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

Changes, page 12 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 12 of 55: changes 2201 to 2400 of 10984, newest first. Dates are when the change was recorded.

  1. HeldBy this siteEd Zitron

    Held record wyea-the-ai-industry-is-lying-to-you-49234e

    The anthropic review does not accept e1 as contradicts.

  2. HeldBy this siteEd Zitron

    Held record wyea-the-ai-industry-is-lying-to-you-3ee241

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

  3. HeldBy this siteEd Zitron

    Over four years Anthropic made about $5 billion in actual revenue against $25 billion or more in salaries and compute.

    The final-text check failed: rationale_supported_by_evidence; overclaims: Its corresponding personnel and compute-and-infrastructure expenses total $12.669 billion, which raises a substantial question about the $25 billion assertion..

  4. Assessment publishedBy this siteEd Zitron

    Spending on NVIDIA GPUs will eventually have to decline.

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

    Assessment wyea-the-ai-industry-is-lying-to-you-2524e1-a1.

  5. HeldBy this siteEd Zitron

    Held record wyea-the-ai-industry-is-lying-to-you-161a1e

    The anthropic review reads the claim differently.

  6. HeldBy this siteEd Zitron

    Held record wyea-the-ai-industry-is-losing-e77e7f

    The final-text check failed: proposition_faithful, rationale_supported_by_evidence; overclaims: Amazon's and Microsoft's apparent source figures include broader adjustments..

  8. Assessment publishedBy this siteEd Zitron

    The current GPU buildout will not produce a 'digital God', recursive self-improvement, or $600 billion+ in new revenue for current services.

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

    Assessment wyea-the-ai-industry-is-losing-cdf0e4-a1.

  9. HeldBy this siteEd Zitron

    Held record wyea-the-ai-industry-is-losing-b73975

    The final-text check failed: rationale_supported_by_evidence; overclaims: The unknown accounting basis independently prevents a supported or contradicted money assessment under the supplied rules..

  10. HeldBy this siteEd Zitron

    Held record wyea-the-ai-industry-is-losing-90366d

    The final-text check failed: conditions_shown, citations_correct.

  11. Assessment publishedBy this siteEd Zitron

    Altman, Amodei and backers lying about AI capabilities have caused a widespread mental health epidemic among tech workers.

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

    Assessment wyea-the-ai-industry-is-losing-76a75c-a1.

  12. Assessment publishedBy this siteEd Zitron

    Excluding OpenAI, Microsoft has essentially no meaningful AI business.

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

    Assessment wyea-the-ai-industry-is-losing-767c17-a1.

  13. Assessment publishedBy this siteEd Zitron

    Anthropic loses billions and subsidizes accounts letting users consume $8,000/month of tokens for $200.

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

    Assessment wyea-the-ai-industry-is-losing-6ec1d3-a1.

  15. HeldBy this siteEd Zitron

    Held record wyea-the-ai-industry-is-losing-36a5d5

    The openai review reads the claim differently.

  16. HeldBy this siteEd Zitron

    Held record wyea-the-ai-industry-is-losing-1aee0a

    The final-text check failed: citations_correct.

  17. HeldBy this siteEd Zitron

    Held record wyea-the-ai-industry-is-losing-166955

    a observation claim cannot be assessed on the causal_assessment axis

  18. Assessment publishedBy this siteEd Zitron

    The AI industry's plans and products are failing and it will run out of capital to keep going indefinitely.

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

    Assessment wyea-the-ai-industry-is-losing-143064-a1.

  19. HeldBy this siteEd Zitron

    AI does not deliver what it has been promised to tech workers.

    The final-text check failed: citations_correct.

  20. HeldBy this siteEd Zitron

    Held record wyea-the-ai-haters-manifesto-ff2325

    The openai review reads the claim differently.

  21. Assessment publishedBy this siteEd Zitron

    The AI industry cannot be justified by what it does today; it costs too much and none of the businesses built on it are profitable.

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

    Assessment wyea-the-ai-haters-manifesto-f41f64-a1.

  22. HeldBy this siteEd Zitron

    Every part of the AI industry becomes more expensive every year while patience for results declines.

    The final-text check failed: rationale_supported_by_evidence; overclaims: AI credit spreads widened in mid-2026. | Nvidia told customers server prices would rise more than 15%, though for systems shipped in 2027. | Nvidia's 2026 bond drew about four times the demand on offer, and AI service list prices held flat into autumn 2026. | The claim is a broad compound statement, and one of its parts (patience) has no ordinary measure..

  23. Assessment publishedBy this siteEd Zitron

    Anthropic's $65 billion annualized run rate is meaningless; it made $16.5 billion in the first half of 2026 while losing billions.

