Anthropic didn't lose $80B. Some complex securities tied to their stock went up in value by $72B, which they have to count as a loss, even though it's not cash out the door.
As far as Google's mark to market profit on Anthropic shares, everyone realizes that's not a sustainable source of income. Which is probably one reason why it's trading at about 15 times profits. If you back out the $99B of (unrealized) equity gains in anthropic and others, Google is trading at 25 times the last 12m earnings.
People should really look at cash flow statements instead of just earnings. Anthropic published one and their cash flow was positive for the first half of this year.
That doesn't mean all the circular financing won't be a problem, but people keep drawing the wrong conclusions from the negative earnings, which are relevant to investors but not so much to the viability of the business.
Wasn’t that also influenced by their discounted compute with SpaceX? I was under the impression the cash flow positive news was highly dependent on a lot of limited-time discounting that wouldn’t last.
What I don't understand about the circular financing is how does it not violate antitrust? All these supposed rivals own huge stakes in each other. Are they really competitors?
Keiretsu - perfectly acceptable business relationship preferences that in no way require governments to regulate their GDP growth / campaign donation black box machine.
Strange how you complain about this for circular flows when of course the GOAT of circular flows is the money system itself
govt prints money (FED) -> government -> banks -> (economy) -> tax -> government
And the second "great circle"
govt loans money (FED) (money that doesn't exist, which isn't a problem if you can print money) -> banks -> (economy) -> loan repayment -> money gets "unprinted"/destroyed
Everything about the economy is a great circle. You will also understand after a while, if you wish for the economy to remain stable, this must be the case. You can't change it without either hyperinflation and extreme "capital flight" (between quotes because of course in that circumstances capital will run into gold or other physical resources), by either vastly overheating or stopping the economy.
And yes, most instances are of the overheating kind. The only case of stopping the economy was accidental (despite many warnings), in other words, yes, it was Republicans. But last time that happened, even republicans could decide to change course. Of course, only once huge damage was done.
Way more people complain about this too, if you pay attention to all the US inflation and national debt anxieties. It’s clearly unsustainable and driving interest rates up steadily.
People complain about the imbalance between economy supply and economy demand leading to inflation and debt concerns.
The complaint isn’t that the economy is “circular”, how else could it work with 3 or more people engaged in commerce? Neither of those things get fixed on different ways to manage the money supply either, it’s not as if they didn’t have national debts or inflation while on the gold standard.
It’s just a different format. Providing liquidity in the hope that more is generated from the economy. Except it’s doing more harm than good other than in true crisis. Other times it’s just creating a bubble, in national debt and in AI.
The actual control is with the "Board of Governors" which is a government agency (ie. the president controls them, subject to approval by the US senate). The banks structure is to make banks cooperate, it does not give them any kind of control over anything.
That is from the increase in the share price, because they owe that in shares and the share price is higher now. On the other side is the IPO, it balances the books but it doesn't count as an asset today, whereas the promised shares count as a debt today.
people should look at dividends rather than anything else; but the people who matter at these scales are just dick measuring and don't care about anything to do with the actual economics of the system.
They're not really "complex securities"; they're just convertible bonds. The lender gets to choose if they want to be paid back (1) in a fixed note value (like a bond), or (2) with shares of the company (like equity)—if it IPO's. (I.e., the value is the max() of the two. ReLU-shaped).
I don't understand the accounting aspects, and they seemed very weird to me, when I tried to read about was happening with Anthropic. Basically: the rules says the convertible debt has to be valued at the predicted IPO price—because that's the value of the thing (Anthropic stock) Anthropic will be obligated to pay. At the point they re-calculated this debt, the valuation went up drastically, so the rules say that counts as a very large quarterly "loss". But: on the other side of the scale, the increase in predicted IPO price does not count as a quarterly "profit"—even though they should cancel out. Even though the increase in share price exactly cancels the size of the debt they're paying back—the debt denominated as a fixed number of shares.
Well they can’t have it both ways on the accounting front.
These complex securities they took a mark on only exist as a way for them to access compute at deep deep discount by issuing suppliers lottery tickets.
Post IPO when they don’t have lottery tickets to issue as easily the numbers should be interesting.
This is correct. Out of the $80.7B loss, $72.4B was due to changes in the fair value of their convertible notes (and similar financial arrangements).
Going a bit meta, I'm curious why we don't observe a faster convergence of opinions here. I assume that HN users are well-versed in AI tech, and the current generation of LLMs will happily break down the relevant information in Form S-1. Nevertheless, we see a chaotic storm of opinions, with wildly varying S/N ratios.
Is it a limit of current LLMs, or is the bottleneck perhaps human nature?
Yes but those convertible notes only exist to grant them access to compute they would otherwise need to pay fair market value for.
If I drive a taxi making $200/day and need $150/day of gas.. but convince Shell to take a convertible note instead of $150.. Do I make a) $200/day profit, b) $50/day profit, or c) $200/day - (change_in_value[convertible_note]) in profit?
Some are arguing that C is somehow fake, but its obviously more real than A!
Human nature. Obviously we know some things are just true. It's common sense! Why would I invite a non-deterministic hallucination machine that can't think to challenge my world view with something inconvenient like facts?
