Pundit ClaimsAI claims by Gary Marcus and Ed Zitron, with the evidence

Changes, page 14 of 55

Every substantive change to the register, newest first. Speaker events record what the people covered said or changed. Site events record what this site published, corrected, flagged, or withdrew. The two never share a label.

Substantive changes (10984)

Page 14 of 55: changes 2601 to 2800 of 10984, newest first. Dates are when the change was recorded.

  1. HeldBy this siteEd Zitron

    Held record wyea-make-fun-of-them-74ee9d

    The anthropic review found the criteria unfair to the speaker's words.

  2. Assessment publishedBy this siteEd Zitron

    OpenAI's consumer product loses billions of dollars.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-make-fun-of-them-695d7d-a1.

  3. Assessment publishedBy this siteEd Zitron

    AI agents are fundamentally unworkable because LLMs are inherently unreliable and break down on multi-step tasks.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-make-fun-of-them-3ecb07-2-a1.

  4. Assessment publishedBy this siteEd Zitron

    Very few of ChatGPT's 500 million users pay, and user numbers outside ChatGPT are not there.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-make-fun-of-them-28afc4-a1.

  5. Assessment publishedBy this siteEd Zitron

    OpenAI under Altman is nowhere near delivering a functioning agent, let alone anything approaching intelligence.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-make-fun-of-them-0ecb19-a1.

  6. Assessment publishedBy this siteEd Zitron

    Meta is not making any money from generative AI.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-make-fun-of-them-095ddc-a2.

  7. Assessment publishedBy this siteEd Zitron

    The AI bubble has been inflated by hype and lies from Altman and Amodei, amplified by tech media.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-longcon-aa0483-a2.

  8. HeldBy this siteEd Zitron

    Held record wyea-longcon-a05d5c

    e3 is dated 2025-02, which is not inside the period described (2025-01-23 to 2025-02-17); it can only be context

  9. HeldBy this siteEd Zitron

    Held record wyea-longcon-959fea

    The anthropic review reads the claim differently.

  10. HeldBy this siteEd Zitron

    Held record wyea-longcon-8be835

    e2 is dated 2025-02, which is not inside the period described (2025-01-23 to 2025-02-17); it can only be context; e3 is dated 2025-02, which is not inside the period described (2025-01-23 to 2025-02-17); it can only be context

  11. Assessment publishedBy this siteEd Zitron

    User numbers alone tell you nothing about a business's sustainability or profitability, or about how its users use the product.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-longcon-84747c-a1.

  12. Assessment publishedBy this siteEd Zitron

    The tech industry and part of the economy are heading for a crash driven by AI leaders' hype.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-longcon-723509-a2.

  13. HeldBy this siteEd Zitron

    Held record wyea-longcon-6267f9

    The final-text check failed: proposition_faithful, citations_correct.

  14. Assessment publishedBy this siteEd Zitron

    Most of the AI coverage you read is led by companies that benefit financially from you thinking AI is important, and by default all of this coverage mentions OpenAI or ChatGPT.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-longcon-059e5e-a1.

  15. Assessment publishedBy this siteEd Zitron

    AI GPUs are useful for generative AI and little else.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-let-ai-burn-f45d5c-a2.

  16. Assessment publishedBy this siteEd Zitron

    Once the AI hype subsides, it will become obvious that real demand for AI was not strong.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-let-ai-burn-e46913-a2.

  18. Assessment publishedBy this siteEd Zitron

    Cheaper Chinese open-source models already offer performance similar to American labs' premium products.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-let-ai-burn-d06b54-a1.

  19. Assessment publishedBy this siteEd Zitron

    GPUs, unlike fiber or railroads, serve basically one use, which lacks meaningful demand outside subsidies and circular financing.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-let-ai-burn-c65a07-a2.

  20. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-c35f2b

    hedged language is not scored, even on a plain reading

  21. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-c19e2d

    The openai review found the criteria unfair to the speaker's words.

  22. Assessment publishedBy this siteEd Zitron

    Demand for compute beyond OpenAI and Anthropic is not large; they make up 50% of hyperscaler RPOs ($448B).

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-let-ai-burn-afe890-a1.

  23. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-a79b2f

    The openai review found the criteria unfair to the speaker's words.

