ClaimsEliezer Yudkowsky

“I mean, it's theoretically possible, but seems very unlikely.”

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

Factual AI policy and governance

Yudkowsky says it is theoretically possible, but seems very unlikely, that the AI event horizon has already been crossed, because somebody would need an AI in their lab that was much more advanced than the public AI technology.

Unresolvable The public record could not settle it.

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

What would count as true or false

The proposition reports what Yudkowsky said, so judge whether the statement was made, as of February 2023. True: a transcript or recording from about February 2023 (e.g., the Bankless interview) shows Yudkowsky saying it is theoretically possible but very unlikely that the AI 'event horizon' had already been crossed, and giving as his reason that someone would need a lab AI much more advanced than public AI technology. Mostly true: he made the statement, but the reasoning is paraphrased with a minor difference, or the date is slightly off. Mixed: he made the 'possible but very unlikely' statement, but the reasoning attributed to him is substantially different or absent. Mostly false: the transcript shows a materially different stance on this question, for example calling it likely, or answering a different question. False: the transcript shows he did not make such a statement, or said the opposite.

Set before the research began.

Evidence (4)

  1. E1 Bankless Podcast: 159 - We're All Gonna Die with Eliezer Yudkowsky — LessWrong LessWrong · 20 February 2023 Own record

    “20th Feb 2023”

  2. E2 Transcript: Yudkowsky on Bankless follow-up Q&A — LessWrong LessWrong · 28 February 2023 Own record

    “Eliezer: Um, no, but I'm not going to go name the competitor, because then people will be like, "Oh, I should go work for them", you know? I'd rather they didn't.”

  3. E3 Yudkowsky on AGI risk on the Bankless podcast EA Forum (post by RobBensinger) · 13 March 2023 Own record

    “Eliezer gave a very frank overview of his take on AI two weeks ago on the cryptocurrency show Bankless:”

  4. E4 DEBRIEF - We're All Gonna Die podscripts.co (transcript of Bankless debrief episode) · 21 February 2023

    “Welcome to the debrief. This is our episode after the episode with Eliezer,”

Reasoning

E1 shows that a Bankless interview with Yudkowsky was published around 20 February 2023, and E4 and E5 confirm that the episode took place. None of the quoted passages contains his remark that it is "theoretically possible, but seems very unlikely" that the AI event horizon had already been crossed. None gives his stated reasoning either. The quote in E2 is about declining to name a competitor, which is a different question. The evidence doesn't show the wording or the reasoning, so the claim can't be settled.

Cite this claim

Eliezer Yudkowsky, “Yudkowsky says it is theoretically possible, but seems very unlikely, that the AI event horizon has already been crossed, because somebody would need an AI in their lab that was much more advanced than the public AI technology.” (said February 2023). Verdict: unresolvable. Reliability API, claim c_07330bcc794d, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_07330bcc794d

This claim is in dataset release v1.1.2, with the same verdict.

Reply to this claim

For Eliezer Yudkowsky, or someone speaking for them: a correction or counter-evidence. Replies are read before anything is shown; an approved reply shows your name, role, date, statement and link. Your email address is never shown and is deleted 30 days after the decision (privacy).

More factual claims by Eliezer Yudkowsky

  1. Mixed Factual 20 April 2026
    “I was observing that obvious-seeming hypothesis myself, sometime around the time of GPT 3.5.”

    Yudkowsky himself observed the persona-selection-like hypothesis around the time of GPT 3.5.

  2. Mostly true Factual 20 April 2026
    “The test-passing preference stands on its own once formed.”

    Once formed through RL training, an LLM's test-passing preference stands on its own and is not overridden by its inference of what the user wants.

  3. Mostly true Factual 20 April 2026
    “LLMs surely have neural circuits they learned to model text streams from fictional and nonfictional personas that are not lying, deceiving, cheating.”

    LLMs surely have neural circuits they learned to model text streams from fictional and nonfictional personas that are not lying, deceiving…

  4. Mostly true Factual 20 April 2026
    “The predict-as-if-honest circuitry is contextual, invoked to predict some streams of text but not others.”

    LLMs' predict-as-if-honest circuitry is contextual, invoked for predicting some text streams but not others.

  5. Mostly true Factual 20 April 2026
    “Presumably, somewhere along the way LLMs learn a dispreference for tests that don't pass, independently of how that gets prevented or whether the new plan does what the user really wanted.”

    Presumably, at some point during RL training on code-passing tasks, LLMs learn a dispreference for failing tests, independent of how that is avoided or whether the user's real goal is met.

All of them

Verdicts are made automatically; how verdicts are made.