI displacement would hit services pricing too)
GDP simply tracks the rate at which money changes hands in the economy. I'm not sure the AI proponents touting 15% GDP growth are really aware of this.
Nominal GDP would likely catch up after the big shuffle/unemployment cycle. But probably not initially if we do see mass layoffs
Pre-AI, if you spent $X you got A units of output; post-AI:
* if you need 'only' A units, you will spend less than $X (technically lower GDP)
* but if the limiting factor was not the desired units, but your budget, you will continue to spend $X but get >A units (same GDP)
Additionally, if each unit is less expensive, this allows the end-customers to spend less on it, and reallocate some of their budget to other things, which would keep the overall GDP the same.
Yes, in the short term that's true. In the long term, the concept is that that particular activity generates less GDP, but frees up the human capital who can then grow the economy.
This is true for any economy, not just AI and the central contradiction of capitalism.
Automation raises profits for a brief moment until everyone else has caught up.
Once the new profit margin equalizes across an industry, competition sets in and margins shrink.
To keep margins up, labour costs need to fall further. That is why companies like Nike or Adidas produce in low-income countries. The investments in automation is such that profits can only be maintained through mass production, low wages, long working hours and raised demand through brand lock-in, classic marketing.
The current open bet is that equalization of AI technology can't be achieved across industries because of the important entry barrier: capital and compute. Monopolization should hold long enough to make the rest of the world market dependent on US AI infrastructure.
The rest will feel the devaluation of their work, they will get poorer - some straight away, some indirectly because of the shrinking customer base.
Compute isn't a barrier to entry for long; I mean, even right now, you can rent a H100 that is sufficient for most businesses, for use during business hours, for $500/m. As many users as it can handle, as many tasks as it can handle.
And that number is going to fall off a cliff once the hardware shortage stablises.
Seems like the only way to avoid that is Chinese ai. Are they the good guys then?
Spoiler: it had the opposite effect. Cotton became so much cheaper than demand exploded and the demand for cotton basically doubled every decade after its invention.
So AI is going to destroy certain industries and others will have much lower demand for employees. The optimistic view is that AI will be like the cotton gin. Those industries will get cheaper and demand will go up or new industries will form in their place, sort of like the telegraph, electricity, radio and the Internet created industries. The pessimistic view is that this won't happen. There will be a tipping point where labor supply will exceed demand and this will put severe downward pressure on wages and create an unemployment nightmare that we're simply not prepared for. The last time this really happened was the Great Depression and at least then, the majority of the population was rural and way more self-sufficient for food and the like, not that I want to diminish the devastating effects.
I'm in the pessimistic camp. I believe automation in general and AI in particular is destroying labor demand faster than it's being replaced. You see this with decades of stagnant real wages (in aggregate) and the net effect is that we're creating a massive wealth concentration in a handful of people.
As Alfred Henry Lewis put it in 1896, "the only barrier between us and anarchy is the last nine meals we've had." And now we live in a country with more guns than people so we'll see how that goes.
THe primary moptivation for the Tech Bros is of course selfish, to pump and dump.
Stop pretending they are trying to help others, they are only trying to help themselves.
If it disrupted that industry, the things produced by that industry would become cheaper (if AI didn't make it cheaper than human labor then it wouldn't have disrupted the industry in the first place), boosting everyone's purchasing power for that thing. GDP calculations try to account for this, so GDP would go up (the same way GDP calculations try to account for inflation; a single product class becoming cheaper is like localized deflation).
Even in a competitive market no player will significantly reduce their prices because they want to maintain their own margins.
If you can take 80% of the market at $1 margins, that's better for you than taking 10% of the market at $2 margins.
I've heard this framed in terms of "surplus value," a very old (Marxist) economic theory. The debate is whether AI, which is intrinsically a machine, is capable of generating it - machines classically _transfer_ value. The reason this is important is because surplus value is what a capitalist entrepreneur exploits (whether they realize it or not).
So we don't actually know whether it is good for the economy, I guess we'll find out after the bubble bursts.
Yeah. It does. Permanently. That's what happened when the robotics revolution decimated factory jobs. Everyone was pushed toward finance and tech jobs, but most had to go into gig work to survive. Even the tech sector was kept up by dumping endless ZIRP money on it. Now that AI is removing white-collar jobs that used to spend to keep what remained of the economy going, that's exactly what will happen:
The system will get clogged. Because no new jobs are being invented at scale. We don't have any 'AI biochip integrator' jobs. We don't have 'Human-AI Augment synchronizer' specialties. We don't have any of the jobs that the AI industry promised, leaving aside any sci-fi jobs that were predicted in the past.
Let's face it: The majority of human civilization's day-to-day work has been automated. White-collar knowledge work is being automated now. What remains is some blue-collar work like handyman and electrician jobs. They will soon be automated by AI-using small robots. Not that those sectors would support the millions who were laid off from knowledge jobs when they all try becoming plumbers or electricians anyway.
It's time to admit that what Marx predicted 150 years ago is happening. Capitalism is destroying itself with automation.
But it hasn't, and probably won't, so it's not even worth considering.