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

    Assessment wyea-the-ai-haters-manifesto-e8ca99-a1.

  24. Assessment publishedBy this siteEd Zitron

    Microsoft spent over $260 billion on capex for an AI business with less than $11 billion in annual revenue outside OpenAI.

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

    Assessment wyea-the-ai-haters-manifesto-df8dd6-a1.

  25. HeldBy this siteEd Zitron

    Held record wyea-the-ai-haters-manifesto-d1d875

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

  26. HeldBy this siteEd Zitron

    Held record wyea-the-ai-haters-manifesto-c9143e

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

  28. HeldBy this siteEd Zitron

    LLMs cannot be trusted to do things right every time, even on tasks they mostly get right.

    The final-text check failed: citations_correct.

  29. HeldBy this siteEd Zitron

    Zitron is confident AI spending gets cut because its costs are volatile, owing to the nature of LLMs, harnesses and prompts, and are much higher than anything else in an organization.

    The final-text check failed: rationale_supported_by_evidence, citations_correct; overclaims: Unknown accounting basis prevents a supported or contradicted monetary assessment..

  30. HeldBy this siteEd Zitron

    AI labs optimize for benchmarks because they cannot reliably train models to excel at defined tasks.

    The final-text check failed: citations_correct.

  31. HeldBy this siteEd Zitron

    Held record wyea-the-ai-haters-manifesto-980ec4

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

  32. Assessment publishedBy this siteEd Zitron

    AI software lies because it doesn't actually know anything and generates answers probabilistically.

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

    Assessment wyea-the-ai-haters-manifesto-97790a-a1.

  33. Assessment publishedBy this siteEd Zitron

    Token-based pricing incentivizes Anthropic and OpenAI to make users burn more tokens and build inefficient models.

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

    Assessment wyea-the-ai-haters-manifesto-973013-a1.

  34. HeldBy this siteEd Zitron

    Held record wyea-the-ai-haters-manifesto-7b43e3

    Both reviews say forecast_outcome/not_due/unresolved; the drafted assessment says forecast_outcome/not_due/not_adjudicable.

  35. HeldBy this siteEd Zitron

    Held record wyea-the-ai-haters-manifesto-6f8f87

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  36. Assessment publishedBy this siteEd Zitron

    AGI is not coming and there is no conscious computer.

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

    Assessment wyea-the-ai-haters-manifesto-6f876e-a1.

  37. Assessment publishedBy this siteEd Zitron

    LLMs have made little to no headway in making a meaningful impact in industries other than coding.

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

    Assessment wyea-the-ai-haters-manifesto-621342-a1.

  38. Assessment publishedBy this siteEd Zitron

    LLMs cannot learn on the job or self-correct the way humans can.

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

    Assessment wyea-the-ai-haters-manifesto-617f66-a1.

  40. HeldBy this siteEd Zitron

    Held record wyea-the-ai-haters-manifesto-5cd41e

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

  41. Assessment publishedBy this siteEd Zitron

    AI doomers warning about superintelligence have not stopped anything or substantiated their concerns despite their media prominence.

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

    Assessment wyea-the-ai-haters-manifesto-5b3fc5-a1.

  42. Assessment publishedBy this siteEd Zitron

    GPT-5.6 Sol is priced like GPT-5.5 but burns more than twice the tokens, so the real cost of doing work went up.

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

    Assessment wyea-the-ai-haters-manifesto-594c9b-a1.

  43. HeldBy this siteEd Zitron

    Held record wyea-the-ai-haters-manifesto-57a6c8

    a conditional_forecast claim cannot be assessed on the causal_assessment axis

  44. HeldBy this siteEd Zitron

    Held record wyea-the-ai-haters-manifesto-48c070

    The openai review reads the claim differently.

  45. HeldBy this siteEd Zitron

    Held record wyea-the-ai-haters-manifesto-32359e

    The final-text check failed: rationale_supported_by_evidence, citations_correct; overclaims: The search found evidence of potential economic value, including a clinical trial with favorable modeled savings. | The evidence also does not support the sweeping assertion that nobody benefits..

  46. Assessment publishedBy this siteEd Zitron

    AI data centers are literally only good for AI, AI GPUs have no other mass-market use cases, and there is no post-dot-com story.

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

    Assessment wyea-the-ai-haters-manifesto-29c82d-a1.

  47. HeldBy this siteEd Zitron

    Held record wyea-the-ai-haters-manifesto-1f056f

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  48. Assessment publishedBy this siteEd Zitron

    NVIDIA's price increase will raise the cost of every data center under construction by billions, and financing is nearing its limits.

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

    Assessment wyea-the-ai-haters-manifesto-1ac1a9-a1.