  24. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-7fb04d

    The final-text check failed: citations_correct.

  25. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-7e44d8

    The anthropic review found the criteria unfair to the speaker's words.

  26. Assessment publishedBy this siteEd Zitron

    There will be no bailout for banks' AI data center investments, and none is needed.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-let-ai-burn-7d36fb-a1.

  27. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-694555

    The openai review reads the claim differently.

  28. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-62614c

    The anthropic review reads the claim differently.

  29. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-61d9ba

    The final-text check failed: citations_correct.

  30. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-609e1e

    The final-text check failed: citations_correct.

  32. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-5fe3c4

    The final-text check failed: proposition_faithful, citations_correct.

  33. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-4ef27a

    The final-text check failed: citations_correct.

  35. Assessment publishedBy this siteEd Zitron

    Lack of data center capacity is not holding back generative AI usage or product launches.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-let-ai-burn-4a62db-a1.

  36. Assessment publishedBy this siteEd Zitron

    When the bubble bursts, the vast majority of big tech will be fine, at least in the short term.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-let-ai-burn-4332c3-a2.

  37. Assessment publishedBy this siteEd Zitron

    There is little compelling evidence that inference services are profitable.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-let-ai-burn-3595f3-a2.

  38. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-2cd717

    The two reviews differ: anthropic historical_observation/not_applicable/mixed, openai historical_observation/not_applicable/unresolved.

  39. Assessment publishedBy this siteEd Zitron

    Generative AI will not lead to AGI.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-let-ai-burn-24be19-a1.

  40. Assessment publishedBy this siteEd Zitron

    If OpenAI cannot go public, SoftBank will face a real liquidity crisis given its $40B+ short-term loans for OpenAI.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-let-ai-burn-24941d-a2.

  41. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-1b20a2

    The openai review found the criteria unfair to the speaker's words.

  42. Assessment publishedBy this siteEd Zitron

    After the bubble bursts, leftover GPUs will not be useful to hobbyists or cheaper to run, and incomplete data centers will not be cheaper to finish.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-let-ai-burn-1a5cfa-a2.

  43. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-1846e3

    The openai review found the criteria unfair to the speaker's words.

  44. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-134c72

    The openai review found the criteria unfair to the speaker's words.

  45. HeldBy this siteEd Zitron

    Held record wyea-let-ai-burn-085c67

    The openai review found the criteria unfair to the speaker's words.

  46. Assessment publishedBy this siteEd Zitron

    Giving $42B to OpenAI or Anthropic would not fix their business models or create the $400B+ annual revenue needed to justify NVIDIA's GPU sales through 2027.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-let-ai-burn-0261d0-a2.

  47. Assessment publishedBy this siteEd Zitron

    The author sees no scenario in which costs for Anthropic's and OpenAI's services come down or their rate limits are eased.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-i-will-never-respect-a-website-f8e605-a1.

  48. Assessment publishedBy this siteEd Zitron

    As capacity limits are hit, Anthropic and OpenAI will degrade user experience as much as they can via model degradation or rate limits.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-i-will-never-respect-a-website-f28d3d-a2.

  49. Assessment publishedBy this siteEd Zitron

    LLMs cannot and will not do what AI executives' job-loss and superintelligence rhetoric claims.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-i-will-never-respect-a-website-f10257-a1.

  50. HeldBy this siteEd Zitron

    Held record wyea-i-will-never-respect-a-website-eb8ad7

    The anthropic review found the criteria unfair to the speaker's words.

  51. Assessment publishedBy this siteEd Zitron

    LLMs will always hallucinate, so their outputs are not trustworthy.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-i-will-never-respect-a-website-d4324f-a1.

  52. Assessment publishedBy this siteEd Zitron

    LLMs cannot be deterministic.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-i-will-never-respect-a-website-d4324f-2-a1.

  53. Assessment publishedBy this siteEd Zitron

    For OpenAI and Anthropic subscribers, rate limits will make current workflows impossible and prices will increase.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-i-will-never-respect-a-website-d24847-a1.

  54. Assessment publishedBy this siteEd Zitron

    Every LLM product is sold under false pretenses about long-term sustainability and possibility.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-i-will-never-respect-a-website-c4e27f-a1.