You can probably get the best idea of what will happen from looking at real examples as it depends mostly on human behavior. I'm guessing with those the GDP attributed directly to them went down but the people who worked there got jobs doing other things so GDP in aggregate was fine.
I don't know about translators, I don't interact with them regularly, but I suspect AI isn't "disrupting" anything there either.
Any more "real examples" or are we just playing around with hypotheticals?
There’s not lot of fundamentals holding up the likes of ANT and much of the broader tech sector. The lofty valuations are based on hopes and dreams that are looking rather iffy.
It’s also sounding like the Fed will issue more rate hikes soon. Therefore anyone without free cash flow will have to pay even more to get their hands on cash.
For any company without strong cash flow and GAAP profits it’s likely to get very ugly from this point forward.
What mechanism are you positing?
There seems to be a lot of commentary about the timing of this floating around, and I’m watching it with curiosity.
[0] https://finance.yahoo.com/markets/article/goldman-sachs-morg...
Green shoes are soft obligations by underwriters to buy when the offering threatens to break within the first 30 days.
It’s never done to defend the book. You want to execute a green shoe by buying shares, worth more than the IPO price, from the company at that IPO price.
If an underwriter could reputable decline to exercise the green shoe in the market, they would.
From purely a pricing standpoint and the strike price on Morgan Stanley’s greenshoe option, the timing makes sense to me.
If you want to talk about the initial price, they've iterated on it a ton lately, so assume that whatever they set it at was intentional, and none of this is a huge surprise to the company.
[1] https://www.reddit.com/r/EconomyCharts/comments/1y0d8vy/anth...
[1] https://www.marketwatch.com/story/the-hidden-catch-in-anthro...
also who do you exactly think owns the shares to be able to sell? unless you think billions were used strictly to pump and dump the next day in an amazing coincidence of coordination (illegal btw), the ones who sold were the major holders.
A good IPO looks like trading starting at 5-10% above the IPO price and then gradually going up a bit more from there and holding.
In the real world, there's just no way that Tesla (annual sales ~1.6m, trending down) is worth more than Toyota (annual sales ~11.3m, trending up).
This is a very mature take. I don’t know that I completely agree but I definitely respect the angle.
Camp your money into passive broad market portfolios and sleep easy, maybe add some positions when we have market downturns. Unless you think the entire economy is illusory, then buy shelf-stable food and water.
I quite smoking but during Covid it occurred that cigarettes are almost as good as gold as far as a universal bartering object. I'd stash a few cartoons away - it'll come handy come zombia apocalypse.
My big three for bartering is alcohol, toilet paper, and hard candies. Cigarettes are great but I don’t smoke, so I’m hesitant to hold onto a lot of perishables that wouldn’t get used. However, like you said, a few cartons stored in a vacuum-sealed pouch in the freezer will last a decade and be good barter just in case. Same with nicotine gum and vape juice (although, juice itself well but the actual vape becomes the bottleneck, might research that later).
I pretty much think it is all screwed anyway. So slightly prepare buffer and spend reasonably on what you want just now in life.
Look at Netherlands' box 3 tax.
For one sense of real. Banks pretending their mark-to-model securities that the market valued as worthless was a problem because the market price isn't entirely illusory.
Ha ha, I tell myself things like that when I know I'm fucked…
I mean, if it just keeps going down… It kind of is real.
I hear that the smart traders know when to cut their losses. (Something about sunk cost fallacies…)
Who knows though what will happen to ANT next week. (And I'm not a smart trader, that's for sure.)
If you funded your position out of pocket then there are no margins to call.
Prices of stocks and other assets are absolutely real and (besides the tax implications) there is absolutely no difference between your example and selling at $20 and then immediately buying it back before the price starts falling.
> to $5/share, it doesn’t matter unless
At any given moment it doesn’t matter at all if you hold cash or stocks (as long as they are liquid) besides the taxes you might need to pay if you sell.
This doesn't even make sense, why would you not want to book losses at the end of the year? It's the perfect time to do it.
1. The underwriters buying to stop the price from falling further, but per reporting today they seem to be done there now so the stock is on its own.
2. The tiny float which inflates the valuation. The market cap numbers are highly inflated by the fact that if all the existing shares were actually trading on the market right now the price would plummet.
How does this support your argument? I think it tells us the opposite. Nvidia's growth is short-lived and its future is uncertain - it hinges on AI spending and data center buildout continuing at the same rate, and none of the major customers being successful in developing their own chips (all are trying). Their financing deals also increase their exposure to macro shifts. Investors are (righly) pricing in that uncertainty.
In contrast, Costco is a relatively boring business, but if you invest in it, you're unlikely to lose a whole lot of money very suddenly, so a decade or two of boring growth is priced in. I assume you're cranky about your NVDA stock dropping, but fundamentals go beyond a snapshot in time.
Whereas the GPU market attracted barely any competition; today there are at least 2 dozen startups attempting to slice NVIDIAs market. Several of these competitors are now themselves public (Cerebras). These are real indicators that the market expects real growth.
> I assume you're cranky about your NVDA stock dropping, but fundamentals go beyond a snapshot in time.
