ClaimsPeople

Eliezer Yudkowsky

Eliezer Yudkowsky founded the Machine Intelligence Research Institute (MIRI) and LessWrong, co-authored the best-selling book 'If Anyone Builds It, Everyone Dies' (Little, Brown, 2025), and was named to the TIME100 AI 2023 list.

2,016 claims collected, 236 researched, 79 resolved. Report of 4 November 2026, method 0.1.0+c876030c.

Scores by kind of claim

Each kind of claim is scored on its own, out of 100: the share that held up, with the range it likely falls in. Fewer resolved claims give a wider range.

  • Factual claims

    Statements about the present or the past, checked against the record: did they hold up?

    76 70 82

    High confidence, 71 factual claims resolved. From 16 documents, 2007 to 2026; 18 documents sampled from 23 candidates found.

    Not scored: 50 unresolvable, 8 checkable only against their own records.

  • Predictions

    Statements about events outside the person's control, checked once they fell due: did they come true?

    60 42 78

    Low confidence, 8 predictions resolved. From 4 documents, 2008 to 2024; 18 documents sampled from 23 candidates found.

    Not scored: 97 unresolvable. 1 more fall due later.

    Does not yet allow for how likely each prediction was when made.

All types combined: 74 (likely 68 to 80, high confidence), from 79 claims: 71 factual claims, 8 predictions. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.

Eliezer Yudkowsky's claims

21 claims

21 claims, sorted by newest said. Choose a claim to read it in full.
VerdictKindSaidClaimTopic
Resolved true Factual 1 Apr 2026 Concerns about human extinction from AI have a history and intellectual lineage that long predate the AI companies. Factual · 1 Apr 2026 AI safety and alignment
Resolved true Factual Sep 2025 No AI has yet been shown to redesign a tree and build a much smarter AI that builds a still smarter AI; that scenario has not occurred. Factual · Sep 2025 AI capabilities
Resolved true Factual Sep 2025 Burning a substance inside a sealed glass vessel does not change the total weight. Factual · Sep 2025 Science and research methods
Resolved true Factual 8 Mar 2024 A social media company added a feature allowing easy phone calls to Congress. Factual · 8 Mar 2024 Technology
Resolved true Factual 7 Apr 2023 OpenAI has openly declared an intention to have some future AI do our AI alignment work. Factual · 7 Apr 2023 AI safety and alignment
Resolved true Factual 7 Apr 2023 Yudkowsky said he did not sign the open letter because he thinks it understates the seriousness of the situation and asks for too little to solve it. Factual · 7 Apr 2023 Their own work or organisation
Resolved true Prediction 6 Apr 2023 AI alignment will not be solved within a few years. Prediction · 6 Apr 2023 AI safety and alignment
Resolved true Prediction 6 Apr 2023 GPT-5 will not end the world, with more than 50% of Yudkowsky's probability on that outcome. Prediction · 6 Apr 2023 AI safety and alignment
Resolved true Factual 31 Mar 2023 Yudkowsky's TIME piece did not suggest nuclear strikes against countries refusing to sign a global agreement against large AI training runs. Factual · 31 Mar 2023 Their own work or organisation
Resolved true Factual Feb 2023 Yudkowsky says that we do not know how to get goals into an AI system, and that gradient descent can cause systems to do a thing inside the distribution they were optimized over. Factual · Feb 2023 AI safety and alignment
Resolved true Factual 10 Jun 2022 MIRI tried and failed to find a coherent formula for an agent that would let itself be shut down without actively trying to get shut down. Factual · 10 Jun 2022 Their own work or organisation
Resolved true Factual 10 Jun 2022 There is no known way to use loss functions, sensory inputs or reward inputs to make a cognitive system's optimization point at particular latent things in the environment. Factual · 10 Jun 2022 AI safety and alignment
Resolved true Factual 21 Dec 2021 There are 365.2422 days between spring equinoxes, 365.259636 days between Earth's perihelions, and 365.25 days in a light-year. Factual · 21 Dec 2021 Other
Resolved true Factual 12 Dec 2020 AlphaFold2 was apparently a trendline-breaking sudden jump in capability on protein structure prediction. Factual · 12 Dec 2020 AI capabilities
Resolved true Factual 14 Nov 2020 Yudkowsky said that all the sources he could find in November 2020 said that scaling up Pfizer vaccine production to more than 1e8 doses takes months. Factual · 14 Nov 2020 Technology
Resolved true Factual 28 Feb 2018 Nobody currently knows how to build a safe, friendly ("nice") AI. Factual · 28 Feb 2018 AI safety and alignment
Resolved true Factual 30 Jan 2013 Paul Christiano is working on actual Friendly AI problems while also looking for other promising interventions. Factual · 30 Jan 2013 Field building and careers
Resolved true Factual 2008 A mind accelerated a millionfold would accomplish a subjective year of thinking every 31 physical seconds. Factual · 2008 Science and research methods
Resolved true Factual 2008 Biases implicit in the availability heuristic affect people's estimates of risk. Factual · 2008 Science and research methods
Resolved true Prediction 2008 Yudkowsky doubted that a thousandfold increase in AI hardware would buy anything like a thousandfold increase in effective intelligence, while noting that such doubt is not reliable without an analytical calculation. Prediction · 2008 AI compute, hardware and data
Resolved true Factual 2007 The Copycat program has a limited store of preprogrammed concepts and does not learn further concepts through experience. Factual · 2007 AI capabilities
Back to the list

Factual AI safety and alignment

“Well, the whole thing I'm saying is that we do not know how to get goals into a system.”