  55. HeldBy this siteEd Zitron

    Held record wyea-i-will-never-respect-a-website-adba18

    The openai review reads the claim differently.

  56. Assessment publishedBy this siteEd Zitron

    Every AI agent story is about code generation, summarization, or generating content from an information source of uncertain trustworthiness.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-i-will-never-respect-a-website-a5d7f8-a1.

  58. HeldBy this siteEd Zitron

    Held record wyea-i-will-never-respect-a-website-79663e

    The openai review found the criteria unfair to the speaker's words.

  59. Assessment publishedBy this siteEd Zitron

    The attack on Sam Altman's home was a direct result of the AI industry's deceptive scare campaign.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-i-will-never-respect-a-website-38adf6-a1.

  60. Assessment publishedBy this siteEd Zitron

    On June 1 2026, Codex rate limits will be halved.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-i-will-never-respect-a-website-3312fd-a1.

  61. HeldBy this siteEd Zitron

    Held record wyea-i-will-never-respect-a-website-1f0c4e

    a observation claim cannot be assessed on the causal_assessment axis

  62. HeldBy this siteEd Zitron

    Held record wyea-i-will-never-respect-a-website-10274b

    The final-text check failed: proposition_faithful.

  64. Assessment publishedBy this siteEd Zitron

    When the AI bubble ends, many people will ask how it happened and will want a bailout of AI companies.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-hyperscale-normalization-eeab64-a2.

  65. Assessment publishedBy this siteEd Zitron

    Outside rentals to Anthropic and OpenAI, AI GPUs generate very little economic value, with few signs that will change.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-hyperscale-normalization-e166c8-a2.

  66. Assessment publishedBy this siteEd Zitron

    LLMs create outputs based on a summary of past data and pass that off as 'intelligence' in a way that coddles the user.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-hyperscale-normalization-d9f876-a1.

  67. HeldBy this siteEd Zitron

    Held record wyea-hyperscale-normalization-d5781f

    The openai review reads the claim differently.

  68. HeldBy this siteEd Zitron

    Held record wyea-hyperscale-normalization-cd0a17

    The final-text check failed: citations_correct.

  69. HeldBy this siteEd Zitron

    Held record wyea-hyperscale-normalization-c4c109

    The openai review reads the claim differently.

  70. HeldBy this siteEd Zitron

    Held record wyea-hyperscale-normalization-c3a122

    The final-text check failed: proposition_faithful.

  71. HeldBy this siteEd Zitron

    Held record wyea-hyperscale-normalization-c24ec4

    The openai review found the criteria unfair to the speaker's words.

  72. Assessment publishedBy this siteEd Zitron

    Basically every AI data center Zitron has seen uses environmentally disastrous behind-the-meter gas turbines.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-hyperscale-normalization-c11d4d-a1.

  74. Assessment publishedBy this siteEd Zitron

    Excluding OpenAI and Anthropic, there is not much demand for AI services.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-hyperscale-normalization-97efb9-a1.

  76. HeldBy this siteEd Zitron

    Held record wyea-hyperscale-normalization-7a6a66

    a observation claim cannot be assessed on the causal_assessment axis

  77. HeldBy this siteEd Zitron

    Held record wyea-hyperscale-normalization-4919e0

    The openai review found the criteria unfair to the speaker's words.

  78. Assessment publishedBy this siteEd Zitron

    The data center under the Anthropic-Lambda deal will not exist for years, if ever.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-hyperscale-normalization-4919e0-2-a1.

  79. HeldBy this siteEd Zitron

    Held record wyea-hyperscale-normalization-45e18a

    The openai review reads the claim differently.

  80. HeldBy this siteEd Zitron

    Held record wyea-hyperscale-normalization-33c6ff

    The two reviews differ: anthropic historical_observation/not_applicable/unresolved, openai historical_observation/not_applicable/mixed.

  81. HeldBy this siteEd Zitron

    Held record wyea-hyperscale-normalization-20bb29

    The two reviews differ: anthropic historical_observation/not_applicable/unresolved, openai historical_observation/not_applicable/contradicted.

  83. Assessment publishedBy this siteEd Zitron

    GPT-4 did not autonomously hire a TaskRabbit worker or take the actions described; it was prompted and humans carried them out.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-f60bb5-a1.