People are emotionally invested in a deflation and slowdown in AI because they feel threatened by AI. They'll cling to reeds of hope and wilfully misread financial stats to make themselves feel better.
The tech blowout had little to do with Anthropic's filing.
The numbers from multiple companies show a high CapEx investment in a product that has yet to be proven as profitable as the very things it is hurting (i.e. ads). End of story.
Didnt it start with Korean stocks first? It feels like a wave, it hits there first then the US tech stocks. Korea index has stabilized for a few days now, but hasn't recovered what it lost.
It doesn’t help Google to create an AI-powered business which turns $1B in profit annually if it results in them losing a search/ads business which turns $5B in profit annually.
The CapEx spend for AI not only makes AI itself more expensive (and thus less profitable), it’s now driving everything else in technology to be more expensive (and thus less profitable).
AI is a profitability tax on the entire tech sector and it remains unclear if the profit AI will deliver is going to be worth that.
I would say they should slow down on build-out to balance supplier/labor availability with available cash flows, but I suspect all the Big Tech companies are worried about falling behind the other and being put in the position to need to rent out compute.
It’s more or less certain that Anthropic is already profitable if you discount the losses due to becoming too valuable.
What you are saying here
> The CapEx spend for AI not only makes AI itself more expensive (and thus less profitable),
There’s nothing inherently unique about this. It’s just supply and demand - the only problem is that the ecosystem of raw materials can’t keep up with the demand. It will catch up soon and the prices will stabilise.
Your only problem seems to be this aspect only: that RAM production hasn’t increased according to demand.
No I don’t agree they should slow down. I think RAM and other things should rather speed up to react appropriately to the demand.
It’s like saying the iPhone is not profitable because too many people want it and its price is going up.
> Last year, for instance, Anthropic lost $80 billion, and Amazon had negative $42 billion free cash flow. The entire big tech complex, it seems, has moved from minting cash to burning it. And investors are, well, not panicking exactly, but wondering if the bubble is ending
Is this really the quality of such newsletters? I’m not asking rhetorically because I’d like to also have trustable sources to understand the underlying reality. I actually spent time understanding this part and the 80B dollar loss thing is completely misrepresented here (also went viral in other media).
"The shift in politics is even starker. In the 2010s and early 2020s, big tech was well-liked by most people. But AI is absolutely loathed by the public. These companies have never faced real public dissatisfaction, they all imagine themselves to be the good guys, with the credibility of having created magical search engines and smartphones. But that era is gone. They are now akin to the bankers on Wall Street in 2008, and politicians are winning elections by saying they hate AI as much as the voters do."
Yeah I see this too. That we're not innovating any more also doesn't help make our industry look good
You can't claim that the tech industry isn't innovating. You can say that innovation effort has shifted away from many things that still need it. The issue I have is that the companies and products are so large and complex that they are extremely difficult to compete with.
I'm in agreement with the general point but was big tech well liked in the early 2020s?
Facebook scandals had exploded already, Google had already long abandoned "don't be evil", nobody ever liked Microsoft, and I think sentiment towards Apple is somewhat unchanged.
Tech at least in my eyes peaked around 2018-19. That also coincided with the first more mainstream public-aware introduction of vision and NLP systems into tech products. A few years later it all went downhill for the tech companies.
was that change organic, though? every single shitrag went berserk about AI once the so-called journalists had realized they were about to lose their jobs to it.
I think there is a massive selection criteria bias, AI has been a batshit insane productivity boost for me but I’ve been so busy with it that I’ve generally stopped posting online - I only do so now because I’m out of quota. I no longer see any point in arguing with people - I’ve got better things to do, like argue with AI. I used to scour the Earth looking for smart people to talk to and now I don’t even have to leave my apartment.
In 2014 I moved from London to SF to start a big tech job. In my first day there I accidentally wondered into a massive rally, I asked one of the participants what it was about and they said they were protesting all these foreign tech workers coming to SF. It felt a bit weird to end up in the middle of an ‘anti-me’ rally - especially one so passionate and energetic. A few protesters pressed me on where I was from and what I do, I lied and told them I was an investment banker. They told me that they suppose that was ok. I specifically told them that because not long ago it was finance being the target of protests and I curious to see if they had somehow moved on - if this was a principled stance or if it was being astroturfed.
These protests ended up getting pretty violent, people were throwing rocks through bus windows, throwing up on cars.
Don’t worry I didn’t stay long. I loved the Bay Area but did not like SF all that much.
I believe some of the busses were occupied when they were attacked and the lack of injuries was largely due to good fortune. I could believe a claim that this was done by false flag agitators intent on undermining the protests.
First, ChatGPT has been the most downloaded app for two years straight, the idea that AI is loathed by the general public is pure delusion, propped up by manipulated surveys and people living in a bluesky bubble.
Second, Big tech wasn't well liked by the early 2020s, that ended in the early 2010s when the NYT editorial board took a top down decision to cover big tech negatively.
Third, the idea that they were innovating before but aren't now is insane. AI is the biggest innovation in tech since the transistor and possibly the only true innovation since the GUI.
Respectfully, download numbers are not an adequate proxy for opinion. The number of companies mandating AI use plays in significantly. There are highly ploarized bubbles of opinion fwiw. In the end, it requires actual controlled polling of opinion to have a satisfactory understanding.