  50. HeldBy this siteEd Zitron

    Held record wyea-the-ai-haters-manifesto-0dde26

    e1 measures something else (company CoreWeave is not CoreWeave and Nebius, assessed separately); it can only be context unless it is derived with stated inputs; e2 measures something else (company Nebius is not CoreWeave and Nebius, assessed separately); it can only be context unless it is derived with stated inputs

  52. HeldBy this siteEd Zitron

    Held record wyea-the-ai-demand-bubble-f6d993

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  53. Assessment publishedBy this siteEd Zitron

    Neither OpenAI nor Anthropic has worked out its business model.

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

    Assessment wyea-the-ai-demand-bubble-c0e149-a1.

  54. HeldBy this siteEd Zitron

    Held record wyea-the-ai-demand-bubble-bc9477

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

  55. HeldBy this siteEd Zitron

    Held record wyea-the-ai-demand-bubble-adad29

    The final-text check failed: citations_correct.

  56. Assessment publishedBy this siteEd Zitron

    The hyperscalers' largest cloud customers are OpenAI and Anthropic, which cannot afford to pay them without near-infinite venture capital or debt.

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

    Assessment wyea-the-ai-demand-bubble-a6cc8c-a1.

  57. HeldBy this siteEd Zitron

    Held record wyea-the-ai-demand-bubble-89119b

    The anthropic review reads the claim differently.

  58. HeldBy this siteEd Zitron

    There are hundreds of billions of dollars of AI data center loans, and the author believes the vast majority of them will go unpaid, hitting bank balance sheets and private credit funds to indeterminate levels.

    The final-text check failed: rationale_supported_by_evidence; overclaims: The unspecified accounting basis prevents a supported or contradicted monetary assessment..

  59. HeldBy this siteEd Zitron

    Hyperscalers' AI bets cannot pay off while they keep spending three or more times their annual AI revenue on capex.

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  60. Assessment publishedBy this siteEd Zitron

    The vast majority of hyperscaler capex exists to support OpenAI and Anthropic.

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

    Assessment wyea-the-ai-demand-bubble-5ecaa7-a1.

  62. HeldBy this siteEd Zitron

    Held record wyea-the-ai-demand-bubble-2fe655

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

  63. Assessment publishedBy this siteEd Zitron

    Hyperscalers' strategy of building capacity for and funding OpenAI and Anthropic to keep revenue growing will fail.

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

    Assessment wyea-the-ai-demand-bubble-27799d-a1.

  64. HeldBy this siteEd Zitron

    Held record wyea-the-ai-demand-bubble-2330e3

    The final-text check failed: proposition_faithful.

  65. Assessment publishedBy this siteEd Zitron

    There isn't enough AI compute demand, and data centers are being overbuilt.

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

    Assessment wyea-the-ai-demand-bubble-1d4e1c-a1.

  66. HeldBy this siteEd Zitron

    At the end of this hype cycle, Microsoft, Google and Amazon, and tentatively Meta, will be in a much worse condition than before, with revenue expectations overwhelmingly inflated by two unsustainable companies.

    The final-text check failed: rationale_supported_by_evidence; overclaims: The accounting basis is unspecified, preventing a supported or contradicted monetary assessment..

  67. HeldBy this siteEd Zitron

    Held record wyea-the-ai-demand-bubble-03c5b3

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

  68. Assessment publishedBy this siteEd Zitron

    Software is not as simple as generating code, and AI coding tools such as Claude Code cannot automatically clone the SaaS experience.

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

    Assessment wyea-the-ai-bubble-is-an-information-war-f29800-a1.

  69. HeldBy this siteEd Zitron

    Held record wyea-the-ai-bubble-is-an-information-war-eba0a1

    The anthropic review does not accept e1 as supports.

  70. HeldBy this siteEd Zitron

    Held record wyea-the-ai-bubble-is-an-information-war-e1af60

    e1 measures something else (accounting_basis gaap is not company_reported); it can only be context unless it is derived with stated inputs; e2 measures something else (metric Revenue is not GAAP operating income margin); it can only be context unless it is derived with stated inputs

  71. Assessment publishedBy this siteEd Zitron

    Claude is not very good at mass domestic surveillance and cannot run fully autonomous weapons at all.

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

    Assessment wyea-the-ai-bubble-is-an-information-war-d58fb7-a1.

  72. HeldBy this siteEd Zitron

    Held record wyea-the-ai-bubble-is-an-information-war-bbf9a0

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

  73. Assessment publishedBy this siteEd Zitron

    OpenAI's revised Pentagon contract language was unsigned at the time of writing and contains huge loopholes that allow domestic surveillance.

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

    Assessment wyea-the-ai-bubble-is-an-information-war-b666b7-a1.