  84. Assessment publishedBy this siteEd Zitron

    AI agents are not autonomous, do not replace jobs or coders, and will not do so, because probabilistic models are poor at taking precise actions.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-f27777-a1.

  85. HeldBy this siteEd Zitron

    Held record wyea-how-to-argue-with-an-ai-booster-e52c31

    The openai review found the criteria unfair to the speaker's words.

  86. HeldBy this siteEd Zitron

    Held record wyea-how-to-argue-with-an-ai-booster-e0a202

    e1 is dated 2025-07, which is not inside the period described (2025-07-17 to 2025-08-25); it can only be context

  87. HeldBy this siteEd Zitron

    Held record wyea-how-to-argue-with-an-ai-booster-d46281

    The anthropic review found the criteria unfair to the speaker's words.

  88. Assessment publishedBy this siteEd Zitron

    Inference costs rose because model developers hit diminishing returns and could only improve models by having them burn more tokens through reasoning.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-d06e78-a1.

  89. Assessment publishedBy this siteEd Zitron

    Apart from construction spending, AI data centers do not drive much economic growth.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-c69921-a1.

  90. Assessment publishedBy this siteEd Zitron

    AI models keep getting larger and more expensive.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-a265e4-a1.

  91. Assessment publishedBy this siteEd Zitron

    Generative AI companies, despite ample funding and attention, have not produced a product with a defined, meaningful mass-market use case.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-a0927e-a1.

  92. Assessment publishedBy this siteEd Zitron

    Model developers have only managed to make new models more effective through reasoning models that burn extra tokens via test-time compute.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-8a4b43-a2.

  93. Assessment publishedBy this siteEd Zitron

    AI model developers are out of ideas and will not escape the path of improving models by burning more tokens.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-80bd07-a1.

  94. HeldBy this siteEd Zitron

    Held record wyea-how-to-argue-with-an-ai-booster-7dcec2

    The anthropic review found the criteria unfair to the speaker's words.

  96. HeldBy this siteEd Zitron

    Held record wyea-how-to-argue-with-an-ai-booster-772381

    the result turns on our interpretation rather than the speaker's words, so it cannot be scored without a stated plain reading

  97. HeldBy this siteEd Zitron

    Held record wyea-how-to-argue-with-an-ai-booster-67c5f4

    The openai review found the criteria unfair to the speaker's words.

  98. HeldBy this siteEd Zitron

    Held record wyea-how-to-argue-with-an-ai-booster-679acd

    The anthropic review found the criteria unfair to the speaker's words.

  99. HeldBy this siteEd Zitron

    Held record wyea-how-to-argue-with-an-ai-booster-60b11f

    The final-text check failed: citations_correct.

  100. Assessment publishedBy this siteEd Zitron

    Even if OpenAI spends heavily tailoring ChatGPT for government use, it will not become essential, because LLMs are not very useful.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-501ea0-a2.

  101. HeldBy this siteEd Zitron

    Held record wyea-how-to-argue-with-an-ai-booster-46cfa0

    The anthropic review does not accept e2 as supports.

  102. HeldBy this siteEd Zitron

    Held record wyea-how-to-argue-with-an-ai-booster-3cf87f

    a observation claim cannot be assessed on the causal_assessment axis

  103. Assessment publishedBy this siteEd Zitron

    Counting all its funding plus 30% of Amazon's capex since 2024, Anthropic's true cost is about $77.1 billion.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-39aff3-a1.

  104. Assessment publishedBy this siteEd Zitron

    For the most part, the companies buying AI GPUs are not going to go bust once the AI bubble bursts.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-37eeb6-a2.

  105. Assessment publishedBy this siteEd Zitron

    Enterprise demand will not rescue generative AI companies, because enterprises have already tried the technology and it is not working.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-379584-a2.

  106. Assessment publishedBy this siteEd Zitron

    Nobody will build the self-learning, self-teaching AI agent that AI 2027 assumes.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-323580-a1.

  107. Assessment publishedBy this siteEd Zitron

    AI GPUs will not be adopted for another mass-scale use case, because GPUs cannot easily switch tasks.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-1e5ae0-a1.

