“I found that the claim of “solution” was pretty misleading, as I summed up in a tweet, and they refused to correct their misleading presentation:”
Our reading, the sentence that was checked:
In 2019 OpenAI's claim that its robot 'solved' the Rubik's cube was misleading, and OpenAI refused to correct it.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Erdosgate, paragraph 12.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-erdosgate-a7e202-a1
We recorded 5 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2025-10-21. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
From the speaker
- back to 2019
- a robot that had “solved” the Rubik’s cube
- the claim of “solution” was pretty misleading
- they refused to correct their misleading presentation
Our interpretation
- Assess whether the presentation materially implied that the learned system determined the solution moves when its learned contribution was physical manipulation.
- Assess the alleged refusal separately: it requires evidence of a correction request and a decision to reject it.
- An unchanged page or public disagreement alone does not prove refusal.
- The target period is 2019, as explicitly stated.
Counts as supported if: The presentation materially misrepresented the learned capability, and records establish that OpenAI rejected a request to correct that presentation.
Counts as contradicted if: The relevant presentation clearly communicated the capability and limitations, or records establish that OpenAI accepted and made the requested correction. A disclosure elsewhere is counterevidence but may not settle the presentation's overall impression.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Unresolved
The record was checked, and the evidence available by the cutoff does not settle it.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
The technical distinction behind Marcus's criticism is real: the learned system handled sensing and manipulation, while a separate algorithm supplied the moves. OpenAI also disclosed that distinction and imperfect success in its public materials. Contemporary criticism documents concern about the promotion, but the admitted evidence does not establish a refused correction request. The compound claim remains unresolved.
Limitations
- Disagreement over wording does not by itself prove a factual misrepresentation.
- Failure to find a documented refusal is not evidence that no refusal occurred.
- Potentially relevant forum responses could not be verified through read_page.
Evidence (3 items)
Goes against the claim: OpenAI,
generated by Kociemba’s algorithm.
Why it matters here: The announcement expressly assigns move generation to a separate algorithm and also reports imperfect success. These disclosures weigh against treating the complete presentation as concealing the system's actual capability.
- Where
- Footnote A
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Context only: OpenAI et al., author-hosted paper,
sensing and control
Why it matters here: The paper confirms the technical distinction underlying Marcus's criticism. It also shows that the researchers publicly disclosed that distinction.
- Publisher
- OpenAI et al., author-hosted paper
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: The Next Web,
the PR behind it “weapons-grade.”
Why it matters here: Independent contemporary criticism gives context for concern about the promotion. It does not establish either factual misrepresentation or a refusal to correct it.
- Publisher
- The Next Web
- Kind of source
- News report
- Describes events of
- Public from
What would settle it
A dated correction request and OpenAI response, together with contemporaneous versions of the specific presentation at issue.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Erdosgate
I found that the claim of “solution” was pretty misleading, as I summed up in a tweet, and they refused to correct their misleading presentation:
- Where
- paragraph 12
- Published
- 2025-10-21
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-erdosgate-a7e202-a1: Current. Outcome: Unresolved. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.