The Hater's Guide To The AI Bubble
We recorded 38 claims from this post. 15 were checked against evidence. 23 can’t be checked. 38 more are held. Read the post at its source.
Claims from this post (38)
In the order they appear in the post. Each quotation is the speaker’s words in this post. A label such as “Supported” or “Can’t be checked” is the claim’s current assessment; its page has the evidence and the reasons.
Assessment: Can’t be checkedSaid
Observation
“In short, I believe the AI bubble is deeply unstable, built on vibes and blind faith, and when I say "the AI bubble," I mean the entirety of the AI trade.”
Our reading: The AI bubble, meaning the whole AI trade, is deeply unstable and built on vibes and blind faith.
No horizon
Assessment: UnresolvedSaid
Observation
“We’re in a god damn bubble, it is so obvious we’re in a bubble, it’s been so obvious we’re in a bubble, a bubble that seems strong but is actually very weak, with a central point of failure.”
Our reading: AI is obviously in a bubble that looks strong but is very weak, with a central point of failure.
No horizon Evidence as of
Limitation: The cited measurements precede the target market dates and are context only.
Assessment: Can’t be checkedSaid
Conditional forecast
“If they keep their promises, by the end of 2025, Meta, Amazon, Microsoft, Google and Tesla will have spent over $560 billion in capital expenditures on AI in the last two years, all to make around $35 billion.”
Our reading: If they keep their promises, by end of 2025 these five companies will have spent over $560 billion in AI capex over two years to earn around $35 billion in AI revenue.
Deadline passed Evidence as of
Only if: if they keep their capex promises
Assessment: Can’t be checkedSaid
Forecast
“Google's $1.1 billion in workspace service revenue came from a forced price-hike on those who use Google services to run their businesses, meaning that this is likely not a number that can significantly increase without punishing them further.”
Our reading: Google's $1.1 billion Workspace AI revenue came from a forced price hike and probably cannot grow much without further punishing customers.
No horizon
Qualifier: without punishing business customers further
Assessment: UnresolvedSaid
Observation
“Someone's gonna get mad at me for saying this, but I believe that Meta is simply burning cash on generative AI.”
Our reading: Meta is simply burning cash on generative AI.
No horizon Evidence as of
Limitation: The reporting period and cash-flow meaning are evaluator interpretations.
Assessment: Can’t be checkedSaid
Forecast
“There is a shareholder vote for Tesla to potentially invest in xAI, which will probably happen, allowing Musk to continue to pull leverage from his Tesla stock until the company's decaying sales and brand eventually swallow him whole.”
Our reading: Tesla investing in xAI via a shareholder vote will probably happen.
No horizon
Qualifier: pending shareholder vote
Assessment: UnresolvedSaid
Causal claim
“To be clear, I am not saying that any of the Magnificent 7 are going to die — just that five companies' spend on NVIDIA GPUs largely dictate how stable the US stock market will be.”
Our reading: Five companies' spending on NVIDIA GPUs largely determines how stable the US stock market is, though none of the Magnificent 7 is necessarily going to die.
No horizon Evidence as of
Limitation: The cited assessments predate the target date and are context only.
Assessment: Can’t be checkedSaid
Observation
“Yet after that, generative AI feels more like a feature of cloud infrastructure rather than infrastructure itself.”
Our reading: Generative AI feels more like a feature of cloud infrastructure than infrastructure itself.
No horizon
Qualifier: feels more like
Assessment: Can’t be checkedSaid
Observation
“Businesses powered by generative AI do not seem to be popular.”
Our reading: Businesses powered by generative AI do not appear to be popular.
No horizon
Assessment: Can’t be checkedSaid
Forecast
“Cursor misled its customers, and its current revenue is, as a result, highly unlikely to stay at this level.”
Our reading: Cursor misled customers, and its current revenue is highly unlikely to stay at its current level.
No horizon
Assessment: UnresolvedSaid
Observation
“I realize it's likely a little boring hearing about software as a service, but this is the only place where generative AI can really make money.”
Our reading: Software as a service is the only place generative AI can really make money.
No horizon Evidence as of
Limitation: The revenue-versus-profit interpretation materially affects the assessment.