  108. Assessment publishedBy this siteEd Zitron

    The real cost of AI inference has risen, and claims that it is falling are false.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-15df5f-a2.

  109. Assessment publishedBy this siteEd Zitron

    Attributing 30% of Microsoft's capex since 2024 and adding OpenAI's $41.4B in funding, OpenAI's true cost is about $82 billion.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-0dc738-a1.

  110. Assessment publishedBy this siteEd Zitron

    The cost of inference is not trending downward and has not gone down yet, though he says it maybe could go down.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-to-argue-with-an-ai-booster-064588-a2.

  111. Assessment publishedBy this siteEd Zitron

    Popular home pages on the internet are full of AI-generated misinformation, among other online harms users face.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-how-654cb5-a1.

  112. HeldBy this siteEd Zitron

    Held record wyea-how-17b51b

    The openai review found the criteria unfair to the speaker's words.

  113. Assessment publishedBy this siteEd Zitron

    Even with racks of Blackwell GPUs, generative AI models are not going to get better.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-godot-isnt-making-it-ef7b25-a1.

  114. Assessment publishedBy this siteEd Zitron

    The revenue expected from generative AI is not coming.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-godot-isnt-making-it-d2786c-a1.

  115. Assessment publishedBy this siteEd Zitron

    Advertising will not stop OpenAI's losses, which will eventually drown the company.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-godot-isnt-making-it-b1f398-a1.

  116. Assessment publishedBy this siteEd Zitron

    OpenAI is a terrible business burning $5 billion to make $3.4 billion, with no proof it can reduce costs.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-godot-isnt-making-it-97c288-a2.

  117. Assessment publishedBy this siteEd Zitron

    Generative AI will not have a transformative iPhone moment.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-godot-isnt-making-it-6ade4f-a1.

  118. Assessment publishedBy this siteEd Zitron

    Generative AI is going nowhere, and at some point Google, Microsoft or Meta will pull back on AI capital expenditures.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-godot-isnt-making-it-668c80-a1.

  119. Assessment publishedBy this siteEd Zitron

    Before hyperscalers pull back, there will be many desperate stories claiming AI gains can come from outside training new models.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-godot-isnt-making-it-640fc9-a2.

  120. HeldBy this siteEd Zitron

    Held record wyea-godot-isnt-making-it-4e76a9

    the conclusive event e1 is not outcome evidence in the packet

  121. HeldBy this siteEd Zitron

    Held record wyea-godot-isnt-making-it-4b8c03

    The final-text check failed: proposition_faithful.

  122. Assessment publishedBy this siteEd Zitron

    Chips will not bring down inference costs quickly enough for AI companies to turn their finances around.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-godot-isnt-making-it-4278f1-a2.

  123. Assessment publishedBy this siteEd Zitron

    When the AI bubble pops, the lack of other hyper-growth markets will cause a bloodbath in big tech stocks.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-godot-isnt-making-it-407eda-a2.

  124. Assessment publishedBy this siteEd Zitron

    At some point a hyperscaler will pull back and the AI era will end.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-godot-isnt-making-it-36d73e-2-a1.

  125. Assessment publishedBy this siteEd Zitron

    None of the big AI companies ever had a plan for generative AI beyond throwing GPUs at it.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-godot-isnt-making-it-19f6d7-a1.

  126. Assessment publishedBy this siteEd Zitron

    Meaningful generative AI products are not coming.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-godot-isnt-making-it-04f4bd-a1.

  127. Assessment publishedBy this siteEd Zitron

    Subsidized AI token pricing is ending soon, exposing customers to the true costs.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-four-horsemen-of-the-aipocalypse-df2ee7-a1.

  128. Assessment publishedBy this siteEd Zitron

    NVIDIA is selling years' worth of GPUs in advance of installable capacity, and those GPUs are being warehoused.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-four-horsemen-of-the-aipocalypse-deeb9e-a1.

  129. HeldBy this siteEd Zitron

    Held record wyea-four-horsemen-of-the-aipocalypse-d4d2d6

    The openai review found the criteria unfair to the speaker's words.