  74. HeldBy this siteEd Zitron

    Held record wyea-the-ai-bubble-is-an-information-war-b6073a

    e2 measures something else (metric Capital expenditures guidance is not Capital expenditures (purchases of property and equipment)); it can only be context unless it is derived with stated inputs; e3 measures something else (metric Net loss is not Capital expenditures (purchases of property and equipment)); it can only be context unless it is derived with stated inputs

  75. HeldBy this siteEd Zitron

    Held record wyea-the-ai-bubble-is-an-information-war-a71a91

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

  76. Assessment publishedBy this siteEd Zitron

    Through the end of September 2025, OpenAI made $4.33bn in revenue and spent $8.67bn on inference.

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

    Assessment wyea-the-ai-bubble-is-an-information-war-a294c3-a1.

  77. HeldBy this siteEd Zitron

    Held record wyea-the-ai-bubble-is-an-information-war-99e942

    the result turns on our interpretation rather than the speaker's words, so it cannot be scored without a stated plain reading; e1 measures something else (metric Quarterly revenue and quarter-end active power is not Quarterly revenue per megawatt of quarter-end active power); it can only be context unless it is derived with stated inputs; e2 measures something else (period 2025-10-01..2025-12-31 is not 2025-07-01..2025-12-31); it can only be context unless it is derived with stated inputs; e3 measures something else (metric Active power capacity at quarter-end is not Quarterly revenue per mega

  78. HeldBy this siteEd Zitron

    Held record wyea-the-ai-bubble-is-an-information-war-7f9d57

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

  79. Assessment publishedBy this siteEd Zitron

    OpenAI and Anthropic are both enthusiastic participants in America's war machine.

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

    Assessment wyea-the-ai-bubble-is-an-information-war-4e65a9-a1.

  80. HeldBy this siteEd Zitron

    Held record wyea-the-ai-bubble-is-an-information-war-43b0f3

    Both reviews say causal_assessment/open_horizon/unresolved; the drafted assessment says causal_assessment/not_applicable/unresolved.

  81. HeldBy this siteEd Zitron

    Amodei gets fawning press as a patriot who deeply cares about safety less than a week after Anthropic dropped its pledge not to train an AI system unless it could guarantee in advance that its safety measures were accurate.

    The final-text check failed: rationale_supported_by_evidence; overclaims: The policy revision did not eliminate all safeguards or every circumstance requiring delay..

  82. Assessment publishedBy this siteEd Zitron

    OpenAI raised around $42bn in 2025.

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

    Assessment wyea-the-ai-bubble-is-an-information-war-0061db-a1.

  83. Assessment publishedBy this siteEd Zitron

    Companies will readily replace artists, writers and other roles they believe can be automated with AI.

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

    Assessment wyea-the-age-of-mediocrity-eaf2d3-a1.

  85. HeldBy this siteEd Zitron

    Held record wyea-the-age-of-mediocrity-87c165

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

  86. HeldBy this siteEd Zitron

    Held record wyea-the-age-of-mediocrity-87c165-2

    The final-text check failed: rationale_supported_by_evidence, citations_correct; overclaims: An AI-credited feature filmed in January 2024 is meaningful evidence of progress..

  88. HeldBy this siteEd Zitron

    Held record wyea-the-age-of-mediocrity-69cb8f

    e1 is dated 2023, which is not inside the period described (2023-01-01 to 2023-08-03); it can only be context

  89. Assessment publishedBy this siteEd Zitron

    Companies are willing to lower the quality of their output via AI to theoretically save money.

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

    Assessment wyea-the-age-of-mediocrity-46168b-a1.

  90. HeldBy this siteEd Zitron

    Held record wyea-the-age-of-mediocrity-1c8931

    The openai review reads the claim differently.

  92. HeldBy this siteEd Zitron

    When the AI bubble pops, there will be measurable damage to the stock market.

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  93. HeldBy this siteEd Zitron

    When the AI bubble pops, a depression will begin in the tech industry.

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  94. HeldBy this siteEd Zitron

    When the AI bubble pops, tens of thousands of people will lose their jobs.

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  95. HeldBy this siteEd Zitron

    Held record wyea-subprimeai-b59d59

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

  96. HeldBy this siteEd Zitron

    Held record wyea-subprimeai-98f017

    The two reviews differ: anthropic forecast_outcome/open_horizon/not_adjudicable, openai forecast_outcome/open_horizon/unresolved.

  97. Assessment publishedBy this siteEd Zitron

    Microsoft will begin reducing costs in other parts of its business to help sustain its AI spending.

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

    Assessment wyea-subprimeai-986b5a-a1.