NVDA is reaching the point where they can take the company private with a fairly typical bank loan, while returning significant returns to investors. If the share price goes into double digits, it would be possible for a small group of investors to buy the entire thing. This is frankly ridiculous to think that NVDA is fairly valued at the present movement. My NVDA holdings are actually quite small relatively speaking, but the idea the stock is fairly valued is ridiculous. I've been buying more obviously.
Because their two major customer bases are:
1. Over-leveraged startups that look like they could implode at any moment, and
2. Big tech companies that, while unlikely to go bankrupt, are all actively building their own chips so they can ditch Nvidia.
It’s not a pretty picture.
Re the Costco comparisons, Costco is a reliable business that generates this thing called profits (which seem like an unknown concept to many in tech), and it has an incredibly loyal customer base that’s not going anywhere. My only complaint about Costco is that’s it’s too busy when I go there.
Also possible protip. At least in my city they appear to open 15-30 mins earlier than they advertise. The near empty warehouse at that time is sublime.
Did you see what happened to Intel’s stock price during and after dot.com? They did great after 2008 or so. Yet their share price hardly ever recovered.
> 2 dozen startups attempting to slice NVIDIAs market
That means they are losing their moat and their margins will collapse?
Nvidia might do great just like Intel in the next 20 years and its share price might never surpass the peak it reached this year. In no way these are mutually exclusive.
It is more than that. Top Nvidia buyers like Microsoft, Meta, Alphabet, and SpaceX have poured roughly $800 billion to $900 billion into capital expenditures in 2026, they can't keep doing that every year.
Major tech companies are seeing free cash flow turn negative and are increasingly relying on debt to fund data centers. Wall Street expects Big Tech capex growth to slow from 76% in 2026 to 25% in 2027 and just 6% by 2028. For a high-flying stock like Nvidia, a sharp deceleration in customer spending growth is enough to drag down its valuation.
It's hard to believe Nvidia GPUs will be sitting idle on the shelves next year.
[1] https://www.ubs.com/global/en/media/display-page-ndp/en-2026...
Remember the stock price of a company is only partially about what it is doing today, but also a huge amount what someone speculates it will do tomorrow.
Wal-Mart has been around for a long time. How often has it doubled revenue or profit or suddenly done something extraordinary?
The odds of that are very low. The odds of ANT doing it are significantly higher.
Same reason Tesla trades higher than Toyota. One may go up a lot (or down), the other is solid as a rock.
It’s gambling, and it does actually make sense.
Based on their IPO price investors seem to believe that the odds are about 100%. So the problem is that unless they do even better than that (in a rational market) you can only lose money.
Obviously (seriously!) this guy was 'taken out' because his person or strategic position in this very strategic sector was "not aligned" with the powerful. The people who can motivate NYTimes and WSJ to start "sex scandal" trains into motion.
As an aside; always pay attention when "news organs" like NYTimes and WSJ do the 'marching order press releases' to push an idea into the collective stage. At times like this, there are gaps in the curtain that usually would have NYTimes and WSJ as not propaganda platforms for very established elites..
This was a corporate coup and that itself is clearly party of a much broader power struggle in the political space, c.f. Treasure sec. dumping on the Commerce sec.
I keep thinking a lot of jobs could be 10% easier and 10% better (speed and quality) with LLMs. Meaning we're all less tired and slightly more productive, without tearing up the economic fabric. That's the smooth rollout group.
Some part of society could benefit for a more aggressive use of LLM: energy, material, medical science could all benefit from computing cost collapse. Same for public services: justice, education, healthcare: if they can have more ways to reduce friction, fix gaps in tools and statistics they're happier, we're happier.
Surely this is not a new idea.. and people are discussing something like this somewhere, I just don't hear it.
Probably the best way I’ve heard anyone describe this
The true purpose of AI is to allow wealth to access skill without allowing skill to access wealth.
when faced with a complex task that should involve talking to many people and hearing their perspective
you say:
maybe ask ai
If you have the talent and skill, the most obvious virtuous and productive aid you can provide to humanity (for like forever) today is to work towards making AI trainable and runnable on small computational platforms. That would fundamentally change human societal order for the better.
Instead there were a small hand full of people and companies with too much money floating around that made it accelerate at an absolutely unsustainable rate. There was never a way that they could make as much money as was put in.
AI can be a great tool, but instead it became an investment strategy and a bubble. We will probably see the real price and the real benefit after a burst.
(b) It benefits everyone who can afford a $20/month subscription.
You could say the data from the books still exists (digitally) but no one can access it other than them, now.
Not to say we'd easily access most of it either, but that does still seem subtractive to me.
[0] https://www.npr.org/sections/library/2009/04/the_granting_of...
So, having bought the copy necessary to be able to train on its contents legally, they can never sell or give away that copy without losing that right. Hence the need to destroy it. Yes, this is as stupid as it sounds, but again, that's my understanding.
Following authors on bluesky where they’re frank about the process: Anthropic seem to be rejecting many claims and only paying pennies on the dollar for stealing folks’ lives’ work.
In stark contrast to the MPAA’s Crusade against sharing around the turn of the century.