  130. HeldBy this siteEd Zitron

    Held record wyea-four-horsemen-of-the-aipocalypse-c45343

    The final-text check failed: proposition_faithful, rationale_supported_by_evidence; overclaims: This earlier, narrower snapshot neither verifies nor contradicts the claimed April 1–21, 2026 figures: 114GW scheduled through end-2028, with 15.2GW under construction, including 5GW scheduled for 2026 [e1]..

  133. HeldBy this siteEd Zitron

    Held record wyea-four-horsemen-of-the-aipocalypse-8c93f5

    The openai review found the criteria unfair to the speaker's words.

  134. HeldBy this siteEd Zitron

    Held record wyea-four-horsemen-of-the-aipocalypse-8b733d

    e2 measures something else (metric Cumulative sales visibility through 2027 is not Cumulative Blackwell and Rubin sales visibility); it can only be context unless it is derived with stated inputs; e3 is decisive for a financial claim but states no metric scope

  135. HeldBy this siteEd Zitron

    Held record wyea-four-horsemen-of-the-aipocalypse-805c11

    The anthropic review found the criteria unfair to the speaker's words.

  136. Assessment publishedBy this siteEd Zitron

    Microsoft's GitHub Copilot changes are driven by its weekly running costs nearly doubling since January 2026.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-four-horsemen-of-the-aipocalypse-7ecdaf-a1.

  137. Assessment publishedBy this siteEd Zitron

    At least 4.6GW of the 15.2GW of AI data center capacity under construction is for OpenAI and at least 4GW for Anthropic.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-four-horsemen-of-the-aipocalypse-67610a-a2.

  138. HeldBy this siteEd Zitron

    Held record wyea-four-horsemen-of-the-aipocalypse-56fa28

    The anthropic review reads the claim differently.

  139. Assessment publishedBy this siteEd Zitron

    The vast majority of AI services currently subsidize their subscriptions.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-four-horsemen-of-the-aipocalypse-53d623-a1.

  140. HeldBy this siteEd Zitron

    Held record wyea-four-horsemen-of-the-aipocalypse-499ff5

    The final-text check failed: proposition_faithful, rationale_supported_by_evidence, citations_correct; overclaims: Goldman's April 15, 2026 report describes one unnamed industry datapoint in which engineering inference costs were approaching 10% of engineering headcount cost [e1]. | The figure concerns engineering headcount costs, not all employee costs, and says 'approaching' 10%..

  141. Assessment publishedBy this siteEd Zitron

    Further price increases or service degradations from Anthropic and OpenAI would indicate they are running low on cash.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-four-horsemen-of-the-aipocalypse-419d8f-a2.

  142. Assessment publishedBy this siteEd Zitron

    AI subscription services almost always lose money on users unless they use the subscription lightly.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-four-horsemen-of-the-aipocalypse-29ab9d-a1.

  143. Assessment publishedBy this siteEd Zitron

    By the end of 2026, most AI services will move some or all customers to token-based billing.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-four-horsemen-of-the-aipocalypse-21f43b-a1.

  144. HeldBy this siteEd Zitron

    Held record wyea-four-horsemen-of-the-aipocalypse-119c98

    The anthropic review reads the claim differently.

  146. Assessment publishedBy this siteEd Zitron

    Zuckerberg gave up on the metaverse as soon as AI became the new hype trend.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-fire-mark-zuckerberg-938237-a1.

  147. HeldBy this siteEd Zitron

    Held record wyea-fire-mark-zuckerberg-63e75b

    hedged language is not scored, even on a plain reading

  148. Assessment publishedBy this siteEd Zitron

    Less than two years after renaming itself, Meta is pivoting again from the metaverse to AI as its largest investment.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-fire-mark-zuckerberg-03c2d7-a1.

  149. Assessment publishedBy this siteEd Zitron

    There is an AI bubble, and Altman himself inflated it.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-false-prophet-c24972-a1.

  150. HeldBy this siteEd Zitron

    Held record wyea-false-prophet-994acd

    The openai review does not accept e3 as supports.

  151. HeldBy this siteEd Zitron

    Held record wyea-false-prophet-82fc07

    The openai review found the criteria unfair to the speaker's words.

  152. HeldBy this siteEd Zitron

    Held record wyea-false-prophet-7bd619

    The openai review found the criteria unfair to the speaker's words.