Additionally, I'm not sure how you square NYT coverage being more critical of generally worse behavior as "top down" without a citation, care to point at anything to substantiate that statement?
Innovative, in some part sure. But lately the running jokes have become "Uber, but for $X", which hints at imaginative stagnation. AI has been the largest shake up in a decade (or more), but it is hard for Average Joe to grasp what is changing.
People are being led around by the nose. Cui bono? 2000 years and most people still can't manage to ask that question reflexively. AI hate is a parochial narrative being pushed by people who stand to lose influence. Examined objectively, there are reasons to be cautious when adopting AI tools but a knee-jerk rejection of everything AI is a purely irrational response that has been drilled into people through repetition, conformity and media influence.
Tech companies were never the "good guys" any more than they are the "bad guys" now. Those are just colored filters thrown in front of the lime light to shift your perception. As soon as tech or AI works to their benefit then suddenly they'll be the good guys again and no one will care they were the bad guys before. People just go where the algo leads them, believe what their influencers believe, think what their peers think.
People hate AI because culture is being turned into slop in front of our very eyes. There's clear evidence that people collectively reject:
1. AI games
2. AI Art
3. AI Music
4. AI literature in most forms, except for factual text like search
Or at least, they won't pay money for it
AI hatred isn't being pushed by people with a narrative. I can't use AI for my game, because it'll sell £0 if I do. That's the net result of millions of other people telling me that they despise AI and won't purchase my product if it contains any AI, and I'm just making a business decision
Also google and some other big tech companies were widely loved for a while. Microsoft/gates managed to entirely rehabilitate their image at one point, and google was seen as one of the good guys. Those days died with the cambridge analytica scandal, and there's no evidence its coming back
And why do those millions of other people have that opinion? Have they arrived at it of their own accord?
Yes, you can create garbage with AI tools. You can create garbage using Blender, Unreal Engine, Unity, GIMP, Ableton, iMovie and Word as well and there are mountains of garbage created with those tools that nobody pays for or likes. If you have some talent and put the same effort into creating art with AI that you do with a digital paint program then it won't be garbage. AI isn't a magic wand that you can wave around and get perfect results, it can save you time just like using an image editor saves you a lot of time compared to drawing and painting by hand but it won't do all of the work for you. It's a tool and a poor craftsman that blames his tools for his failure.
My point exactly that big tech was loved for a while and then hated and then loved again and hated again. Have they truly changed that much that often or is it your perception that has changed?
> Yes, you can create garbage with AI tools. You can create garbage using Blender, Unreal Engine, Unity, GIMP, Ableton, iMovie and Word
Do you not see any difference between creating garbage in these ways?
People have created low-quality things since the dawn of humanity, using all sorts of tools, professional and not. However, before AI, making most types of garbage still required some human work, it wasn't infinitely scalable. And what was automatable ended up being so horrible that most people could easily dismiss it on the spot, or for them to even be filtered out algorithmically. On the internet, it wasn't perfect, but the signal-to-noise ratio was still at a level where you reliably discern the signal part.
AI changed this. Even a beginner can set one up to run almost fully unattended - and while what comes out won't be good, it also won't be immediately apparent as such, especially to the general public. More and more I find myself being fooled by a website link, a YouTube video thumbnail or a paragraph in an article. AI lets people create garbage that can impersonate real effort decently well, and they've used that to great extents. As we speak, thousands of agents are working day and night to pump every online platform with their 'content', without a single person in the world to care about what they put out. They're taking over and removing any good signal we had, everything becomes diluted with trash on a scale we have never seen. People notice that. Many are clueless about AI, and social influence plays a role, but they're not being misled and have their opinions manipulated. They wouldn't enjoy most of this content regardless.
In addition to the dilution of good information, they care about appearances and how AI lets you create an appearance of caring about something. The cheap food stand seemed honest when they put up an unassuming print of a simple Word document, and now seems dishonest when they put up an AI-generated mess that looks nothing like the real product and desperately tries to seem more refined than they are. Their manager was at least honest when he replied with a lazy 'k', and now he's lying about caring by dumping out a bullet point list of AI-generated text for a response that we know he himself hasn't read. It's the dream product for scammers, grifters and people who simply don't care.
Can you use AI to help create something good? Sure, but anyone doing that still needs to take their time and put care into it. Even if AI makes them faster, by the time they finish, 10TB of more slop has already been produced to drown them out, because that can be fully automated. Effectively 99.9% of AI content ends up being garbage, so of course people are going to reject it. Most probably don't want the few nuggets of gold for the ocean of waste.
People hate these outputs, as well as the wholesale enclosure of their commons (training on public data ignoring their copyrights and the social contracts that enabled them). Additionally, despite how it might appeal to a subset of humanity, "we're going to take all of the jobs" is not a fantastic sales pitch.
How can the output both be so much worse than what a human could produce while simultaneously taking all the jobs and displacing the human made data it was trained on?
If everyone hates AI output and refuses to buy anything made with it then in what way is it a threat to people who don't use AI in their work?
Are these talking points perhaps crafted to benefit a small group of people who will lose the edge they currently have to a larger number of people enabled by these new tools?