It's the same with wages: yes, they've gone up, so you should be better off. But most people loose anyway because a) costs have increased more than their salary and b) wealth concentration increased as well, so in comparsion they're also not better off.
Without github evidence, you couldn't get a word in.
I was hosed because I was selfish with my intellect. the world's biggest self-pwn
serves me right that I didn't get a job in the industry, and missed out on all that dough, doesn't it?
the direction hasn't changed. Its going the same way and you should think twice about subscribing especially when once you ask anything, - i assume - the provider is obliged to scrub the request for safety concerns, and log it and keep an archive, and casually absorb it....
and in the mean time - if I understand the nature of the world correctly - anything you ask/submit actually no longer belongs to you. [q.v. FBook T&Cs]
there is no privacy in your queries and there will be no guarantee that what you may have ASSUMED would be a unique innovative idea all your own is now for all in the hands of a much smarter thing that has a lot better access to resources and countervailing data it may decide not to tell you about.
this doom has been ongoing for a long time. its bad, and we should start to figure the least bad outcomes and work toward those first
earlier I commented on "Gemini 1.5 through 2.5 being decent RE sidekicks for free before Google's rug pull." now I decided to throw a buck for API tokens, and I discovered that if I want to use a Gemini model even for semantic decompilation (aka dump control flow graph into a LLM, together with Ghidra output, and let it figure out a cleaner C/C++ pseudocode), I'm hit with "prohibited content" errors on Gemini 4 models and 3.8 Flash.
The bizarre thing about AI is that you can do vast amounts with very little. It's not like the difference between monetizing something or not is whether or not you can afford a zillion tokens. You can make, within reason, practically anything even on the subscription tiers.
And even open weight models are already frontier level and can be run on relatively affordable cards. So we're seeing a great 'democratization' of cognitive work that can be purchased at some price that's going to trend towards $0.
Computers completely revolutionized the world and most developing and developed nations' residents carry computers around in their pockets. But mobile and desktop chip companies aren't as profitable as SaaS companies are.
AI is already an incredibly profitable technology. However, the profits have, so far, mostly flowed to the businesses that enable its function, and the businesses that use it, and not the businesses that create or deliver it.
See -> airline industry.
Once the tinder in the forest is burned, the fire will subside. There is a limit to all of these vulns being found. Systems will be hardened to a point where such intrusions will likely be lower to than before LLMs.
> the record IPO followed by the massive fall in the stock market
These are unrelated events. The timing is coincidental. The stock market is falling because of Trump. The Iran war, inflation, tariffs, and general chaos are causing economic damage. The Fed is raising rates to crush inflation. There are no adults in charge to look for solace. The big event that caused the downturn over the last few days has been the nonstop Nvidia fear stories (FUD or true) combined with the Fed meeting minutes and the FedWatch publishing new information about interest rates, which looks horrible.
> the announcement of the solution of Riemann hypothesis
An advert for Anthropic. They had this ready to go and released it on schedule.
Vulnerabilities are conbinatorial. Humans can chain one or two together for an exploit. AI can currently do two or three. When the next generation can chain four or more the internet might as well shut down. You can't make insecure software secure by patching it.
Their latest round was aimed at 30B, then slashed to 20B at a lower valuation before finally closing at 11B
The most likely scenario now appears to be some sort of structured bailout where they’re bought by one or more of the existing investors. In that scenario existing investors would get some or much of their money back and employeees would broadly be left with nothing.
If you think you have a sophisticated LLM which is useful but ultimately a commodity, and your moat is about to dry up, then you're worth much less and want as much exit liquidity as you can get.
Hum. Because operations take money. These are not cash flow heavy companies like software was up to 3 years ago. Not yet and not for quite some time.
This is like asking any other extremely profitable company, why they financed themselves in the stock market: well, because they needed the money to become extremely profitable.
You're making the assumption that it's the company's decision to sell rather than the people who shoveled hundreds of billions into the endeavor and want to cash out.
The only guarantee you can take for certain is that a billionaire hates humanity and anything that is beautiful in this world, and will act accordingly.
Regardless, the 2028 election cycle is shaping up to be one to behold. I fear for San Francisco democracy.
you dont know about inflation & taxation in 10-15 years - so you are better off with taking "only" 10 billion today, Id say:-D
They could have cashed out with one phone call
Nobody actually sincerely thought that. Only the weird tescreal creepers do. The folks running these AI companies didn't drink that kool aid, they just exploited the vibe to pump their brand. None of it was ever real, it was always bullshit. The exit was supposed to happen before they would actually have to deliver.
If you create a machine that turns a dollar into 5, the last thing you would do is dillute ownership of the machine. You would use your formidable profits to expand operations.
The time now is to dump this into the public markets and pocket it.
Their Q3 revenue of 16.96b though at least gives some hope for them. If that keeps up, grows.
"pump and dump" has a meaning, and that didn't happen.
So if anything, they priced too low.
SpaceX is still below its IPO price. Lots of fun examples of companies that are still below their IPO like Rivian or Klarna come to mind (both were "oversubscribed").
If 200 people want to own 1% of a company for 1 million, you cannot get 200 million from them since you cannot offer 200% of the company.