  153. Assessment publishedBy this siteEd Zitron

    As things get desperate, Altman will use sheer force of will and push OpenAI to grow and sell into as many industries as possible.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-false-prophet-7a6d66-a1.

  154. HeldBy this siteEd Zitron

    Held record wyea-false-prophet-70449a

    The openai review found the criteria unfair to the speaker's words.

  155. Assessment publishedBy this siteEd Zitron

    When the AI house of cards collapses, Altman will blame external factors and remain a billionaire.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-false-prophet-6ae0e8-a2.

  156. Assessment publishedBy this siteEd Zitron

    Generative AI will not become the incredible supercomputer Altman has promised.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-false-prophet-590350-a1.

  157. HeldBy this siteEd Zitron

    Held record wyea-false-prophet-585cfe

    The final-text check failed: citations_correct.

  158. HeldBy this siteEd Zitron

    Held record wyea-falling-with-style-3fb5ac

    e1 is dated 2023-04, which is not inside the period described (2023-03-01 to 2023-04-13); it can only be context; e2 is dated 2023-04, which is not inside the period described (2023-03-01 to 2023-04-13); it can only be context; e1 is decisive for a financial claim but states no metric scope; e2 is decisive for a financial claim but states no metric scope; e3 is decisive for a financial claim but states no metric scope

  159. HeldBy this siteEd Zitron

    LLMs are and always will be incapable of reasoning or consciousness.

    The final-text check failed: citations_correct.

  160. HeldBy this siteEd Zitron

    Held record wyea-expectations-versus-reality-fe4f4f-2

    The anthropic review found the criteria unfair to the speaker's words.

  161. Assessment publishedBy this siteEd Zitron

    So-called autonomous agents are effectively chatbots answering questions from documents, a decade-old technology.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-expectations-versus-reality-f2bbd3-a1.

  162. Assessment publishedBy this siteEd Zitron

    Rabbit will try to charge its roughly 100,000 customers a subscription fee to stay afloat.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-expectations-versus-reality-f0e72a-a1.

  163. HeldBy this siteEd Zitron

    Held record wyea-expectations-versus-reality-f0c979

    The openai review found the criteria unfair to the speaker's words.

  164. Assessment publishedBy this siteEd Zitron

    Generative AI hallucinates because it lacks consciousness and any ability to learn or know anything.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-expectations-versus-reality-eda15a-a1.

  165. HeldBy this siteEd Zitron

    Held record wyea-expectations-versus-reality-df1f59

    The anthropic review does not accept e1 as contradicts.

  166. Assessment publishedBy this siteEd Zitron

    The Rabbit R1 is essentially an Android app in a box.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-expectations-versus-reality-bc4034-a2.

  167. Assessment publishedBy this siteEd Zitron

    If an app like Uber changes its website layout or flows, the corresponding Rabbit R1 functionality would break.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-expectations-versus-reality-aabc01-a2.

  168. Assessment publishedBy this siteEd Zitron

    Setting aside its security concerns, Rabbit's Large Action Model looks like it barely functions.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-expectations-versus-reality-a59892-a1.

  169. Assessment publishedBy this siteEd Zitron

    OpenAI failed to build a more efficient model for Microsoft in late 2023, and GPT-4 performs worse than its predecessor at some tasks.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-expectations-versus-reality-78badf-a1.

  170. HeldBy this siteEd Zitron

    Held record wyea-expectations-versus-reality-72e412

    The anthropic review does not accept e1 as supports.

  171. Assessment publishedBy this siteEd Zitron

    Sora will not succeed in replacing filmmakers because it is an impractical and ineffective solution.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-expectations-versus-reality-6a9b84-a1.

  172. HeldBy this siteEd Zitron

    Held record wyea-expectations-versus-reality-69f58f

    hedged language is not scored, even on a plain reading

  173. HeldBy this siteEd Zitron

    The Rabbit R1 lacks basic features such as timers, reminders, calendars, photos and email access.

    The final-text check failed: citations_correct.

  174. HeldBy this siteEd Zitron

    Rabbit must store user credentials in original form, which is insanely dangerous.

    The final-text check failed: citations_correct.