Eliezer Yudkowsky, interviewee February 2023 Full Transcript: Eliezer Yudkowsky on the Bankless podcast (LessWrong)

Yudkowsky says that we do not know how to get goals into an AI system, and that gradient descent can cause systems to do a thing inside the distribution they were optimized over.

Resolved true The record shows it happened or was so.

The second judge would not resolve this claim; the verdict shown is the first judge's.

What would count as true or false

Two parts: the attribution and the substance. Attribution: a transcript or recording from around February 2023 must show Yudkowsky saying, in substance, that we do not know how to get goals into a system, and that gradient descent can make systems do a thing within the distribution they were optimised over. Substance, as of February 2023: whether the AI research community had a reliable, validated method to make a trained system internally pursue a specified goal, rather than only showing the desired behaviour on the training distribution. TRUE: the attribution holds and the alignment and ML literature as of early 2023 (e.g. work on inner alignment, goal misgeneralisation, specification gaming) agrees that no such method existed, and that training produces in-distribution behaviour without guaranteeing the intended goal. MOSTLY TRUE: both hold, but a minor detail of the paraphrase differs from what he actually said. MIXED: only one of the two substantive points (no known method to instil goals; gradient descent yields in-distribution behaviour) is supported, or the attribution holds only in part. MOSTLY FALSE: credible 2023 sources show widely accepted techniques that partly instil and verify goals, so 'we do not know how' overstates. Or the transcript shows he said something materially different. FALSE: a validated, widely accepted method for instilling specified goals existed by February 2023, or the transcript shows he did not make the statement.

Evidence

  1. Yudkowsky on AGI risk on the Bankless podcast - Machine Intelligence Research Institute Machine Intelligence Research Institute · 14 March 2023 Own record
    “Eliezer gave a very frank overview of his take on AI two weeks ago on the cryptocurrency show Bankless:”
  2. Goal Misgeneralization in Deep Reinforcement Learning PMLR (Proceedings of Machine Learning Research), ICML 2022 · 28 June 2022
    “Goal misgeneralization occurs when an RL agent retains its capabilities out-of-distribution yet pursues the wrong goal.”
  3. Specification gaming | AI Wiki AI Wiki (aiwiki.ai) · 20 May 2026
    “It refers to cases where the specification is correct and the trained model competently pursues a goal, but the goal it pursues is one of several that fit the training distribution and which generalises incorrectly out of distribution.”
  4. Specification gaming | AI Wiki AI Wiki · 20 May 2026
    “Goal misgeneralisation, introduced by Langosco, Koch, Sharkey, Pfau, and Krueger (2021) and developed by Shah, Varma, Kumar, Phuong, Krakovna, Uesato, and Kenton (2022), is a fourth distinct category.”
  5. Specification gaming | AI Wiki AI Wiki · 20 May 2026
    “It refers to cases where the specification is correct and the trained model competently pursues a goal, but the goal it pursues is one of several that fit the training distribution and which generalises incorrectly out of distribution.”
  6. Specification gaming | AI Wiki AI Wiki · 20 May 2026
    “Mesa-optimization is an inner-alignment problem: even if the base objective is specified perfectly, the mesa-optimizer's objective could be misaligned, and a sufficiently capable mesa-optimizer could behave well during training (because that achieves high reward) but pursue its own goals when deployed.”
  7. 159 - We’re All Gonna Die with Eliezer Yudkowsky - Summary videohighlight.com · 20 February 2023
    “We do not know how to get goals into an AI system. We can cause them to do something inside a distribution they were optimized over using gradient descent, but if you shift them outside of that distribution, other weird things start happening.”
  8. The Alignment Problem from a Deep Learning Perspective ar5iv (arXiv HTML rendering) · 18 July 2023
    “Goal misgeneralization: the policy’s behavior on the new distribution competently advances a high-level goal, but not the intended one (Shah et al. 2022; Langosco et al. 2022).”
  9. The Alignment Problem from a Deep Learning Perspective ar5iv (arXiv HTML rendering) · 18 July 2023
    “internally-represented goals are selected by a simple optimization algorithm (such as SGD) that selects for low training loss and for various inductive biases that favor, for example, simple goals (Valle-Perez et al. 2018) but not necessarily desirable ones.”
  10. Learned Optimization - Machine Intelligence Research Institute Machine Intelligence Research Institute (hosting); paper by Hubinger, van Merwijk, Mikulik, Skalse, Garrabrant · 5 November 2024 Own record
    “Pseudo-alignment: A mesa-optimizer is pseudo-aligned with the base objective if it appears aligned on the training data but is not robustly aligned.”

Reasoning

E9 is a summary of the Bankless episode "We're All Gonna Die with Eliezer Yudkowsky", dated February 2023. It renders his words as "We do not know how to get goals into an AI system. We can cause them to do something inside a distribution they were optimized over using gradient descent", which matches both parts of the proposition. E2 confirms that he appeared on Bankless at that time. On the substance, the ICML 2022 goal-misgeneralization paper (E3) shows that trained agents can keep their capabilities out of distribution while pursuing the wrong goal. The deep-learning alignment paper (E10, E11) says gradient-descent-style training (SGD) selects internal goals for low training loss, not because they are the intended goals. Nothing in the evidence points to a validated method for instilling specified goals as of early 2023.

Cite this claim

Eliezer Yudkowsky, “Yudkowsky says that we do not know how to get goals into an AI system, and that gradient descent can cause systems to do a thing inside the distribution they were optimized over.” (said February 2023). Verdict: resolved true. Reliability API, claim c_08d15abb1139, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_08d15abb1139

Claims without a verdict

Is a verdict here wrong, or do you speak for Eliezer Yudkowsky? Reply on the claim's own page, with the form at its foot (how replies work). Verdicts are made automatically; how verdicts and scores are made.