Not really an accurate view, and the conspiracy angle isn't helping.
The output could be fine in itself, except people (arbitrarily, some could say) value the human production. Displacement is simply a matter of volume: the sleepless machine can churn indefinitely while people can not. The thing is too, most outputs being mass produced are of a substantially lower quality than people (and what is possible).
A problem is people are being tricked into it, or being given little choice. If 90% of companies use AI (as in, mandated from leadership), what "choice" is there? If people recognize the 10% remaining and do business with them exclusively, economics will take over and force either price hikes or the 10% to grow. It's playing chicken.
Based on what the trends seem to be, people will unknowingly keep taking in products with AI (mostly out of not understanding the choice).
As for the enclosure point which is important: People make things for their enjoyment. When people try to profit off their work, they rightfully get put off. The deal is "if you can view this for free, it isn't meant to be sold. If it is sold, I deserve to have my time compensated". AI training/scraping has violated this agreement, and funnels the money to a third party to the people who made the valuable originating art. If it was not valuable, then why not train AI on its own output? Since that appears to be a "problem", there must be something about what humans create that has value AI cannot produce itself.
>Are these talking points perhaps crafted to benefit a small group of people who will lose the edge they currently have to a larger number of people enabled by these new tools?
That flavour of conspiratorial thinking is one of the reasons for polarization. Sometimes, people prefer the human touch. This applies mostly to art and other topics lumped into "the Humanities".
Hope that clears things up for you, or at least sheds light on why people might not see things in the way you do.
> people (arbitrarily, some could say) value the human production
This is a different argument, going into it would muddy the waters. It would devolve into a debate about how much and what type of input and individual perception. It's also been rehashed a million times since Duchamp thought to put a mass produced urinal on display as a fountain.
> most outputs being mass produced are of a substantially lower quality than people (and what is possible)
Different purposes and mass production does not eliminate the market for artisanal work, it replaces only the limited batch work done by hand previously.
> A problem is people are being tricked into it...
God help anyone who learns how the sausage is made.
> "if you can view this for free, it isn't meant to be sold. If it is sold, I deserve to have my time compensated"
This is mostly a matter for the courts but as long as the works are not being reproduced in whole unattributed or in violation of whatever license they held then it will mostly be allowed. If you look for it, you'll notice that artists "borrow" from each other constantly with no attribution and no payment to the originator. All but the grossest plagiarism goes entirely unpunished because everyone realizes that no one can stand alone. Is it different now that it is being made available as a service? Yes, there are now more legal ramifications due to an audit trail of scraping and purchases combined with the much larger output that can be measured and compared. Does it destroy the commons? No, people will use the outputs in their own work which will be published and become part of the commons just like any new style or media.
> that appears to be a "problem"
It's a solved problem. Synthetic training data is a valuable addition now. It requires a way to validate or rank it but that's not any worse than cleaning and filtering human data.
> conspiratorial thinking
An excellent smokescreen but we've been living in a soup of mass media propaganda for over a hundred years now so there's no denying that influence campaigns exist. Denying that humans act in coordinated ways to influence the opinion of others is absurd when you can get a degree in it, have it as a job title and build a trillion dollar industry on it.
>This is a different argument, going into it would muddy the waters. It would devolve into a debate about how much and what type of input and individual perception. It's also been rehashed a million times since Duchamp thought to put a mass produced urinal on display as a fountain.
On the contrary, it demonstrates an unwillingness to grapple with inconvenient problems, as does the rest of your reply.
>'s a solved problem. Synthetic training data is a valuable addition now. It requires a way to validate or rank it but that's not any worse than cleaning and filtering human data.
Untrue for artistic outputs oddly enough.
>build a trillion dollar industry on it.
Or selectively quote in the comment section to dullen the need to address issues.
Anthropic didn't lose $80B. Some complex securities tied to their stock went up in value by $72B, which they have to count as a loss, even though it's not cash out the door.
As far as Google's mark to market profit on Anthropic shares, everyone realizes that's not a sustainable source of income. Which is probably one reason why it's trading at about 15 times profits. If you back out the $99B of (unrealized) equity gains in anthropic and others, Google is trading at 25 times the last 12m earnings.
People should really look at cash flow statements instead of just earnings. Anthropic published one and their cash flow was positive for the first half of this year.
That doesn't mean all the circular financing won't be a problem, but people keep drawing the wrong conclusions from the negative earnings, which are relevant to investors but not so much to the viability of the business.
Wasn’t that also influenced by their discounted compute with SpaceX? I was under the impression the cash flow positive news was highly dependent on a lot of limited-time discounting that wouldn’t last.
What I don't understand about the circular financing is how does it not violate antitrust? All these supposed rivals own huge stakes in each other. Are they really competitors?
Zaibatsu - Clearly antitrust.
Keiretsu - perfectly acceptable business relationship preferences that in no way require governments to regulate their GDP growth / campaign donation black box machine.
The entire economy relies on circular flows. That’s why we look at something called “net profit”.
This made me chuckle, thanks.