  175. Assessment publishedBy this siteEd Zitron

    Generative AI is extremely expensive because even simple GPT-4 prompts require complex computation on $10,000 GPUs.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-expectations-versus-reality-474ae0-a1.

  176. Assessment publishedBy this siteEd Zitron

    The Playwright scripting Rabbit uses is not AI.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-expectations-versus-reality-401cf2-a1.

  181. Assessment publishedBy this siteEd Zitron

    Over the next few quarters Anthropic's non-cancellable obligations will weigh on its balance sheet, the opposite of profitability, regardless of performance.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-exit-liquidity-f6a179-a1.

  182. HeldBy this siteEd Zitron

    Held record wyea-exit-liquidity-ede7e8

    The final-text check failed: rationale_supported_by_evidence; overclaims: Together, these disclosures support the claim under the contract's reading of changing business economics (e1)..

  184. Assessment publishedBy this siteEd Zitron

    A large chunk of Anthropic's and OpenAI's revenue depends on their customers continually raising capital.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-exit-liquidity-d51cb2-a1.

  186. HeldBy this siteEd Zitron

    Held record wyea-exit-liquidity-b0394d-2

    e3 is decisive for a financial claim but states no metric scope

  187. Assessment publishedBy this siteEd Zitron

    Anthropic's revenue concentration among large customers is certainly higher than the generous estimate of nearly one-quarter.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-exit-liquidity-ae378b-a2.

  188. Assessment publishedBy this siteEd Zitron

    Anthropic and OpenAI depend on venture-backed startups for 80% of enterprise revenue, and Anthropic used discounts to pump short-term revenue.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-exit-liquidity-93e0c7-a1.

  190. HeldBy this siteEd Zitron

    Held record wyea-exit-liquidity-91236f

    e1 measures something else (period 2026-07-01..2026-09-30 is not 2022-11..2026-09); it can only be context unless it is derived with stated inputs; e2 measures something else (period 2022-10-01..2022-12-31 is not 2022-11..2026-09); it can only be context unless it is derived with stated inputs

  191. Assessment publishedBy this siteEd Zitron

    Investors in Anthropic's IPO and Meta's stock sale will be left holding the bag.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-exit-liquidity-8d89d4-a2.

  192. HeldBy this siteEd Zitron

    Held record wyea-exit-liquidity-8a7bc6

    e1 is dated 2026-10, which is not inside the period described (2026-01-01 to 2026-10-28); it can only be context; e1 measures something else (metric Minimum compute commitment to Microsoft is not Minimum compute commitments to Google, Amazon and Microsoft, including the share not yet recognized); it can only be context unless it is derived with stated inputs; e2 measures something else (metric Minimum compute commitment to Google is not Minimum compute commitments to Google, Amazon and Microsoft, including the share not yet recognized); it can only be context unless it is derived with stated i

  193. HeldBy this siteEd Zitron

    Held record wyea-exit-liquidity-771313

    The final-text check failed: proposition_faithful.

  194. HeldBy this siteEd Zitron

    Held record wyea-exit-liquidity-729aa8

    The final-text check failed: proposition_faithful; overclaims: Anthropic will pay Broadcom at least $5B (possibly $10B) in 2027, rising to as much as $30B a year..

  195. Assessment publishedBy this siteEd Zitron

    Up to $2.31B (7%) of Anthropic's roughly $33.08B 2026 revenue through Q3 came from other companies distilling its models.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-exit-liquidity-7033f0-a1.

  196. HeldBy this siteEd Zitron

    Held record wyea-exit-liquidity-700a4f

    The anthropic review reads the claim differently.

  197. Assessment publishedBy this siteEd Zitron

    A large share of Anthropic's customers are unprofitable, unsustainable AI startups.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-exit-liquidity-6f995b-a2.

  198. HeldBy this siteEd Zitron

    Held record wyea-exit-liquidity-5da7bd

    hedged language is not scored, even on a plain reading

  199. Assessment publishedBy this siteEd Zitron

    AI data centers are not getting built and depend on endless high-interest debt.

    First assessment published after both reviews and the final-text check passed.

    Assessment wyea-exit-liquidity-460f5b-a1.

  200. HeldBy this siteEd Zitron

    Held record wyea-exit-liquidity-43f4ea

    a observation claim cannot be assessed on the causal_assessment axis