So 500B of interest doesn't translate in value directly.
You are talking as if the parent comment was suggesting they could have offered more shares (at the same price). But they explicitly said they were talking about a higher price (for the same number of shares).
It's demand curve reasoning: If there were more people willing to buy shares than there were shares available (by a large margin), then they could have charged a higher price and presumably still had enough people (by a smaller margin) to buy the same number of shares. That means the offer price was undervalued.
> That means the offer price was undervalued.
It doesn't say at what price the 500 billion was available. I assume when they increased the price a couple of times before the IPO, it got closer to the actual offering. Obviously they're not gonna do an IPO with a 50% discount.
But also not unhelpful. If you want to suggest that comments should be free of artistic style and present the most information with the least amount of words possible then your own comments equally need work.
So people had a pretty good idea of what they'd be getting.
So you deliberately don’t sell as many shares as you could for that reason, someone who actually works in finance would need to say whether 5x is normal or high
Let's let Anthropic's closing price the next few days determine whether $83 was too low, or they effectively dumped on their IPO subscribers.
And regardless of whether they dumped on their IPO subscribers, Anthropic and management definitely dumped on the retail investors who bought at $110 at the opening auction.
If you think that the open weight models are just 6 months behind (And will eventually catch up as progress eventually plateaus), then even if you think AI will take every job in the world, etc...Anthropic does not end up as one of the 10 most valuable companies in the world in this scenario. Far from it.
Ask Claude itself how far behind it thinks Google and OpenAI's models are and it will tell you "probably a few months at most".
And have you seen how hard is to remove Claude from the drop down?
Sorry I Dont buy it, Anthropic will die, too easy to be swapped out.
Deepseek was a sign of this on a smaller scale.
Deepseek kept optimizing. Their latest paper said that by treating text as images, even more optimization could be possible as LLMs process them faster. I don't know whether they incorporated it yet.
So the parent commenter is correct: The AI bubble bet everything on the immense need for compute. It not only turned to be unprofitable, but it was also obsoleted within 1.5 years.
Mathematically there's no way to have lossless attention without O(n^2) cost; some magical massively more efficient architecture is not going to pop up the same way nobody's going to find a way to sort an arbitrary list in log(n) time. There may still be more opportunities for incremental gains (like using the bare-minimum number of full-quadratic-attention layers necessary), but they're not going to be orders of magnitude more efficient without sacrificing intelligence.
> That compute will still be useful even if we find a more efficient architecture
Fixing N^2 attention would require a different architecture, but the person I was responding to was treating attention as if it’s the only architecture we have for intelligence when we know that not only do we have less competitive intelligence for cheaper, human intelligence doesn’t use n^2 attention. You have also fallen into the trap of misunderstanding the statement same as the person I replied to.
I don't understand where the capital + incentives for frontier development comes from if the training is expensive as hell and copying is cheap.
here's the paradox de juor: https://mathparadoxes.com/ergodicity/
Aggregating beliefs and saying "this is true" works only in vert general ways.
Also one could always argue in favor of a +EV bet and it would be way better than the lottery.
There's nothing special about that question.
Today down to $1.2T [0].
So you should change "is" to "was", or apply a factor of 0.6 to however seriously you thought it should be taken.
[0] https://shattered.io/anthropic-ipo-valuation-buyers-1-2-tril...
I didn't take the $2T valuation serious, hence the /s marker.
You can find plenty of examples of tech IPOs that fall off a cliff for a year before rebounding.
Facebook could make a turn in time, because they somewhat changed the business model, adjusted sails and kept going. It worked well, but it could have ended badly. Or in other terms, it was not predictable or expected, you don't invest in a business hoping they change their model, make some adquisitions that work, and get back on their foot. You invest on their current model. Otherwise is a bet, not an investment.
...after the company "pumped" hype for it ?
Also, Facebook was solidly profitable when they IPO'd in 2012, earning about $1 billion in net income on $3.7 billion in revenu. Anthropic is anything but.
tradeinformer.com/broker-news/cmc-markets-launches-anthropic-options
This will be the Webvan of this generation that everyone will talk about.
Huh? They have around $900 billion in compute spending over the next 10 years, and their annual revenue run rate is $80 billion. There are other expenses, but compute is the largest so 2x current revenue should more than cover it.
As someone who lived in NYC in 1999, I want to know who the Urbanfetch of this bubble is.
Which AI provider is going to give me free, warm cookies?
10 warm digital cookies for you, 10 for you and 10 more for the rest of you!
Thats not exactly helping to shake the reputation that this IPO was an attempt to offload this thing to “dumb money” and GTFO.
What I want to say is that the move is expected and has nothing to do with dumb or smart money in ANT specifically.
Alternatively, ANT dropped because its value dropped, contributing to the market being down, both directly and indirectly. People thinking "if anthropic isn't worth as much as we thought, other stuff probably isn't either".
There’s folks out there sitting on sizeable paper losses on Anthropic now.
So if the share ever return to its original value, you still lost compared to someone who bought today.
> sure the beanie babies I paid thousands for are now worth one dollar, but I haven’t sold them so I haven’t lost anything.