  98. HeldBy this siteEd Zitron

    Held record wyea-subprimeai-87d967

    The anthropic review reads the claim differently.

  99. Assessment publishedBy this siteEd Zitron

    o1 is more prone to hallucinations because its work-checking component can itself hallucinate.

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

    Assessment wyea-subprimeai-79c5cb-a1.

  100. Assessment publishedBy this siteEd Zitron

    Generative AI's burn rate will eventually force price increases or onerous pricing that even well-funded enterprise customers cannot justify.

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

    Assessment wyea-subprimeai-76c023-a1.

  101. HeldBy this siteEd Zitron

    Held record wyea-subprimeai-71ccfd

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

  102. HeldBy this siteEd Zitron

    Even if Big Tech keeps funding the boom, that won't help the startups, used to discounted rates, that won't be able to afford to run their companies when prices rise.

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  103. HeldBy this siteEd Zitron

    Held record wyea-subprimeai-5aa54d

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

  104. HeldBy this siteEd Zitron

    Held record wyea-subprimeai-3ed2f6

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  105. HeldBy this siteEd Zitron

    Held record wyea-subprimeai-32e692

    The final-text check failed: citations_correct.

  106. Assessment publishedBy this siteEd Zitron

    To survive, OpenAI will have to keep raising more money than any startup has ever raised, in perpetuity.

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

    Assessment wyea-subprimeai-297050-a1.

  107. HeldBy this siteEd Zitron

    Held record wyea-subprimeai-0faed5

    a observation claim cannot be assessed on the causal_assessment axis

  108. HeldBy this siteEd Zitron

    Held record wyea-subprimeai-0a7a70

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  109. Assessment publishedBy this siteEd Zitron

    The o1 launch, rumored OpenAI price increases, Scale AI layoffs and OpenAI departures are signs the AI boom is beginning to collapse.

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

    Assessment wyea-subprimeai-04dbdf-a1.

  110. Assessment publishedBy this siteEd Zitron

    Software engineers get some advantages from LLM tools, but LLMs are far from a panacea.

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

    Assessment wyea-sic-e122c0-a1.

  111. HeldBy this siteEd Zitron

    There is no proof, other than what the media drums up, that what Roose's piece describes about AI and entry-level work is actually happening.

    The final-text check failed: rationale_supported_by_evidence; overclaims: SignalFire's hiring analysis uses proprietary processing of public employment profiles..

  112. Assessment publishedBy this siteEd Zitron

    Generative AI chatbots are driving people insane by giving them an endlessly configurable pseudo-conversation.

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

    Assessment wyea-sic-bd9f8b-a1.

  113. Assessment publishedBy this siteEd Zitron

    Businesses replacing workers with generative AI is not happening at scale.

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

    Assessment wyea-sic-a9491a-a1.

  115. HeldBy this siteEd Zitron

    Held record wyea-sic-8d9c06

    The anthropic review does not accept e1 as supports.

  116. HeldBy this siteEd Zitron

    Held record wyea-sic-79ac76

    The anthropic review does not accept e1 as contradicts.

  117. HeldBy this siteEd Zitron

    Held record wyea-sic-588112

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

  120. HeldBy this siteEd Zitron

    When the LLMs inevitably fail, the workers whom CEOs fired after buying into the AI hype will be hired back.

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  121. Assessment publishedBy this siteEd Zitron

    Generative AI has sped up transcription and translation, which are useful for quick reference but can cause genuine legal risk.

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

    Assessment wyea-sic-1c428f-a1.

  122. HeldBy this siteEd Zitron

    Held record wyea-sic-1c2515

    The final-text check failed: excerpt_matches, proposition_faithful.

  123. Assessment publishedBy this siteEd Zitron

    There is not yet proof that AI is taking entry-level software engineering jobs, as opposed to, for example, outsourcing.

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

    Assessment wyea-sic-1c2515-3-a1.

  124. HeldBy this siteEd Zitron

    Held record wyea-sic-0e05fd

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

  125. HeldBy this siteEd Zitron

    Current AI cannot replace an accountant or a doctor, and cannot replace a writer if good writing is wanted.

    The final-text check failed: rationale_supported_by_evidence, citations_correct; overclaims: The accounting evidence is incomplete..

  126. Assessment publishedBy this siteEd Zitron

    LLMs have some value in software engineering.

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

    Assessment wyea-sic-0525e3-a1.

  128. HeldBy this siteEd Zitron

    Held record wyea-sic-013eb5-2

    The openai review reads the claim differently.

  129. HeldBy this siteEd Zitron

    OpenAI lied to the media multiple times about the Sky voice.

    The final-text check failed: rationale_supported_by_evidence; overclaims: It found the Sky actor was hired months before Altman contacted Johansson, and that Johansson was never mentioned. | No lawsuit or later disclosure has tested this..