Assessment: Can’t be checkedSaid
Observation
“These are the single-most-popular consumer software subscriptions in the world — and OpenAI's 15.5 million subscribers suggest that it can't rely on them for the kind of growth that would actually make the company worth $300 billion (or more).”
Our reading: OpenAI's 15.5 million subscribers suggest it can't rely on them for the kind of growth that would make the company worth $300 billion or more.
No horizon
Qualifier: suggest
Assessment: UnresolvedSaid
Observation
“And worse still, it has no path to profitability, and it’s not even anything new!”
Our reading: Perplexity has no path to profitability, and it is not even anything new, being a search engine.
No horizon Evidence as of
Limitation: Financial context comes from a secondary summary of reporting that could not be read.
Assessment: Can’t be checkedSaid
Observation
“In any case, it doesn't seem like you can really build a consumer AI startup that makes anything approaching a real company.”
Our reading: Apart from ChatGPT, it does not seem possible to build a consumer AI startup that becomes anything like a real company.
No horizon
Qualifier: other than ChatGPT
Assessment: Can’t be checkedSaid
Causal claim
“As a result, it's very, very difficult for a company to build something unique.”
Our reading: Because all generative AI companies rely on LLMs with inherent limits, it is very hard for any of them to build something unique.
No horizon
Qualifier: because they all rely on LLMs with inherent limits
Assessment: Can’t be checkedSaid
Observation
“My Rot-Com Bubble thesis suggests we're out of hypergrowth ideas, and yeah, I think we're out of ideas related to Large Language Models too.”
Our reading: He thinks the tech industry is out of ideas related to Large Language Models, and his Rot-Com Bubble thesis suggests it is out of hypergrowth ideas generally.
No horizon
Qualifier: my thesis suggests
Assessment: Can’t be checkedSaid
Rhetoric
“The term "agent" is one of the most egregious acts of fraud I've seen in my entire career writing about this crap, and that includes the metaverse.”
Our reading: Calling AI products "agents" is one of the most egregious frauds the author has seen.
No horizon
Assessment: Can’t be checkedSaid
Observation
“In any case, every single company you see using the word agent is trying to mislead you.”
Our reading: Every company that uses the word "agent" is trying to mislead you.
No horizon
Assessment: Can’t be checkedSaid
Observation
“Nevertheless, none of these products are autonomous agents, and anybody using the term agent likely means "chatbot."”
Our reading: None of the 'agent' products he discussed (such as Salesforce Agentforce, ChatGPT agent, Glean, ServiceNow and coding agents) are autonomous agents, and anybody using the term 'agent' likely means 'chatbot'.
No horizon
Scope: these products (those discussed)
Assessment: UnresolvedSaid
Observation
“OpenAI and Anthropic both lose billions of dollars a year after revenue, and their stories do not mirror any other startup in history, not Uber, not Amazon Web Services, nothing.”
Our reading: OpenAI and Anthropic each lose billions of dollars a year after revenue, and their stories do not mirror any other startup in history, including Uber and AWS.
No horizon Evidence as of
Limitation: OpenAI's detailed accounting adjustments are unavailable in the cited report.
Assessment: Can’t be checkedSaid
Conditional forecast
“the most important company in the entire AI industry needs to convert by the end of the year or it's effectively dead”
Our reading: If OpenAI does not convert to a for-profit structure by the end of 2025, it will be effectively dead.
Deadline passed Evidence as of
Only if: OpenAI does not convert by the end of the year (2025)
Assessment: Can’t be checkedSaid
Conditional forecast
“I've said a lot of this before, which is why I'm not harping on about it, but the most important company in the entire AI industry needs to convert by the end of the year or it's effectively dead, and even if it does, it burns billions and billions of dollars a year and will die without continual funding.”
Our reading: If OpenAI does not convert to a for-profit by the end of 2025, it is effectively dead.
Deadline passed Evidence as of
Only if: if it fails to convert to for-profit by end of year
Assessment: ContradictedSaid
Observation
“As I wrote earlier in the year, there is really no significant adoption of generative AI services or products.”
Our reading: There is no significant adoption of generative AI services or products.
No horizon Evidence as of
Limitation: The decisive count is company-reported through CNBC, not independently audited.