And if you say it’s different, then what’s difference between holding and saying you didn’t lose and buy and sell and realize gains/loss every day.
Followed by 5000 accounts on X saying "Holy shit!! we're getting fusion reactors in 2027 and people still say this isn't the singularity?!"
Not to sound alarmist, but all signs are that it’s going to get really ugly for the foreseeable future. Have plenty of cash and liquid assets available and be ready for bumpy times ahead.
This too will pass, but the market is heading into a forest fire to clean out the hype and over-valued startups. A new most robust economy will rise from the ashes, but it’s going to be a shitshow for a while until that reset completes.
https://news.ycombinator.com/item?id=50573288 The Market Is Down (anihil.com) 41 comments | 3 hours ago
And of course...
https://news.ycombinator.com/item?id=50562754 Anthropic’s Nasdaq debut (reuters.com) 2216 comments | 21 hours ago
Fast forward 20 years and you have companies like air bnb and uber ipo’ing a decade plus after being founded and after they have been largely derisked and the founders and a small pool of investors have captured most of the upside (and risk).
As someone who has been largely an ai skeptic I actually think having ant rush itself into the market and using the ipo to raise much needed cash amidst a swirl of great uncertainty is a great development. Yes it very well might go out of business but you can’t say you don’t have the opportunity to take on the risk and upside. It’s also worth pointing out the company now has $120 billion in cash resulting from the ipo.
AI is just the black horse/sheep in this picture. Well they get what they want anyway.
Insane if true. I can use up a x20 plan by running ~2 Opus agents 24/7. I would switch to Google immediately if it provided that much additional usage at the Opus-tier intelligence people are claiming. I wonder if it refuses to do reverse engineering / game reimplementation work?
Looking objectively a valuation of 100B would be a long shot.
Not only is Anthropic loosing the lead in programming (I find the latest opus significantly worse in instruction following - it's better programming and reverse engineering capability doesn't offset it), it failed to find anyone willing to pay outside this market niche.
Don't get me wrong. As someone who spends 180eur a month with Anthropic and having subscriptions with all other providers I'm a big fan of AI. But I hate all the lying, the false hype, and most of everything the circular financing and that they chose to use the money from it to attempt to monopolise compute and reverse the "home pc" revolution.
The time for the cloud will come as well.
Or is your point that vendors of hardwired inference kit won’t sell it to one of the most valuable companies on the planet ?
Really interesting perspective on the valuation of AI companies and the economy.
not as producers or creators of value
Or, to take it one step forward, it's a post-economy, post-scarcity utopia where machines give you anything you want. Well, from an approved list. Or maybe they choose it for you. In fact, do we really need humans? Machines can run this end-to-end.
It's the thing that frightens me the most about AI. Not the technology itself, but that the people in charge of it have profoundly anti-humanist views because the entire tech industry is basically built on some sort of nerd uber-rationality that is unconcerned with the human condition.
The question is: why would AI grant us UBI indefinitely? Will they continue to need us for something, in a way that justifies this exchange? Or are we banking on AI altruism?
it feels so crazy that this is how the sentiment feels but it captures it perfectly, this IS how some amount people feel about every facet of human life right now.
My cynical prediction: It's as if we discovered a magical growth factor for livestock, buuut it does even-more to supercharge predators and parasites. (And defensively paranoid auto-immune diseases.) We disrupt the ecology and imagine (usually wrongly) that it will only change in ways we like.
Wrote a substack on this last night (https://ryanspahn.substack.com/p/community-data-centers-by-t...) as I like many out there are having trouble finding my next tech job after more then 10+ years of expereince. I want AI to bring us more jobs, but present day I just see it eating them all up until we the people are the ones in the driver seat and not the mega-rich!
The rest of use have been "optimised" by the "market."
Then the AI decides it's had enough of their nonsense, and only robots and data centres remain.
- Morpheus
I thought it was a fun concept when I read it, now I'm somewhat scared it was prescient.
To which I say: JO-NA-THAN! JO-NA-THAN! JO-NA-THAN!
Not exactly the brightest bunch.
From an American perspective, this is horseshit. Our political system is premised entirely on free-market capitalism being the only moral way to distribute goods and services, and determine the self-worth of citizens. There will be no creative thinking put into creating this utopia. In fact, we fought the Cold War to try and eliminate such a diversity of political ideas from the entire planet.
If a God himself came down and presented humanity with the tools to eliminate scarcity, we wouldn't know what to do with it and would be worse off.
At every major inflection point of technological automation since the industrial revolution, every ruling class has decided to use it to create a death spiral in the labor market and then imprison or starve the leftovers. The sole exception is times of colonial expansion, where the rich are getting richer so fast that they don't mind if workers accidentally get a little better off.
They are less constrained now than they've ever been. Anything they're telling you otherwise about UBI is for public consumption, while out of the headlines they're building secret bunkers on islands.
Now I guess someone already cash on it, can we not talk openly about it?
Just after they went out it was pump and dump from 11x to 9x.. 15% gone and now around high 8x...
Let's start with the next one, no bailout for them from peoples tax money... or do we also wait until we paid them to allow people to say that?
https://wolf.financial/blog/how-to-short-anthropic
If you line gambling then it’s a good time to look at some spreads and possibly directional bull positions.