[dead]
Well rules stopped mattering to BigCo many hears ago
Strange how you complain about this for circular flows when of course the GOAT of circular flows is the money system itself
govt prints money (FED) -> government -> banks -> (economy) -> tax -> government
And the second "great circle"
govt loans money (FED) (money that doesn't exist, which isn't a problem if you can print money) -> banks -> (economy) -> loan repayment -> money gets "unprinted"/destroyed
Everything about the economy is a great circle. You will also understand after a while, if you wish for the economy to remain stable, this must be the case. You can't change it without either hyperinflation and extreme "capital flight" (between quotes because of course in that circumstances capital will run into gold or other physical resources), by either vastly overheating or stopping the economy.
And yes, most instances are of the overheating kind. The only case of stopping the economy was accidental (despite many warnings), in other words, yes, it was Republicans. But last time that happened, even republicans could decide to change course. Of course, only once huge damage was done.
Way more people complain about this too, if you pay attention to all the US inflation and national debt anxieties. It’s clearly unsustainable and driving interest rates up steadily.
People complain about the imbalance between economy supply and economy demand leading to inflation and debt concerns.
The complaint isn’t that the economy is “circular”, how else could it work with 3 or more people engaged in commerce? Neither of those things get fixed on different ways to manage the money supply either, it’s not as if they didn’t have national debts or inflation while on the gold standard.
It’s just a different format. Providing liquidity in the hope that more is generated from the economy. Except it’s doing more harm than good other than in true crisis. Other times it’s just creating a bubble, in national debt and in AI.
Government doesn't print money. FED is a private institution.
Every FED governor is appointed by POTUS.
The actual control is with the "Board of Governors" which is a government agency (ie. the president controls them, subject to approval by the US senate). The banks structure is to make banks cooperate, it does not give them any kind of control over anything.
Re: positive Cashflow
Was that the one where they removed a bunch of big positions to make it positive (was it training cost, I forgot)?
The data published in their S1 says they took a $693M operational loss in the first half of this year.
That is from the increase in the share price, because they owe that in shares and the share price is higher now. On the other side is the IPO, it balances the books but it doesn't count as an asset today, whereas the promised shares count as a debt today.
They are cash flow positive, $3.4 billion in H1.
That means they're making money.
people should look at dividends rather than anything else; but the people who matter at these scales are just dick measuring and don't care about anything to do with the actual economics of the system.
It's more tax efficient for investors to sell stock than receive dividends. A myopic focus on dividends is economically illiterate.
They're not really "complex securities"; they're just convertible bonds. The lender gets to choose if they want to be paid back (1) in a fixed note value (like a bond), or (2) with shares of the company (like equity)—if it IPO's. (I.e., the value is the max() of the two. ReLU-shaped).
https://en.wikipedia.org/wiki/Convertible_bond
I don't understand the accounting aspects, and they seemed very weird to me, when I tried to read about was happening with Anthropic. Basically: the rules says the convertible debt has to be valued at the predicted IPO price—because that's the value of the thing (Anthropic stock) Anthropic will be obligated to pay. At the point they re-calculated this debt, the valuation went up drastically, so the rules say that counts as a very large quarterly "loss". But: on the other side of the scale, the increase in predicted IPO price does not count as a quarterly "profit"—even though they should cancel out. Even though the increase in share price exactly cancels the size of the debt they're paying back—the debt denominated as a fixed number of shares.
Well they can’t have it both ways on the accounting front.
These complex securities they took a mark on only exist as a way for them to access compute at deep deep discount by issuing suppliers lottery tickets.
Post IPO when they don’t have lottery tickets to issue as easily the numbers should be interesting.
This is correct. Out of the $80.7B loss, $72.4B was due to changes in the fair value of their convertible notes (and similar financial arrangements).
Going a bit meta, I'm curious why we don't observe a faster convergence of opinions here. I assume that HN users are well-versed in AI tech, and the current generation of LLMs will happily break down the relevant information in Form S-1. Nevertheless, we see a chaotic storm of opinions, with wildly varying S/N ratios.
Is it a limit of current LLMs, or is the bottleneck perhaps human nature?
Yes but those convertible notes only exist to grant them access to compute they would otherwise need to pay fair market value for.
If I drive a taxi making $200/day and need $150/day of gas.. but convince Shell to take a convertible note instead of $150.. Do I make a) $200/day profit, b) $50/day profit, or c) $200/day - (change_in_value[convertible_note]) in profit?
Some are arguing that C is somehow fake, but its obviously more real than A!
Human nature. Obviously we know some things are just true. It's common sense! Why would I invite a non-deterministic hallucination machine that can't think to challenge my world view with something inconvenient like facts?
People are emotionally invested in a deflation and slowdown in AI because they feel threatened by AI. They'll cling to reeds of hope and wilfully misread financial stats to make themselves feel better.
The tech blowout had little to do with Anthropic's filing.
The numbers from multiple companies show a high CapEx investment in a product that has yet to be proven as profitable as the very things it is hurting (i.e. ads). End of story.
Didnt it start with Korean stocks first? It feels like a wave, it hits there first then the US tech stocks. Korea index has stabilized for a few days now, but hasn't recovered what it lost.
The filing showed that it was profitable. The biggest expenses were because the stock valuation of Anthropic went really high.
I asked this to ChatGPT “if anthropics valuation didn’t rise and stayed at 200B what would their profit be on q2”
The answer is around 1B of profit.