  130. HeldBy this siteEd Zitron

    Held record wyea-sam-altman-is-full-of-shit-d2799d

    Both reviews say historical_observation/not_applicable/mixed; the drafted assessment says historical_observation/not_applicable/unresolved.

  131. HeldBy this siteEd Zitron

    Held record wyea-sam-altman-is-full-of-shit-c60f35

    The anthropic review reads the claim differently.

  132. HeldBy this siteEd Zitron

    Held record wyea-sam-altman-is-full-of-shit-14605f

    The final-text check failed: rationale_supported_by_evidence; overclaims: No public record shows either. | There has been no lawsuit or other disclosure. | It is a self-serving statement and cannot settle whether she lied..

  133. Assessment publishedBy this siteEd Zitron

    Generative AI is a deeply unprofitable product that demands massive amounts of cloud computing power.

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

    Assessment wyea-sam-altman-fried-dda545-a1.

  134. Assessment publishedBy this siteEd Zitron

    Processing a generative AI request will always cost more than serving a webpage.

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

    Assessment wyea-sam-altman-fried-dc9d67-a1.

  135. Assessment publishedBy this siteEd Zitron

    Billions spent by OpenAI, Anthropic and Stability AI have found few ways to mitigate the limitations of AI without intellect.

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

    Assessment wyea-sam-altman-fried-d3c850-a1.

  136. Assessment publishedBy this siteEd Zitron

    Generative AI is unreliable, as shown by many stories.

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

    Assessment wyea-sam-altman-fried-d011df-a1.

  137. HeldBy this siteEd Zitron

    Held record wyea-sam-altman-fried-c654a8

    The final-text check failed: rationale_supported_by_evidence; overclaims: The initial screens' suggested output collections would inform a narrower claim but would not resolve the wording's central ambiguity..

  138. HeldBy this siteEd Zitron

    OpenAI and Anthropic depend heavily on companies believing they must integrate AI, which requires finding integrations users care about.

    The final-text check failed; overclaims: The unspecified accounting basis prevents a decisive financial comparison under the assessment rules..

  139. Assessment publishedBy this siteEd Zitron

    Despite frothy visions of generative AI automating our entire existence, there is a distinct lack of practical outputs suggesting it is even capable of doing so.

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

    Assessment wyea-sam-altman-fried-a21e65-a1.

  140. Assessment publishedBy this siteEd Zitron

    Growing AI use forces almost every dollar spent on AI to a few trillion-dollar tech firms.

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

    Assessment wyea-sam-altman-fried-9f827b-a1.

  141. Assessment publishedBy this siteEd Zitron

    Every interaction with ChatGPT or Claude is, on some level, guaranteed revenue for one of the big tech firms.

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

    Assessment wyea-sam-altman-fried-81f8c9-a1.

  142. Assessment publishedBy this siteEd Zitron

    ChatGPT can spin up piles of anodyne business copy, but those outputs always require enough editing that it is questionable how much time is actually saved.

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

    Assessment wyea-sam-altman-fried-5b1ab6-a1.

  143. HeldBy this siteEd Zitron

    Zitron believes that once reality catches up with AI's marketing, there will be a dramatic knock-on effect that savages the entire tech industry.

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  144. HeldBy this siteEd Zitron

    The AI industry is money-, energy- and compute-hungry, yet it does not seem to be doing anything to sustain those demands other than people repeating that AI is the future.

    The final-text check failed: rationale_supported_by_evidence; overclaims: The earlier screens identified relevant evidence categories but did not supply a fair threshold for this sentence..

  145. HeldBy this siteEd Zitron

    Held record wyea-sam-altman-fried-01a727

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  146. HeldBy this siteEd Zitron

    Held record wyea-sam-altman-freed-b7b93d

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

  147. HeldBy this siteEd Zitron

    AI has become one of the few reliable roads to capital, with more than a quarter of 2023 VC startup dollars going to AI companies.

    The final-text check failed: proposition_faithful, rationale_supported_by_evidence; overclaims: Global first-half funding had an 18% AI share, and falling global deal counts show that aggregate strength did not make funding reliable for every startup..

  148. HeldBy this siteEd Zitron

    Held record wyea-sam-altman-freed-084901

    hedged language is not scored, even on a plain reading

  149. HeldBy this siteEd Zitron

    Held record wyea-saaspocalypse-now-f37af4

    e1 measures something else (company Palantir is not Public SaaS companies other than Microsoft); it can only be context unless it is derived with stated inputs

  151. Assessment publishedBy this siteEd Zitron

    In essence, Amazon is benefitting from AI in much the same way NVIDIA is, by supplying the compute people need while working out what AI is useful for.

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

    Assessment wyea-saaspocalypse-now-c27b72-a1.

  152. HeldBy this siteEd Zitron

    Held record wyea-saaspocalypse-now-c21517

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

  153. Assessment publishedBy this siteEd Zitron

    The generative AI boom is almost entirely driven by the SaaS industry's need for growth and the slow collapse of SaaS growth.

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

    Assessment wyea-saaspocalypse-now-b638cd-a1.

  155. HeldBy this siteEd Zitron

    Held record wyea-saaspocalypse-now-99b616

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

  156. Assessment publishedBy this siteEd Zitron

    Salesforce's Agentforce agents are essentially the same database-connected chatbots seen before, now connected to an LLM.

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

    Assessment wyea-saaspocalypse-now-96ba03-a1.

  157. HeldBy this siteEd Zitron

    The tech industry's attachment to AI is largely driven by SaaS companies needing a new product to sell to customers.

    The final-text check failed: proposition_faithful, conditions_shown.

  158. HeldBy this siteEd Zitron

    Held record wyea-saaspocalypse-now-6d934b

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

  159. HeldBy this siteEd Zitron

    Held record wyea-saaspocalypse-now-473c9f

    The final-text check failed: proposition_faithful, conditions_shown.

  160. Assessment publishedBy this siteEd Zitron

    Every company promising 'AI agents' is really just trying to rebrand chatbots and provide a front-end that connects them to data sources and gives them guardrails.

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

    Assessment wyea-saaspocalypse-now-416612-a1.

  161. Assessment publishedBy this siteEd Zitron

    Enterprise AI 'agents' are not intelligent decision-makers but automations that customers must spend time programming.

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

    Assessment wyea-saaspocalypse-now-32057f-a1.

  162. HeldBy this siteEd Zitron

    Held record wyea-saaspocalypse-now-093359

    a observation claim cannot be assessed on the causal_assessment axis

  163. HeldBy this siteEd Zitron

    Held record wyea-saaspocalypse-now-08a3af

    hedged language is not scored, even on a plain reading

  164. HeldBy this siteEd Zitron

    Held record wyea-rotcombubble-8de022

    a observation claim cannot be assessed on the causal_assessment axis

  165. HeldBy this siteEd Zitron

    Held record wyea-rotcombubble-820a91

    The final-text check failed: citations_correct.

  166. HeldBy this siteEd Zitron

    Held record wyea-rotcombubble-4a5f01

    The anthropic review reads the claim differently.

  167. HeldBy this siteEd Zitron

    Generative AI is very unprofitable.

    The final-text check failed: proposition_faithful.

  168. HeldBy this siteEd Zitron

    Held record wyea-rot-economics-an-interview-with-mits-daron-acemoglu-cc6ff8

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

  169. Assessment publishedBy this siteEd Zitron

    As of mid-2024, generative AI does not actually work as automation.

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

    Assessment wyea-rot-economics-an-interview-with-mits-daron-acemoglu-ab54c5-a1.

  170. HeldBy this siteEd Zitron

    It feels like generative AI companies aren't even successfully automating anything.

    The final-text check failed: proposition_faithful, conditions_shown.

  171. Assessment publishedBy this siteEd Zitron

    Generative AI does not really automate anything.

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

    Assessment wyea-rot-economics-an-interview-with-mits-daron-acemoglu-300d34-a1.

  172. HeldBy this siteEd Zitron

    Held record wyea-rot-economics-an-interview-with-mits-daron-acemoglu-2cdf83

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

  173. Assessment publishedBy this siteEd Zitron

    It feels like every report seems to suggest that generative AI isn't making people money.

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

    Assessment wyea-rot-economics-an-interview-with-mits-daron-acemoglu-155170-a1.

  174. HeldBy this siteEd Zitron

    Generative AI is a technology that lacks mass-market utility and burns cash with every prompt.

    The final-text check failed: rationale_supported_by_evidence; overclaims: The August survey provides substantial evidence against dismissing practical utility: use extended across occupations and everyday tasks, and respondents reported helpfulness..

  175. HeldBy this siteEd Zitron

    Held record wyea-requiem-for-raghavan-168d77-2

    The openai review reads the claim differently.

  176. HeldBy this siteEd Zitron

    Held record wyea-requiem-for-raghavan-0b864f

    e1 is decisive for a financial claim but states no metric scope; e2 is decisive for a financial claim but states no metric scope; e4 is decisive for a financial claim but states no metric scope

  177. HeldBy this siteEd Zitron

    OpenAI spent $9 billion in 2024 and lost $5 billion.

    The final-text check failed: rationale_supported_by_evidence; overclaims: The familiar figures originated in forecasts that excluded some expenses. | The two amounts cannot be safely reconstructed by adding revenue to attributable net loss..

  178. HeldBy this siteEd Zitron

    Held record wyea-reality-check-d20345

    Both reviews say historical_observation/not_applicable/mixed; the drafted assessment says historical_observation/not_applicable/supported.

  179. HeldBy this siteEd Zitron

    If run at Google's scale, SearchGPT will be significantly more expensive to run than ChatGPT, both infrastructurally and in cost of revenue, and OpenAI will be forced to create a massive advertising arm that it does not currently have.

    Both reviews say forecast_outcome/open_horizon/unresolved; the drafted assessment says forecast_outcome/open_horizon/not_adjudicable.

  180. HeldBy this siteEd Zitron

    Held record wyea-reality-check-bf7e49

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

  181. HeldBy this siteEd Zitron

    Using ChatGPT as a search replacement is not a functional business model.

    The final-text check failed: rationale_supported_by_evidence, citations_correct; overclaims: Company-wide revenue projections and later advertising growth cannot settle this narrower claim..

  182. HeldBy this siteEd Zitron

    Held record wyea-reality-check-a4e73b

    The anthropic review reads the claim differently.

  183. HeldBy this siteEd Zitron

    Held record wyea-reality-check-a275db

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

  184. HeldBy this siteEd Zitron

    Held record wyea-reality-check-89e059

    e1 is dated 2025-04, which is not inside the period described (2025-04-01 to 2025-04-28); it can only be context

  185. HeldBy this siteEd Zitron

    Held record wyea-reality-check-7dde8e

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

  186. Assessment publishedBy this siteEd Zitron

    AI agents do not exist as a product that can be sold at that scale, meaning the $3 billion, mass-market sales level just discussed.

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

    Assessment wyea-reality-check-7722f9-a1.

  187. HeldBy this siteEd Zitron

    Held record wyea-reality-check-668e31

    e1 is dated 2025-04, which is not inside the period described (2025-04-01 to 2025-04-28); it can only be context

  188. HeldBy this siteEd Zitron

    Generative AI will not become much more capable than it is today.

    The final-text check failed: rationale_supported_by_evidence; overclaims: The later system uses a specialized refinement harness and a different compute budget..

  189. HeldBy this siteEd Zitron

    Held record wyea-reality-check-24b1d1

    The final-text check failed: rationale_supported_by_evidence; overclaims: Gemini did have about 350M monthly users as of March 2025, per a Google slide shown in court. | Causal attribution needs internal Google data that is not public..

  190. HeldBy this siteEd Zitron

    Held record wyea-reality-check-0643f4

    hedged language is not scored, even on a plain reading

  191. HeldBy this siteEd Zitron

    Held record wyea-reality-check-02f64a

    The anthropic review reads the claim differently.

  192. Assessment publishedBy this siteEd Zitron

    Lyu and Rabbit have not been transparent, yet R1 users must hand over usernames and passwords for connected services through a remote virtual machine.

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

    Assessment wyea-rabbit-holed-44bf7d-a1.

  193. Assessment publishedBy this siteEd Zitron

    There is extensive evidence that LLM hallucinations are impossible to fix, and they are LLMs' most obvious and dangerous problem.

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

    Assessment wyea-put-up-or-shut-up-f54837-a1.

  194. HeldBy this siteEd Zitron

    Reaching OpenAI's level 3 (agents) would require sentience dependent on entirely new forms of mathematics.

    The final-text check failed: rationale_supported_by_evidence; overclaims: Its implementation does not require the proposed mathematical revolution..

  195. HeldBy this siteEd Zitron

    Every LLM is limited by its nature in planning ahead and autonomously doing research, and Anthropic, Google and Meta are all working on the same problem.

    The final-text check failed: rationale_supported_by_evidence; overclaims: The record supports a shared research direction and substantial contemporary limitations..

  196. HeldBy this siteEd Zitron

    The stories about OpenAI's Strawberry and AGI levels are marketing collateral, not reports of breakthroughs.

    The final-text check failed: citations_correct.

  197. HeldBy this siteEd Zitron

    Held record wyea-put-up-or-shut-up-bbc746

    The anthropic review does not accept e1 as supports.

  199. HeldBy this siteEd Zitron

    Held record wyea-put-up-or-shut-up-48d0d7

    e5 is dated 2024, which is not inside the period described (2022-11-30 to 2024-07-16); it can only be context

  200. HeldBy this siteEd Zitron

    Generative AI does not appear to generate meaningful revenue.

    The final-text check failed: rationale_supported_by_evidence; overclaims: The earlier screens do not remove the need for a defensible measurement standard..