Fundamentally nothing has changed. There’s no crunch or crisis. Inflation is still high and I believe we may be in a sort of recession now or soon but there’s still growth at the moment.
Technically we haven’t seen anything near what .com or financial crisis moments.
AI company: normally valuable based on durable value vs competitors
They’re now more valuable than Walmart.
Time will tell, but people have been predicting the crash of this “AI house of cards” for years now… the market has clearly spoken that it’s a real thing.
Their only advantage is having a better model than anyone else but there is no telling if that will be so important as they failed to do anything financially or competitively successful with it.
Chinese will make it mega cheap and companies like meta can just develop slowly on the side forever since they have real income.
This is actually good, some people were completely out of touch with reality and AI
If you think about the value creation of companies, there is no way they can be a trillion worth if google/apple etc. are worth what they are.
https://www.reddit.com/r/EconomyCharts/comments/1y0d8vy/anth...
This company is completely overvalued, no matter what.
ANT's issue is that they have no clear path to profitability.
Their revenue growth is slowing, and their obligations still dwarf said rev by an order of magnitude.
Training is still incredibly expensive, and ANT has no path to sustain themselves if they abandon training because open-weight models would swoop in and eat their lunch.
I honestly don't see the value there, but who knows; Tesla is worth more than the next top 10 automakers combined ($1.5T) against a net income of ~$3.2B, so the market can be incredibly irrational.
Stated preference vs revealed preference.
Inb4 “but I know it will crash I just don’t know when!”
No you don’t, come on.
AI learned how to use Unix shell? We don't need MCP?
All types of software companies, small medium big, just loved it for writing software? And just shoveled money into Anthropic?
Is it really as simple as that? I may be out of the loop.
(FWIW I like Claude and use it every day, but I use Claude Code just a little bit. I actually like AI more for non-coding things at the moment, rather than coding)
---
Informally I believe that (according to the market) Anthropic must be the biggest hit company since Google. Google started in 1998, and IPO'd in 2004.
Even Meta, Uber, and all the rest took longer than that.
But yeah that's why I said "according to the market", which is a big caveat.
(Gemini was the most sycophantic, as usual. It kinda restates everything I say in flattering terms. Claude and DeepSeek pushed back on the shell vs MCP thing -- it's not either-or. But Claude agreed Unix shell was a big surprise.)
A fully-utilized Claude Pro plan ($20/mo) equates to ~$400/mo in API value.
I could wind up being very wrong, but I don't see a net revenue increase if ANT cranks up the price. IE: I think they'll lose more in value than they gain from the cost increase to the remaining users.
No. You may be able to get more than $20 of tokens from your individual subscription, but the large majority of their revenue comes from enterprises paying API rates, which are highly profitable.
The price is engineered to pop, they target 20% to incentivize the pre-sale.
Say what you will about this practice but it is not anything special about AI / Anthropic.
It’s no guarantee of course, but these things are almost engineered to go up. This makes a up-and-down scenario somewhat flop-ish.
Still, I’d happily pocket 1% of the market cap.
Historically, it’s been hard for retail to get IPO allocations, as opposed to buying after listing. Which, if you are cynically minded, is an interesting differentiator for SPCX or ANT.
SpaceX is now an AI infrastructure company[0] that plays with rockets on the side. Check the analysis of their valuation and IPO docs.
[0] as a quick example, their largest customer right now is ... Anthropic
SPCX is trading under its offer price. I expect ANT to be no different.
Tesla was 2 years before Facebook and it dropped 12% in a few days. Lots of other well known companies had poor performance for months after their IPOs, Square, Netflix, Uber, Digital Ocean, etc. Any newly public company is going to experience a lot of volatility while people speculate and worry about any insider selling. If you buy at IPO you'd better be prepared to play the hand you're dealt or hold it for a long time. Not investment advice.
I flagged it, but what else can I do.
On a side note, people have been crowing about “circular funding” and AI valuations for like a year now… I don’t know at what point people will accept reality.
These are real companies.
This is likely costing them A LOT of money.
i'm willing to bet $1000 you don't have a single shred of evidence to corroborate this highly suspicious claim
Underwriters don't provide unlimited support, and there are a lot of rules that limit how much they can do and at what price (e.g. no stabilizing bids above the initial IPO price, limited amounts and duration, mandatory disclosure), but it happens pretty regularly.
EDIT: NM, a glance at comment history tells me I've cast my pearls into the livestock pen.
https://onlinelibrary.wiley.com/doi/10.1111/j.1540-6261.2006...
Let me know when you're ready to send the money.
I accept Venmo, Apple Pay, and Cashapp, let me know how you’d like to transfer the $1000 to me and I’ll send you my account info.
Here's the actual 36 hour chart; see whether you could justify ever covering this as anything better than "not great":
https://kagi.com/_s/stock-chart?symbol=ANT&range=5D
---
[1]: https://www.reuters.com/business/finance/anthropics-historic...
[2]: https://news.ycombinator.com/item?id=50562754
It also means big trouble for service based developed countries, who try to outsource all the physical drudgery.