The comment wasn't about binary profitability.
It doesn’t help Google to create an AI-powered business which turns $1B in profit annually if it results in them losing a search/ads business which turns $5B in profit annually.
The CapEx spend for AI not only makes AI itself more expensive (and thus less profitable), it’s now driving everything else in technology to be more expensive (and thus less profitable).
AI is a profitability tax on the entire tech sector and it remains unclear if the profit AI will deliver is going to be worth that.
I would say they should slow down on build-out to balance supplier/labor availability with available cash flows, but I suspect all the Big Tech companies are worried about falling behind the other and being put in the position to need to rent out compute.
It’s not clear what your argument is.
It’s more or less certain that Anthropic is already profitable if you discount the losses due to becoming too valuable.
What you are saying here
> The CapEx spend for AI not only makes AI itself more expensive (and thus less profitable),
There’s nothing inherently unique about this. It’s just supply and demand - the only problem is that the ecosystem of raw materials can’t keep up with the demand. It will catch up soon and the prices will stabilise.
Your only problem seems to be this aspect only: that RAM production hasn’t increased according to demand.
No I don’t agree they should slow down. I think RAM and other things should rather speed up to react appropriately to the demand.
It’s like saying the iPhone is not profitable because too many people want it and its price is going up.
[dead]
If they're referring to the recent tumble, I wouldn't suggest anthropic incited that
coreweave did
i'm sure ant didn't help, but thanks for the fire sale i guess
It's just the beginning. The real collapse will happen right after midterm.
Why? Genuinely interested.
Boy, people sure do love turning data points into stories.
in many ways, its the very thing that makes us people
Maybe! I expect a lot of animals construct narratives to make sense out of the world around them.
did he even renormalize the dataset on the S-1 fill size?
I tried to give this article an honest read but
> Last year, for instance, Anthropic lost $80 billion, and Amazon had negative $42 billion free cash flow. The entire big tech complex, it seems, has moved from minting cash to burning it. And investors are, well, not panicking exactly, but wondering if the bubble is ending
Is this really the quality of such newsletters? I’m not asking rhetorically because I’d like to also have trustable sources to understand the underlying reality. I actually spent time understanding this part and the 80B dollar loss thing is completely misrepresented here (also went viral in other media).
"The shift in politics is even starker. In the 2010s and early 2020s, big tech was well-liked by most people. But AI is absolutely loathed by the public. These companies have never faced real public dissatisfaction, they all imagine themselves to be the good guys, with the credibility of having created magical search engines and smartphones. But that era is gone. They are now akin to the bankers on Wall Street in 2008, and politicians are winning elections by saying they hate AI as much as the voters do."
Yeah I see this too. That we're not innovating any more also doesn't help make our industry look good
You can't claim that the tech industry isn't innovating. You can say that innovation effort has shifted away from many things that still need it. The issue I have is that the companies and products are so large and complex that they are extremely difficult to compete with.
It's not just that we're not innovating on good new products. We're innovating new ways to break the law and exploit human psychology.
Blockchain/NFT hype did a ton of damage to tech's credibility, as well as Uber before it (along with the many Uber for X apps).
I'm in agreement with the general point but was big tech well liked in the early 2020s?
Facebook scandals had exploded already, Google had already long abandoned "don't be evil", nobody ever liked Microsoft, and I think sentiment towards Apple is somewhat unchanged.
Tech at least in my eyes peaked around 2018-19. That also coincided with the first more mainstream public-aware introduction of vision and NLP systems into tech products. A few years later it all went downhill for the tech companies.
Do you not remember when they were dragged in 2015-2016? or 4 years before that when they were heroes instead?
I totally forgot about that. I guess my memory wanted them to be good around that time.
We have all been tricked, unfortunately.
>But AI is absolutely loathed by the public.
by the public populating the echo chambers on reddit and bluesky, perhaps.
9.5 out of 10 normies use it daily now and have no negative opinions about it. the wholesome folx use it too, despite their incessant kvetching.
Most people I talk to use it because it's useful, but also hate what it means and what it is doing to the world.
There definitely has been a broad change in public perception of big tech and the future that it is taking us toward.
was that change organic, though? every single shitrag went berserk about AI once the so-called journalists had realized they were about to lose their jobs to it.
I think there is a massive selection criteria bias, AI has been a batshit insane productivity boost for me but I’ve been so busy with it that I’ve generally stopped posting online - I only do so now because I’m out of quota. I no longer see any point in arguing with people - I’ve got better things to do, like argue with AI. I used to scour the Earth looking for smart people to talk to and now I don’t even have to leave my apartment.
[dead]
I use all sorts of things I hate all the time.
This is not the gotcha you think it is.
People began disliking big tech well before this latest bout of AI.
This is true but the last few years have seen this dislike grow and spread a lot more than it was before - in my personal experience.
This feeling made me leave the tech industry.
It felt like I was helping the bad guys, and my conscience couldn't deal with that.
In 2014 I moved from London to SF to start a big tech job. In my first day there I accidentally wondered into a massive rally, I asked one of the participants what it was about and they said they were protesting all these foreign tech workers coming to SF. It felt a bit weird to end up in the middle of an ‘anti-me’ rally - especially one so passionate and energetic. A few protesters pressed me on where I was from and what I do, I lied and told them I was an investment banker. They told me that they suppose that was ok. I specifically told them that because not long ago it was finance being the target of protests and I curious to see if they had somehow moved on - if this was a principled stance or if it was being astroturfed.
These protests ended up getting pretty violent, people were throwing rocks through bus windows, throwing up on cars.
hey little tip while you assimilate and learn the language - property damage is not violence
Don’t worry I didn’t stay long. I loved the Bay Area but did not like SF all that much.
I believe some of the busses were occupied when they were attacked and the lack of injuries was largely due to good fortune. I could believe a claim that this was done by false flag agitators intent on undermining the protests.
You got a few things wrong:
First, ChatGPT has been the most downloaded app for two years straight, the idea that AI is loathed by the general public is pure delusion, propped up by manipulated surveys and people living in a bluesky bubble.
Second, Big tech wasn't well liked by the early 2020s, that ended in the early 2010s when the NYT editorial board took a top down decision to cover big tech negatively.
Third, the idea that they were innovating before but aren't now is insane. AI is the biggest innovation in tech since the transistor and possibly the only true innovation since the GUI.
Respectfully, download numbers are not an adequate proxy for opinion. The number of companies mandating AI use plays in significantly. There are highly ploarized bubbles of opinion fwiw. In the end, it requires actual controlled polling of opinion to have a satisfactory understanding.
Additionally, I'm not sure how you square NYT coverage being more critical of generally worse behavior as "top down" without a citation, care to point at anything to substantiate that statement?
Innovative, in some part sure. But lately the running jokes have become "Uber, but for $X", which hints at imaginative stagnation. AI has been the largest shake up in a decade (or more), but it is hard for Average Joe to grasp what is changing.
Considering the search engine company is selling a very popular self driving taxi service, they seem to be quite credible and innovative.
[dead]
People are being led around by the nose. Cui bono? 2000 years and most people still can't manage to ask that question reflexively. AI hate is a parochial narrative being pushed by people who stand to lose influence. Examined objectively, there are reasons to be cautious when adopting AI tools but a knee-jerk rejection of everything AI is a purely irrational response that has been drilled into people through repetition, conformity and media influence.
Tech companies were never the "good guys" any more than they are the "bad guys" now. Those are just colored filters thrown in front of the lime light to shift your perception. As soon as tech or AI works to their benefit then suddenly they'll be the good guys again and no one will care they were the bad guys before. People just go where the algo leads them, believe what their influencers believe, think what their peers think.
People hate AI because culture is being turned into slop in front of our very eyes. There's clear evidence that people collectively reject:
1. AI games
2. AI Art
3. AI Music
4. AI literature in most forms, except for factual text like search
Or at least, they won't pay money for it
AI hatred isn't being pushed by people with a narrative. I can't use AI for my game, because it'll sell £0 if I do. That's the net result of millions of other people telling me that they despise AI and won't purchase my product if it contains any AI, and I'm just making a business decision
Also google and some other big tech companies were widely loved for a while. Microsoft/gates managed to entirely rehabilitate their image at one point, and google was seen as one of the good guys. Those days died with the cambridge analytica scandal, and there's no evidence its coming back
And why do those millions of other people have that opinion? Have they arrived at it of their own accord?
Yes, you can create garbage with AI tools. You can create garbage using Blender, Unreal Engine, Unity, GIMP, Ableton, iMovie and Word as well and there are mountains of garbage created with those tools that nobody pays for or likes. If you have some talent and put the same effort into creating art with AI that you do with a digital paint program then it won't be garbage. AI isn't a magic wand that you can wave around and get perfect results, it can save you time just like using an image editor saves you a lot of time compared to drawing and painting by hand but it won't do all of the work for you. It's a tool and a poor craftsman that blames his tools for his failure.
My point exactly that big tech was loved for a while and then hated and then loved again and hated again. Have they truly changed that much that often or is it your perception that has changed?
> Yes, you can create garbage with AI tools. You can create garbage using Blender, Unreal Engine, Unity, GIMP, Ableton, iMovie and Word
Do you not see any difference between creating garbage in these ways?
People have created low-quality things since the dawn of humanity, using all sorts of tools, professional and not. However, before AI, making most types of garbage still required some human work, it wasn't infinitely scalable. And what was automatable ended up being so horrible that most people could easily dismiss it on the spot, or for them to even be filtered out algorithmically. On the internet, it wasn't perfect, but the signal-to-noise ratio was still at a level where you reliably discern the signal part.
AI changed this. Even a beginner can set one up to run almost fully unattended - and while what comes out won't be good, it also won't be immediately apparent as such, especially to the general public. More and more I find myself being fooled by a website link, a YouTube video thumbnail or a paragraph in an article. AI lets people create garbage that can impersonate real effort decently well, and they've used that to great extents. As we speak, thousands of agents are working day and night to pump every online platform with their 'content', without a single person in the world to care about what they put out. They're taking over and removing any good signal we had, everything becomes diluted with trash on a scale we have never seen. People notice that. Many are clueless about AI, and social influence plays a role, but they're not being misled and have their opinions manipulated. They wouldn't enjoy most of this content regardless.