Assessment: Can’t be checkedSaid
Observation
“Any and all stories about AI replacing jobs are intentionally manipulative attempts to boost stock valuations and suggest that models are capable of replacing human workers at scale.”
Our reading: Stories about AI replacing jobs are deliberately manipulative attempts to boost stock valuations; models cannot replace human workers at scale.
No horizon
Assessment: Can’t be checkedSaid
Observation
“To be abundantly clear, the manipulative suggestion here is that these models are autonomous or conscious in some way, which they are not.”
Our reading: AI models are not autonomous or conscious; stories about them scheming are deceptive.
No horizon
Assessment: Can’t be checkedSaid
Causal claim
“Large Language Models are too expensive to run, and have limited abilities beyond the ones I've named previously, and because everybody is running models that all, on some level, do the same thing, it's very hard for people to build really innovative products on top of them.”
Our reading: LLMs are too expensive to run and have limited abilities beyond those he named, and because everybody is running models that all, on some level, do the same thing, it is very hard to build really innovative products on top of them.
No horizon
Qualifier: on some level
Assessment: UnresolvedSaid
Conditional forecast
“OpenAI is Microsoft's largest Azure client — an insanely risky proposition on multiple levels, not simply in the fact that it’s serving the revenue at-cost but that Microsoft executives believed OpenAI would fail in the long term when they invested in 2023 — and Microsoft is NVIDIA's largest client for GPUs, meaning that any changes to Microsoft's future interest in OpenAI, such as reducing its data center expansion, would eventually hit NVIDIA's revenue.”
Our reading: Because Microsoft is NVIDIA's largest GPU client, any changes to Microsoft's future interest in OpenAI, such as reducing its data center expansion, would eventually hit NVIDIA's revenue.
No deadline stated Evidence as of
Only if: if Microsoft's future interest in OpenAI changes, e.g. reducing data center expansion
Assessment: UnresolvedSaid
Observation
“Generative AI products do not provide significant revenue growth, its products are not useful in the way that unlocks significant business value, and the products that have some adoption run at such a grotesque loss.”
Our reading: Generative AI products deliver no significant revenue growth or business value, and those with some adoption run at large losses.
No horizon Evidence as of
Limitation: The strongest customer benefit evidence is company-reported.
Assessment: Can’t be checkedSaid
Observation
“The money isn't there, the users aren't there, every company seems to lose money and some companies lose so much money that it's impossible to tell how they'll survive.”
Our reading: Generative AI lacks money and users, every company appears to lose money, and some lose so much that their survival is in doubt.
No horizon
Assessment: ContradictedSaid
Observation
“Generative AI does not do the things that it's being sold as doing, and the things it can actually do aren't the kind of things that create business returns, automate labor, or really do much more than one extension of a cloud software platform.”
Our reading: Generative AI does not do what it is marketed as doing, and what it can do neither creates business returns nor automates labor.
No horizon Evidence as of
Limitation: The decisive deployment figure is company-reported through a news outlet, not independently measured.
Assessment: UnresolvedSaid
Forecast
“There are limited amounts of capital, but also limited amounts of data centers to actually put GPUs, and on top of that, at some point growth will slow at one of the Magnificent 7, at which point costs will have to come down from things that lose them tons of money, such as generative AI.”
Our reading: At some point growth will slow at one of the Magnificent 7, forcing it to cut money-losing costs such as generative AI.
No deadline stated Evidence as of
Limitation: Revenue growth is an evaluator-selected interpretation of growth.
Assessment: UnresolvedSaid
Observation
“While it theoretically might be, all evidence points to larger models costing more money, especially reasoning-heavy ones like Claude Opus 4.”
Our reading: Although the cost of inference theoretically might be falling, all evidence points to larger models, especially reasoning-heavy ones like Claude Opus 4, costing more money.
No horizon Evidence as of
Limitation: The available studies concern other models and earlier measurements.
Assessment: ContradictedSaid
Forecast
“When? Say OpenAI and Broadcom actually build their ASIC in 2026 (they won't) — how many of them will they build?”
Our reading: OpenAI and Broadcom will not build their custom AI ASIC in 2026.