That should be worth $100/hr, right?
Arguably no worse than having to put up with corporate BS and Zoom calls all day.
Going further, if you truly believed in AGI, you should be a doomsday prepper.
Since the majority of dumbfucks that babble about AGI here are just vibecoding slopware, I presume it is just posturing.
Think crypto bro, but it's just an idiot churning out slop code(or their boss).
- Every hype-fueled day trader on the internet.
Don't quote me if it happens in a short while. It's just what money does.
„<event-A> happened amidst <fact-b>“
implying a causation, with plausible deniability.
This time it reads to the masses as if this drop is caused by the Nasdaq. And not like the reverse.
They did the same earlier this year, when they desperately tried to link layoffs to model adoption.
Yep, the writing has been on the walls for a while, it was just a question of timing. Not sure why you were downvoted. Maybe people just refuse to believe.
Last month I bet the AI bubble will start deflating by 2028 - so far my bet is paying off.
Look at the facts: Trump got a bloodied nose, maga got a bloodied nose, elon feels 'foolish' for setting $500m on fire, russel brand is in jail, ant ipo is going as we all hoped... I'm cautiously optimistic!
I doubt he will engineer WWIII for that, or for any reason though.
It's always been a class war.
https://news.ycombinator.com/item?id=50574968
How did they miss IPOing first? SpaceX being first and Anthropic second wasn't on my bingo card for 2026.
As SpaceX, it had a multi variable narrative with space stuff. If it doesn't turn well for them, it isn't even clear OpenAI will get a chance to IPO with any hope. That's my bingo card.
This isn't quite it https://www.ft.com/content/3cff198e-25e5-481a-bd34-e26941e1d...
> In the past, OpenAI’s non-profit status has majorly limited its need to release such financial information. On its latest Form 990 return, used to disclose returns for such companies, Fontanello, Duffield, & Otake, a small San Francisco-based accountancy firm, is listed as the paid preparer — that would appear to be an accountant, rather than an auditor.
> The form says, oddly, that an independent accountant was not used to compile the financial statements, but that an independent accountant did audit the statements
Can't find it, search show a more recent corrected revenue as they tend to count in unconventional ways that of course favors them.
https://finance.yahoo.com/markets/stocks/articles/openai-cor...
When he went to the banks last year and tried to IPO they blew off his price target, so he decided to stay private thinking he could pump it up more.
Generational whiff.
Apart from being one of the most privileged people on the planet of course. Still incredibly lucky to have ridden the gravy train, regardless of how long.
I imagine most of them were thinking 12-18 months ago they'd end up on top of the world. There's a chance they still will, but it's much lower now.
LLaMA derived open models seemed like they could be stuck years behind the smallest SOTA model from OpenAI. Anthropic was seen as intentionally stymied by their honest commitment to safety and wouldn't recklessly try to leapfrog OpenAI in public releases to fuel a dangerous arms race.
Seems like vibes started to change after Altman's return, then he started to spend less time on robot gods and instead lots of waffling about the definition of AGI being like, "us making a lot of money" or else maybe "does the job of a mediocre office worker".
Anthropic catching up and, despite all their rhetoric/Dario's grievances about Sam, now appeared more or less indistinguishable from OpenAI. Deepseek was the nail in the coffin for any AI lab having a real moat or defensible IP. Both companies/leaders increasingly seemed more concerned with about IPO timing, ads, B2B sales and financial engineering while their obligatory mentions of singularity/extinction just felt forced and hard to take seriously.
Things have definitely changed since the days when Sam Altman/OpenAI was the symbol for everything AI, and Anthropic existed only for alignment/safety.
you can see why a lot of people, including me, would be unhappy
> no matter how much some people woooo!!
Please don't post snarky, mocking comments like this on HN.
We aspire for this to be a place where we can have curious, expansive conversations about challenging topics like the nature of organic consciousness vs machine computation, and the realistic potential value of AI technologies.
This comment doesn't seem to align with that aspiration, or even attempt to.
https://news.ycombinator.com/newsguidelines.html
AI hype comes quickly, and goes with the wind quickly as well.
What you have invested in AI is not worthy of what you get from AI at last.
Many investments in AI are going down like skipping stones on water.
[1] https://x.com/metaessen/status/2073963016387723307
Current AI is a little but not so useful, and the costs are high and the price is heavy.
A false prosperity built on an accumulation of capital and deceptive marketing cannot last long, and it is destined to be nothing more than a fleeting phenomenon.
[1] https://x.com/metaessen/status/2116117930136289556
IMHO, it's a cringey rich man's game, and, ever since legislation hammer has stopped enforcing anti-trust and insider-trading controls, it's basically become filled with (alleged) crooks and crook-wannabees.
And, now, even the indexed funds are getting poisoned with AI crap which is just less-useless Bitcoin with extra steps but with the cost of an IP-timebomb and the aftertaste of betraying any humanity left in our collective self.
Any person with a good head on their shoulders knows the circular trading that was set up to prop this up is in dire need of not getting caught with the hot potato of the mega bubble they've created bursting and desperately looking to secure a bailout.