In addition to the dilution of good information, they care about appearances and how AI lets you create an appearance of caring about something. The cheap food stand seemed honest when they put up an unassuming print of a simple Word document, and now seems dishonest when they put up an AI-generated mess that looks nothing like the real product and desperately tries to seem more refined than they are. Their manager was at least honest when he replied with a lazy 'k', and now he's lying about caring by dumping out a bullet point list of AI-generated text for a response that we know he himself hasn't read. It's the dream product for scammers, grifters and people who simply don't care.
Can you use AI to help create something good? Sure, but anyone doing that still needs to take their time and put care into it. Even if AI makes them faster, by the time they finish, 10TB of more slop has already been produced to drown them out, because that can be fully automated. Effectively 99.9% of AI content ends up being garbage, so of course people are going to reject it. Most probably don't want the few nuggets of gold for the ocean of waste.
People hate these outputs, as well as the wholesale enclosure of their commons (training on public data ignoring their copyrights and the social contracts that enabled them). Additionally, despite how it might appeal to a subset of humanity, "we're going to take all of the jobs" is not a fantastic sales pitch.
How can the output both be so much worse than what a human could produce while simultaneously taking all the jobs and displacing the human made data it was trained on?
If everyone hates AI output and refuses to buy anything made with it then in what way is it a threat to people who don't use AI in their work?
Are these talking points perhaps crafted to benefit a small group of people who will lose the edge they currently have to a larger number of people enabled by these new tools?
Not really an accurate view, and the conspiracy angle isn't helping.
The output could be fine in itself, except people (arbitrarily, some could say) value the human production. Displacement is simply a matter of volume: the sleepless machine can churn indefinitely while people can not. The thing is too, most outputs being mass produced are of a substantially lower quality than people (and what is possible).
A problem is people are being tricked into it, or being given little choice. If 90% of companies use AI (as in, mandated from leadership), what "choice" is there? If people recognize the 10% remaining and do business with them exclusively, economics will take over and force either price hikes or the 10% to grow. It's playing chicken.
Based on what the trends seem to be, people will unknowingly keep taking in products with AI (mostly out of not understanding the choice).
As for the enclosure point which is important: People make things for their enjoyment. When people try to profit off their work, they rightfully get put off. The deal is "if you can view this for free, it isn't meant to be sold. If it is sold, I deserve to have my time compensated". AI training/scraping has violated this agreement, and funnels the money to a third party to the people who made the valuable originating art. If it was not valuable, then why not train AI on its own output? Since that appears to be a "problem", there must be something about what humans create that has value AI cannot produce itself.
>Are these talking points perhaps crafted to benefit a small group of people who will lose the edge they currently have to a larger number of people enabled by these new tools?
That flavour of conspiratorial thinking is one of the reasons for polarization. Sometimes, people prefer the human touch. This applies mostly to art and other topics lumped into "the Humanities".
Hope that clears things up for you, or at least sheds light on why people might not see things in the way you do.
> people (arbitrarily, some could say) value the human production
This is a different argument, going into it would muddy the waters. It would devolve into a debate about how much and what type of input and individual perception. It's also been rehashed a million times since Duchamp thought to put a mass produced urinal on display as a fountain.
> most outputs being mass produced are of a substantially lower quality than people (and what is possible)
Different purposes and mass production does not eliminate the market for artisanal work, it replaces only the limited batch work done by hand previously.
> A problem is people are being tricked into it...
God help anyone who learns how the sausage is made.
> "if you can view this for free, it isn't meant to be sold. If it is sold, I deserve to have my time compensated"
This is mostly a matter for the courts but as long as the works are not being reproduced in whole unattributed or in violation of whatever license they held then it will mostly be allowed. If you look for it, you'll notice that artists "borrow" from each other constantly with no attribution and no payment to the originator. All but the grossest plagiarism goes entirely unpunished because everyone realizes that no one can stand alone. Is it different now that it is being made available as a service? Yes, there are now more legal ramifications due to an audit trail of scraping and purchases combined with the much larger output that can be measured and compared. Does it destroy the commons? No, people will use the outputs in their own work which will be published and become part of the commons just like any new style or media.
> that appears to be a "problem"
It's a solved problem. Synthetic training data is a valuable addition now. It requires a way to validate or rank it but that's not any worse than cleaning and filtering human data.
> conspiratorial thinking
An excellent smokescreen but we've been living in a soup of mass media propaganda for over a hundred years now so there's no denying that influence campaigns exist. Denying that humans act in coordinated ways to influence the opinion of others is absurd when you can get a degree in it, have it as a job title and build a trillion dollar industry on it.
>This is a different argument, going into it would muddy the waters. It would devolve into a debate about how much and what type of input and individual perception. It's also been rehashed a million times since Duchamp thought to put a mass produced urinal on display as a fountain.
On the contrary, it demonstrates an unwillingness to grapple with inconvenient problems, as does the rest of your reply.
>'s a solved problem. Synthetic training data is a valuable addition now. It requires a way to validate or rank it but that's not any worse than cleaning and filtering human data.
Untrue for artistic outputs oddly enough.
>build a trillion dollar industry on it.
Or selectively quote in the comment section to dullen the need to address issues.