Settled before its deadline Evidence as of
Limitation: The decisive evidence is a company disclosure, not an independently reproduced hardware test.
Assessment: Can’t be checkedSaid
Causal claim
“Large Language Models authoritatively state things that are incorrect because they have no concept of right or wrong.”
Our reading: LLMs confidently state incorrect things because they have no concept of right or wrong.
No horizon
Assessment: SupportedSaid
Forecast
“At some point the momentum behind NVIDIA slows.”
Our reading: NVIDIA's momentum will slow at some point.
Settled before its deadline Evidence as of
Limitation: The result depends on interpreting momentum to include a sustained share-price plateau.
Assessment: Can’t be checkedSaid
Conditional forecast
“Perception matters just as much as actual numbers, and sometimes more, and a shift in sentiment could start a chain of events that knocks down the entire house of cards.”
Our reading: A shift in sentiment about NVIDIA could set off a chain of events that brings down the whole AI trade.
No horizon
Only if: a shift in sentiment around NVIDIA
Assessment: UnresolvedSaid
Forecast
“Sam Altman, Dario Amodei, Satya Nadella, Sundar Pichai, Tim Cook, Elon Musk, Mark Zuckerberg and Andy Jassy have overseen a needless, wasteful and destructive economic force that will harm our markets (and by a larger extension our economy) and the tech industry writ large, and when this is over, they must be held accountable.”
Our reading: The AI boom led by these tech CEOs will harm markets, the wider economy and the tech industry.
No deadline stated Evidence as of
Limitation: The market evidence establishes association, not a complete causal estimate.
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Sam Altman, Dario Amodei, Satya Nadella, Sundar Pichai, Tim Cook, Elon Musk, Mark Zuckerberg and Andy Jassy have overseen a needless, wasteful and destructive economic force that will harm our markets (and by a larger extension our economy) and the tech industry writ large, and when this is over, they must be held accountable.”
Our reading: Once the AI bubble is over, the tech CEOs who drove it must be held accountable.
No horizon
Qualifier: when this is over
Held from this post (38)
These records did not pass every check, so nothing about them is published. Held is not a finding.
- The criteria are ambiguous. Our reading, not assessed: If Meta, Amazon, Microsoft, Google, and Tesla keep their spending promises, by the end of 2025 they will together have spent over $560 billion in capital expenditures on AI over 2024 and 2025.
- The two reviews reached different assessments. Our reading, not assessed: OpenAI and Broadcom will not build their custom AI chip (ASIC) in 2026.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The criteria are ambiguous. Our reading, not assessed: The generative AI market is roughly a $50 billion revenue industry posing as a $1 trillion one, with the media's help.
- The two reviews read the claim differently
- The two reviews reached different assessments. Our reading, not assessed: The collapse of the AI bubble will wreck many people's retirements.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The criteria are ambiguous. Our reading, not assessed: NVIDIA's value depends on selling more GPUs every quarter than the quarter before.
- The two reviews reached different assessments. Our reading, not assessed: If any of the five big GPU buyers significantly changes its NVIDIA chip spending, it will eventually hurt the wider market directly and meaningfully.
- The record does not meet the current rules. Our reading, not assessed: Coding agents are not remotely intelligent, and when left to run unchecked they break things and create extra work.
- The two reviews read the claim differently
- The record does not meet the current rules. Our reading, not assessed: Markets are in one of the most brittle situations in economic history, held up by whether four or five companies continue to buy chips that start losing them money as soon as they are installed.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The criteria are ambiguous. Our reading, not assessed: The agents that Dario Amodei says will replace white-collar workers in the next few years will never exist.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The criteria are ambiguous. Our reading, not assessed: OpenAI's ChatGPT agent works very poorly and inconsistently and is probably very expensive to run.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews reached different assessments. Our reading, not assessed: If AI companies successfully move to custom ASICs, the AI trade is still damaged, because NVIDIA needs to keep selling GPUs.
- The two reviews reached different assessments. Our reading, not assessed: Anthropic and OpenAI are the biggest threats to their own customers and will rent-seek from and punish their successful customers.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews reached different assessments. Our reading, not assessed: No generative AI software company is profitable, and none shows hypergrowth like earlier big software companies.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently