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

Changes, page 16 of 37

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 (7383)

Page 16 of 37: changes 3001 to 3200 of 7383, newest first. Dates are when the change was recorded.

  2. HeldBy this siteGary Marcus

    Held record marcus-the-race-for-ai-supremacy-is-over-2b5d63

    e1 is dated 2025-01, which is not inside the period described (2024-12-01 to 2025-01-26); it can only be context

  4. Assessment publishedBy this siteGary Marcus

    Congress should investigate how Meta's open-sourcing of its LLMs helped China catch up.

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

    Assessment marcus-the-race-for-ai-supremacy-is-over-219350-a1.

  5. HeldBy this siteGary Marcus

    Held record marcus-the-race-for-ai-supremacy-is-over-1e522d

    The anthropic screen reads the claim differently.

  6. Assessment publishedBy this siteGary Marcus

    The White House's heavy reliance on Altman may prove a mistake in hindsight.

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

    Assessment marcus-the-race-for-ai-supremacy-is-over-1a687e-a1.

  7. HeldBy this siteGary Marcus

    Held record marcus-the-race-for-ai-supremacy-is-over-15fb68

    The openai review does not accept e4 as supports.

  8. Assessment publishedBy this siteGary Marcus

    US export controls accidentally improved China's AI technical capabilities by forcing efficiency innovations.

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

    Assessment marcus-the-race-for-ai-supremacy-is-over-15c640-a1.

  9. HeldBy this siteGary Marcus

    Held record marcus-the-race-for-ai-supremacy-is-over-03100c

    e2 is dated 2025-01, which is not inside the period described (2025-01-01 to 2025-01-26); it can only be context

  11. HeldBy this siteGary Marcus

    Held record marcus-the-race-between-positive-and-negative-ecb86d

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

  12. Assessment publishedBy this siteGary Marcus

    Existing laws are unlikely to make Microsoft or OpenAI pay for wrong chatbot advice.

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

    Assessment marcus-the-race-between-positive-and-negative-c3c8b6-a1.

  13. Assessment publishedBy this siteGary Marcus

    If GenAI disinformation flips the 2024 election, GenAI companies are unlikely to be held seriously responsible.

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

    Assessment marcus-the-race-between-positive-and-negative-bfa0d7-a1.

  14. HeldBy this siteGary Marcus

    Held record marcus-the-race-between-positive-and-negative-a17ab6

    The anthropic screen doubts the attribution.

  15. Assessment publishedBy this siteGary Marcus

    OpenAI asked the British government to exempt it from copyright laws.

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

    Assessment marcus-the-race-between-positive-and-negative-799f03-a1.

  16. HeldBy this siteGary Marcus

    Held record marcus-the-race-between-positive-and-negative-6251b9

    The anthropic review does not accept e1 as supports.

  17. Assessment publishedBy this siteGary Marcus

    AI companies get rich while society pays the costs.

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

    Assessment marcus-the-race-between-positive-and-negative-3e8256-a1.

  18. HeldBy this siteGary Marcus

    Held record marcus-the-race-between-positive-and-negative-230c51

    The openai screen reads the claim differently.

  20. Assessment publishedBy this siteGary Marcus

    Harmful uses of generative AI (fake books, ad-farm websites, election misinformation, polluted science) are proliferating rapidly.

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

    Assessment marcus-the-race-between-positive-and-negative-0c70cf-a1.

  21. Assessment publishedBy this siteGary Marcus

    One study indicates code quality is going down amid AI-assisted coding.

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

    Assessment marcus-the-race-between-positive-and-negative-0be6dd-a1.

  23. HeldBy this siteGary Marcus

    Held record marcus-the-possibly-coming-ai-backlash-and-e6fde1

    The anthropic screen doubts the attribution.

  24. HeldBy this siteGary Marcus

    Held record marcus-the-possibly-coming-ai-backlash-and-dbd652

    The anthropic screen doubts the attribution.

  25. HeldBy this siteGary Marcus

    Held record marcus-the-possibly-coming-ai-backlash-and-a4b88f

    The anthropic screen doubts the attribution.

  26. HeldBy this siteGary Marcus

    Held record marcus-the-possibly-coming-ai-backlash-and-9280d1

    The openai screen reads the claim differently.

  27. HeldBy this siteGary Marcus

    Held record marcus-the-possibly-coming-ai-backlash-and-8a2b19

    The anthropic screen doubts the attribution.

  29. Assessment publishedBy this siteGary Marcus

    The most important question about AI is how to maintain and grow a just society in the age of AI.

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

    Assessment marcus-the-pope-gets-it-caad16-a1.

  30. Assessment publishedBy this siteGary Marcus

    The most important question about AI is not a technical question.

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

    Assessment marcus-the-pope-gets-it-9e78ab-a1.

  31. HeldBy this siteGary Marcus

    Held record marcus-the-pope-gets-it-1a9134

    The anthropic screen reads the claim differently.

  32. Assessment publishedBy this siteGary Marcus

    Geoffrey Hinton should know better than to treat what LLM researchers create as beings.

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

    Assessment marcus-the-pope-appears-to-understand-ai-fbe983-a1.

  35. Assessment publishedBy this siteGary Marcus

    LLMs are interactive fiction trained to predict the language of actual beings.

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

    Assessment marcus-the-pope-appears-to-understand-ai-b85b53-a1.

  36. Assessment publishedBy this siteGary Marcus

    LLMs differ immensely from humans: LLMs effectively memorize the internet, while humans build mental models through experience of the world.

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

    Assessment marcus-the-pope-appears-to-understand-ai-b7b0d5-a1.

  37. Assessment publishedBy this siteGary Marcus

    It does not follow that LLMs are like people just because they say things people say.

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

    Assessment marcus-the-pope-appears-to-understand-ai-688f63-a1.

  38. Assessment publishedBy this siteGary Marcus

    However rich it is, LLM mimicry of human output proves very little about consciousness, because consciousness is about internal states.

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

    Assessment marcus-the-pope-appears-to-understand-ai-5b607f-a1.

  39. Assessment publishedBy this siteGary Marcus

    Claude's outputs come from a form of mimicry and are not reports of genuine internal states.

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

    Assessment marcus-the-pope-appears-to-understand-ai-43e78e-a1.

  40. Assessment publishedBy this siteGary Marcus

    LLM researchers are not creating beings.

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

    Assessment marcus-the-pope-appears-to-understand-ai-3c2307-a1.

  41. Assessment publishedBy this siteGary Marcus

    OpenAI could end up being seen as the WeWork of AI.

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

    Assessment marcus-the-openai-plot-thickens-fb8a2e-a1.

  42. HeldBy this siteGary Marcus

    Held record marcus-the-openai-plot-thickens-8c6c7e

    The final-text check failed: citations_correct; overclaims: The report dates Saunders's departure to February and Kokotajlo's to April, when both were gone. It identifies their superalignment and governance work, establishing two additional safety-related resignations..

  43. HeldBy this siteGary Marcus

    Held record marcus-the-openai-plot-thickens-848571

    e2 is decisive for a financial claim but states no metric scope; e3 is decisive for a financial claim but states no metric scope; e4 is decisive for a financial claim but states no metric scope; e5 is decisive for a financial claim but states no metric scope; e6 is decisive for a financial claim but states no metric scope

  44. HeldBy this siteGary Marcus

    Held record marcus-the-openai-plot-thickens-627fad

    The openai screen reads the claim differently.

  45. HeldBy this siteGary Marcus

    Held record marcus-the-openai-plot-thickens-3977e6

    The openai screen reads the claim differently.

  46. Assessment publishedBy this siteGary Marcus

    Many key OpenAI people have left, most of them voluntarily.

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

    Assessment marcus-the-openai-plot-thickens-2102c0-a1.

  47. Assessment publishedBy this siteGary Marcus

    OpenAI's prospects of justifying its $80B valuation or raising at a higher valuation look weaker than before.

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

    Assessment marcus-the-openai-plot-thickens-172c6c-a1.

  48. Assessment publishedBy this siteGary Marcus

    OpenAI habitually spins rather than tells the truth.

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

    Assessment marcus-the-openai-board-was-right-ef8d55-a1.

  49. HeldBy this siteGary Marcus

    Held record marcus-the-openai-board-was-right-8c26b1

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

  50. Assessment publishedBy this siteGary Marcus

    People are starting to see through Sam Altman's ruse after he got away with a lot for a long time.

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

    Assessment marcus-the-openai-board-was-right-847f88-a1.

  51. Assessment publishedBy this siteGary Marcus

    The new OpenAI board should recognize that Sam Altman's behavior is inconsistent with a nonprofit pledged to help humanity.

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

    Assessment marcus-the-openai-board-was-right-6be20f-a1.

  52. Assessment publishedBy this siteGary Marcus

    Sam Altman knew the voice resembled Johansson's Her character, so the coincidence claim was a lie.

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

    Assessment marcus-the-openai-board-was-right-4dc2cd-a1.

  53. Assessment publishedBy this siteGary Marcus

    OpenAI intentionally made its voice resemble Scarlett Johansson's.

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

    Assessment marcus-the-openai-board-was-right-353bb1-a1.

  54. Assessment publishedBy this siteGary Marcus

    OpenAI's claim that its voice's resemblance to Scarlett Johansson was coincidental is false.

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

    Assessment marcus-the-openai-board-was-right-30f64e-a1.

  55. Assessment publishedBy this siteGary Marcus

    Those who want to use artists', writers' and actors' work should compensate them and get their permission.

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

    Assessment marcus-the-openai-board-was-right-2d32bc-a1.

  57. Assessment publishedBy this siteGary Marcus

    LLMs are unreliable.

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

    Assessment marcus-the-open-letter-controversy-dcb6c2-a1.

  58. HeldBy this siteGary Marcus

    Held record marcus-the-open-letter-controversy-dcb6c2-2

    e1 is dated 2023-03, which is not inside the period described (2023-03-01 to 2023-03-30); it can only be context

  59. Assessment publishedBy this siteGary Marcus

    Doing nothing about the risks of large AI systems is not acceptable; action should be taken.

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

    Assessment marcus-the-open-letter-controversy-cb3187-a1.

  61. Assessment publishedBy this siteGary Marcus

    None of the top AI labs is remotely transparent.

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

    Assessment marcus-the-open-letter-controversy-b386ab-a1.

  62. HeldBy this siteGary Marcus

    Held record marcus-the-open-letter-controversy-b386ab-3

    The final-text check failed: citations_correct; overclaims: e1 explanation: "The launch documentation describes refusal training, expert adversarial testing feeding into mitigations, and monitoring and enforcement for the deployed model." The e1 text never mentions enforcement. It also presents monitoring for abuse only as something it is 'important' to add, not as a control that is in place..

  63. HeldBy this siteGary Marcus

    Held record marcus-the-open-letter-controversy-b386ab-2

    The anthropic review does not accept e2 as contradicts.

  64. Assessment publishedBy this siteGary Marcus

    AI should not be exempt from regulation.

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

    Assessment marcus-the-open-letter-controversy-a2f203-a1.

  65. Assessment publishedBy this siteGary Marcus

    Doing nothing about AI risk is the most foolish option.

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

    Assessment marcus-the-open-letter-controversy-7f3f9a-a1.

  66. Assessment publishedBy this siteGary Marcus

    AI currently presents a perfect storm of corporate irresponsibility, widespread adoption, lack of regulation and many unknowns.

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

    Assessment marcus-the-open-letter-controversy-765261-a1.

  67. Assessment publishedBy this siteGary Marcus

    LLMs are not close to AGI.

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

    Assessment marcus-the-open-letter-controversy-70026e-a1.

  71. Assessment publishedBy this siteGary Marcus

    There is a real need to rein in the AI systems currently being rolled out at unprecedented scale.

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

    Assessment marcus-the-open-letter-controversy-463ad8-a1.

  73. Assessment publishedBy this siteGary Marcus

    Public backlash against AI will be so strong by 2028 that anti-AI sentiment will probably be a major factor in the 2028 US Presidential election.

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

    Assessment marcus-the-one-issue-unifying-the-american-732b6e-a1.

  74. Assessment publishedBy this siteGary Marcus

    Greed, overhype, and recklessness in the AI industry have spoiled what could have been a great technology.

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

    Assessment marcus-the-one-issue-unifying-the-american-0350ce-a1.

  75. HeldBy this siteGary Marcus

    Held record marcus-the-new-sport-of-misrepresenting-d37f85

    The anthropic screen finds the claim checkable, so it must be graded.

  78. HeldBy this siteGary Marcus

    Held record marcus-the-new-sport-of-misrepresenting-9b3407

    The openai screen reads the claim differently.

  79. Assessment publishedBy this siteGary Marcus

    The AI field has repeatedly found that doing well on a benchmark designed for X does not mean a system has actually solved X.

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

    Assessment marcus-the-new-sport-of-misrepresenting-633bac-a2.

  80. HeldBy this siteGary Marcus

    Held record marcus-the-new-sport-of-misrepresenting-582f3a

    The final-text check failed: rationale_supported_by_evidence; overclaims: Failures to generalize do not establish that no genuine improvement occurred; some gains transferred to independent tests [e2, e3]..

  82. HeldBy this siteGary Marcus

    Held record marcus-the-new-science-of-alt-intelligence-f70bf0

    The anthropic screen reads the claim differently.

  83. Assessment publishedBy this siteGary Marcus

    The persistent problem of discomprehension may signal that AI needs a fundamental rethink.

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

    Assessment marcus-the-new-science-of-alt-intelligence-f493ea-a1.

  86. HeldBy this siteGary Marcus

    Held record marcus-the-new-science-of-alt-intelligence-8cd6fd

    The openai screen reads the claim differently.

  88. Assessment publishedBy this siteGary Marcus

    AI research should draw clues from human and animal cognition without slavishly replicating human intelligence.

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

    Assessment marcus-the-new-science-of-alt-intelligence-4a9f0c-a1.

  90. Assessment publishedBy this siteGary Marcus

    The best path to AGI may not be through data-driven Alt Intelligence approaches.

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

    Assessment marcus-the-new-science-of-alt-intelligence-2b01c1-a1.

  91. Assessment publishedBy this siteGary Marcus

    The AI field should pursue multiple research directions rather than dismissing undeveloped ideas prematurely.

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

    Assessment marcus-the-new-science-of-alt-intelligence-1cffad-a1.

  92. Assessment publishedBy this siteGary Marcus

    Anyone who is candid will recognise that these kinds of errors by current AI systems reveal that something is, for now, deeply amiss.

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

    Assessment marcus-the-new-science-of-alt-intelligence-1ae1e2-a1.

  93. Assessment publishedBy this siteGary Marcus

    AI is still at an early stage of development.

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

    Assessment marcus-the-new-science-of-alt-intelligence-174e93-a1.

  94. HeldBy this siteGary Marcus

    Held record marcus-the-new-science-of-alt-intelligence-1705d9

    The anthropic screen reads the claim differently.

  96. Assessment publishedBy this siteGary Marcus

    OpenAI may yet turn out to be the WeWork of AI.

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

    Assessment marcus-the-need-for-a-president-that-speaks-c95cf3-a1.

  97. Assessment publishedBy this siteGary Marcus

    New approaches to AI are needed.

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

    Assessment marcus-the-need-for-a-president-that-speaks-b3eeb9-a1.

  98. Assessment publishedBy this siteGary Marcus

    The US must elect a President in 2024 who fully grasps AI's transformative impact.

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

    Assessment marcus-the-need-for-a-president-that-speaks-a2975a-a1.

  99. HeldBy this siteGary Marcus

    Held record marcus-the-need-for-a-president-that-speaks-7c5364

    The openai screen reads the claim differently.

  100. Assessment publishedBy this siteGary Marcus

    The US needs a President who stands up to big tech and resists regulatory capture.

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

    Assessment marcus-the-need-for-a-president-that-speaks-70ad38-a1.

  101. HeldBy this siteGary Marcus

    Held record marcus-the-need-for-a-president-that-speaks-5a48ca

    The anthropic screen finds the claim checkable, so it must be graded.

  102. Assessment publishedBy this siteGary Marcus

    The US needs a President who can tell when corporate AI promises are unrealistic, so that AI policy is grounded in reality.

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

    Assessment marcus-the-need-for-a-president-that-speaks-1a060b-a1.

  103. Assessment publishedBy this siteGary Marcus

    The whole generative AI thing is looking more and more like a dud.

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

    Assessment marcus-the-need-for-a-president-that-speaks-10669a-a1.

  105. Assessment publishedBy this siteGary Marcus

    AI will change everything, for good and bad, sometime within the next 5 to 20 years.

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

    Assessment marcus-the-need-for-a-president-that-speaks-033998-a1.

  106. HeldBy this siteGary Marcus

    Held record marcus-the-month-generative-ai-lost-its-e7a329

    The openai screen reads the claim differently.

  107. Assessment publishedBy this siteGary Marcus

    Chinese AI models are rapidly gaining market traction.

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

    Assessment marcus-the-month-generative-ai-lost-its-baaeea-a2.

  108. HeldBy this siteGary Marcus

    Held record marcus-the-month-generative-ai-lost-its-b392c0

    The openai review doubts the attribution.

  109. Assessment publishedBy this siteGary Marcus

    It is not at all clear that Anthropic will actually report the profitable quarter WSJ predicted.

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

    Assessment marcus-the-month-generative-ai-lost-its-b08e0c-a1.

  110. HeldBy this siteGary Marcus

    Held record marcus-the-month-generative-ai-lost-its-af0cf3

    The final-text check failed: proposition_faithful.

  111. HeldBy this siteGary Marcus

    Held record marcus-the-month-generative-ai-lost-its-a948c0

    The anthropic screen reads the claim differently.

  112. HeldBy this siteGary Marcus

    Held record marcus-the-month-generative-ai-lost-its-a2520b

    The final-text check failed: proposition_faithful.

  113. Assessment publishedBy this siteGary Marcus

    Recent developments (IPO delay, stock declines, policy chaos, Chinese competition, commoditization) bode poorly for the US AI industry.

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

    Assessment marcus-the-month-generative-ai-lost-its-8964ae-a1.

  114. Assessment publishedBy this siteGary Marcus

    If his FT argument is even remotely correct, hyperscaling could prove to be among the biggest financial blunders in history.

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

    Assessment marcus-the-month-generative-ai-lost-its-7bef78-a1.

  115. Assessment publishedBy this siteGary Marcus

    OpenAI's IPO plans may fall apart, as WeWork's did.

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

    Assessment marcus-the-month-generative-ai-lost-its-5c44e8-a1.

  117. HeldBy this siteGary Marcus

    Held record marcus-the-month-generative-ai-lost-its-3e7c82

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

  118. HeldBy this siteGary Marcus

    Held record marcus-the-month-generative-ai-lost-its-362eef

    The anthropic review does not accept e5 as contradicts.

  119. HeldBy this siteGary Marcus

    Held record marcus-the-month-generative-ai-lost-its-303234

    The anthropic screen reads the claim differently.

  120. Assessment publishedBy this siteGary Marcus

    The game has changed, and even Polymarket now sees it.

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

    Assessment marcus-the-month-generative-ai-lost-its-1de760-a1.

  121. Assessment publishedBy this siteGary Marcus

    There is perhaps already evidence that OpenAI's troubles are hurting companies it does business with.

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

    Assessment marcus-the-month-generative-ai-lost-its-17a5f8-a1.

  122. Assessment publishedBy this siteGary Marcus

    LLMs are becoming a commodity.

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

    Assessment marcus-the-month-generative-ai-lost-its-092384-a1.

  123. Assessment publishedBy this siteGary Marcus

    OpenAI's IPO delay reflects a lack of confidence.

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

    Assessment marcus-the-month-generative-ai-lost-its-024c3c-a1.

  124. Assessment publishedBy this siteGary Marcus

    Every company that would be subject to SB-1047's stronger regulations is valued in the billions of dollars.

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

    Assessment marcus-the-misguided-backlash-against-californias-e945d0-a1.

  125. HeldBy this siteGary Marcus

    Held record marcus-the-misguided-backlash-against-californias-e7b379

    The anthropic screen doubts the attribution.

  126. HeldBy this siteGary Marcus

    Held record marcus-the-misguided-backlash-against-californias-c41ab9

    The anthropic screen doubts the attribution.

  127. HeldBy this siteGary Marcus

    Held record marcus-the-misguided-backlash-against-californias-c3bf59

    The final-text check failed: rationale_supported_by_evidence; overclaims: These findings support the observation about the early Recall index [e1, e2]..

  128. Assessment publishedBy this siteGary Marcus

    Nobody in the AI industry appears to actually want the regulation they previously hinted at supporting before the Senate.

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

    Assessment marcus-the-misguided-backlash-against-californias-7c61aa-a1.

  129. Assessment publishedBy this siteGary Marcus

    SB-1047 should be strengthened rather than weakened.

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

    Assessment marcus-the-misguided-backlash-against-californias-7b2918-a1.

  130. Assessment publishedBy this siteGary Marcus

    Andrew Ng's critique does not identify which specific SB-1047 provisions would stifle AI model builders.

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

    Assessment marcus-the-misguided-backlash-against-californias-778c5d-a1.

  131. HeldBy this siteGary Marcus

    Held record marcus-the-misguided-backlash-against-californias-6ebe32

    The anthropic screen finds the claim checkable, so it must be graded.

  132. Assessment publishedBy this siteGary Marcus

    The requirements of SB-1047 (safety testing, liability in extreme cases, a shutdown path) are not unreasonable.

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

    Assessment marcus-the-misguided-backlash-against-californias-55c54b-a1.

  133. HeldBy this siteGary Marcus

    Held record marcus-the-misguided-backlash-against-californias-5258e6

    The anthropic screen doubts the attribution.

  135. Assessment publishedBy this siteGary Marcus

    Companies that can afford $100M+ training runs can easily afford SB-1047's paperwork and testing requirements.

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

    Assessment marcus-the-misguided-backlash-against-californias-3b63cd-a2.

  136. HeldBy this siteGary Marcus

    Held record marcus-the-misguided-backlash-against-californias-37b0c4

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

  137. Assessment publishedBy this siteGary Marcus

    SB-1047 is in many ways much weaker than the AI regulation Marcus and Sam Altman discussed at the Senate in 2023.

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

    Assessment marcus-the-misguided-backlash-against-californias-1b2aef-a1.

  138. HeldBy this siteGary Marcus

    Held record marcus-the-mirage-of-visual-understanding-ef5048

    The anthropic screen reads the claim differently.

  139. Assessment publishedBy this siteGary Marcus

    The frontier models described in the study are not AGI.

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

    Assessment marcus-the-mirage-of-visual-understanding-dcd969-a1.

  140. HeldBy this siteGary Marcus

    Held record marcus-the-mirage-of-visual-understanding-c23eda

    The anthropic screen reads the claim differently.

  142. Assessment publishedBy this siteGary Marcus

    Some white collar jobs, such as entry-level coder or market research assistant, may be in jeopardy from AI in the near term.

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

    Assessment marcus-the-mirage-of-visual-understanding-8740a0-a1.

  144. Assessment publishedBy this siteGary Marcus

    A humanoid robot that can't understand the visual world is just a demo and not something you can trust.

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

    Assessment marcus-the-mirage-of-visual-understanding-3c4093-a1.

  145. Assessment publishedBy this siteGary Marcus

    Chat-style search's tendency to hallucinate is likely to accidentally produce medical misinformation.

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

    Assessment marcus-the-long-shadow-of-gpt-f6db7a-a1.

  146. HeldBy this siteGary Marcus

    Held record marcus-the-long-shadow-of-gpt-ebd2a9

    The anthropic screen reads the claim differently.

  147. Assessment publishedBy this siteGary Marcus

    The problem of LLM-generated fake submissions flooding outlets is growing rapidly.

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

    Assessment marcus-the-long-shadow-of-gpt-c8dabf-a2.

  148. Assessment publishedBy this siteGary Marcus

    We have not come to grips with the unknown unknown risks of LLMs.

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

    Assessment marcus-the-long-shadow-of-gpt-b05627-a1.

  150. Assessment publishedBy this siteGary Marcus

    New concerns about LLM harms are emerging almost daily.

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

    Assessment marcus-the-long-shadow-of-gpt-a35687-a2.

  151. Assessment publishedBy this siteGary Marcus

    LLM-generated prose has already disrupted web forums and peer review by flooding outlets with fake submissions.

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

    Assessment marcus-the-long-shadow-of-gpt-944186-a1.

  153. HeldBy this siteGary Marcus

    Held record marcus-the-long-shadow-of-gpt-895f95

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

  154. Assessment publishedBy this siteGary Marcus

    Society needs to prepare for harms caused by large language models.

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

    Assessment marcus-the-long-shadow-of-gpt-895f95-2-a1.

  155. Assessment publishedBy this siteGary Marcus

    Many of the chatbot-enabled attacks he described might be combined with advances in voice-cloning technology, which is already being used for scams.

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

    Assessment marcus-the-long-shadow-of-gpt-22d806-a1.

  156. Assessment publishedBy this siteGary Marcus

    Companion chatbots have already been changed arbitrarily in ways that left some users in serious emotional pain.

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

    Assessment marcus-the-long-shadow-of-gpt-1ed5b4-a1.

  157. Assessment publishedBy this siteGary Marcus

    Marcus does not think we yet know half of what the harms that large language models are almost certain to cause will entail.

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

    Assessment marcus-the-long-shadow-of-gpt-118586-a1.

  158. Assessment publishedBy this siteGary Marcus

    It is not yet known whether LLMs will be net positive or net negative.

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

    Assessment marcus-the-long-shadow-of-gpt-034e6b-a1.

  159. Assessment publishedBy this siteGary Marcus

    The METR graph tells us nothing about which questions computers can reliably answer or when specific questions will be solved.

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

    Assessment marcus-the-latest-ai-scaling-graph-and-why-f8db9d-a1.

  160. Assessment publishedBy this siteGary Marcus

    METR's blog and tweets made claims far beyond what its study supports, dropping the technical paper's qualifications.

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

    Assessment marcus-the-latest-ai-scaling-graph-and-why-f680c2-a1.

  161. HeldBy this siteGary Marcus

    Held record marcus-the-latest-ai-scaling-graph-and-why-f55aad

    The openai screen reads the claim differently.

  162. HeldBy this siteGary Marcus

    Held record marcus-the-latest-ai-scaling-graph-and-why-eb18e0

    The anthropic screen reads the claim differently.

  163. HeldBy this siteGary Marcus

    Held record marcus-the-latest-ai-scaling-graph-and-why-d6e571

    The anthropic screen reads the claim differently.

  164. HeldBy this siteGary Marcus

    Held record marcus-the-latest-ai-scaling-graph-and-why-cf0b29

    The anthropic review reads the claim differently.

  166. HeldBy this siteGary Marcus

    Held record marcus-the-latest-ai-scaling-graph-and-why-c06f05

    The anthropic screen reads the claim differently.

  167. HeldBy this siteGary Marcus

    Held record marcus-the-latest-ai-scaling-graph-and-why-9cf4ba

    The anthropic screen reads the claim differently.

  168. HeldBy this siteGary Marcus

    Held record marcus-the-latest-ai-scaling-graph-and-why-5bb71c

    The final-text check failed: citations_correct.

  169. Assessment publishedBy this siteGary Marcus

    o3 cannot do all tasks humans can do in 1.7 hours, and GPT-3 could not do everything a person can do in 15 seconds.

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

    Assessment marcus-the-latest-ai-scaling-graph-and-why-49a50a-a1.

  170. Assessment publishedBy this siteGary Marcus

    Using the METR graph to predict future AI capabilities is misguided, and its virality reflects people repeating what they want to believe.

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

    Assessment marcus-the-latest-ai-scaling-graph-and-why-3a3eae-a1.

  171. HeldBy this siteGary Marcus

    Held record marcus-the-latest-ai-scaling-graph-and-why-2a2a2b

    The anthropic screen reads the claim differently.

  172. HeldBy this siteGary Marcus

    Held record marcus-the-latest-ai-scaling-graph-and-why-148b02

    The anthropic screen reads the claim differently.

  173. HeldBy this siteGary Marcus

    Held record marcus-the-last-two-weeks-at-openai-da8ecf

    The final-text check failed: proposition_faithful.

  174. HeldBy this siteGary Marcus

    Held record marcus-the-last-two-weeks-at-openai-d315bf

    The openai review reads the claim differently.

  175. Assessment publishedBy this siteGary Marcus

    OpenAI's challenges got worse over the two weeks before late February 2024.

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

    Assessment marcus-the-last-two-weeks-at-openai-924c51-a1.

  176. HeldBy this siteGary Marcus

    Held record marcus-the-last-two-weeks-at-openai-59b61b

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

  177. HeldBy this siteGary Marcus

    Held record marcus-the-last-two-weeks-at-openai-3bdfc7

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

  178. HeldBy this siteGary Marcus

    Held record marcus-the-last-two-weeks-at-openai-142931

    The anthropic review reads the claim differently.

  179. Assessment publishedBy this siteGary Marcus

    Even with reasoning, LLMs still cannot handle distribution shift, as the June 2025 Apple paper confirmed.

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

    Assessment marcus-the-last-few-months-have-been-devastating-8b9119-a1.

  180. HeldBy this siteGary Marcus

    Held record marcus-the-last-few-months-have-been-devastating-80f768

    The anthropic review does not accept e2 as supports.

  181. Assessment publishedBy this siteGary Marcus

    Hopes that LLMs would bring us close to AGI have collapsed over the past few months.

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

    Assessment marcus-the-last-few-months-have-been-devastating-3a1b4d-a1.

  182. HeldBy this siteGary Marcus

    Held record marcus-the-last-few-months-have-been-devastating-208a66

    The anthropic screen reads the claim differently.

  183. HeldBy this siteGary Marcus

    Held record marcus-the-last-few-months-have-been-devastating-029f5e

    e1 is dated 2025-10, which is not inside the period described (2025-10-01 to 2025-10-18); it can only be context

  185. Assessment publishedBy this siteGary Marcus

    Musk's lawsuit plus the AGI-contingent Microsoft license have put OpenAI in the position of needing skeptical experts to deflate its own hype.

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

    Assessment marcus-the-irony-and-the-agony-of-elon-musks-e1c2be-a1.

  186. HeldBy this siteGary Marcus

    Held record marcus-the-irony-and-the-agony-of-elon-musks-9c5480

    The anthropic review does not accept e1 as contradicts.

  188. Assessment publishedBy this siteGary Marcus

    If Musk's lawsuit against OpenAI fails, humanity stands to lose.

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

    Assessment marcus-the-irony-and-the-agony-of-elon-musks-6365d9-a1.

  189. Assessment publishedBy this siteGary Marcus

    In the end, laws may or may not be enough to get OpenAI to honor its initial promises.

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

    Assessment marcus-the-irony-and-the-agony-of-elon-musks-44a9b6-a1.

  190. Assessment publishedBy this siteGary Marcus

    The world would be better if OpenAI returned to its original nonprofit, open, humanity-benefiting mission.

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

    Assessment marcus-the-irony-and-the-agony-of-elon-musks-1ac47c-a1.

  192. Assessment publishedBy this siteGary Marcus

    Amazon is starting to be a receptacle for inaccurate and exploitative quickie histories of recent events.

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

    Assessment marcus-the-imminent-enshittification-of-f28f7a-a2.

  193. HeldBy this siteGary Marcus

    Held record marcus-the-imminent-enshittification-of-f1d5eb

    The anthropic review does not accept e2 as supports.

  194. Assessment publishedBy this siteGary Marcus

    If users are left sifting through floods of useless misinformation, the value of search would fall to zero, which could kill Google.

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

    Assessment marcus-the-imminent-enshittification-of-e7efcc-a1.

  195. Assessment publishedBy this siteGary Marcus

    LLMs are fully capable of making up truly terrible advice, such as a disconcerting recipe from the previous week.

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

    Assessment marcus-the-imminent-enshittification-of-b01cd3-a2.

  196. Assessment publishedBy this siteGary Marcus

    LLM-driven degradation (enshittification) of the internet is already under way.

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

    Assessment marcus-the-imminent-enshittification-of-97563e-a1.

  197. Assessment publishedBy this siteGary Marcus

    Pollution of the internet by LLM-generated content became worse between February and August 2023 than Marcus had expected.

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

    Assessment marcus-the-imminent-enshittification-of-54efd8-a1.

  198. HeldBy this siteGary Marcus

    Held record marcus-the-imminent-enshittification-of-454594

    The anthropic review does not accept e1 as supports.