โTwo graphs, one presented by OpenAI and one by Chollet were misleading.โ
Our reading, the sentence that was checked:
The o3 ARC graphs presented by OpenAI and by Chollet were misleading.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-c39-aff2ef-a1
We recorded 17 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 2024-12-22. 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 releaseโs reference date)
From the speaker
- Two graphs
- one presented by OpenAI and one by Chollet
- were misleading
- omitting results from others like the @jacobandreas lab at MIT
Our interpretation
- The claim concerns the two presentations accompanying the December 20, 2024 o3 announcement.
- The surrounding text supplies a checkable meaning for misleading: omission of relevant prior results exaggerated o3's advantage over the field.
- The assessment must consider each graph's stated scope, labels, and accompanying explanation.
- Omission alone does not establish misleading presentation, and no deceptive intent is inferred.
Counts as supported if: Both original graphs omit relevant contemporaneous results on comparable evaluation settings, and their framing presents a materially larger advance over the field than those results justify.
Counts as contradicted if: The graphs include the relevant results, the proposed comparators are materially incompatible, or their explicit scope and surrounding explanation do not imply the asserted comparison with the field.
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 contemporaneous benchmark record confirms substantial prior results, including an MIT-associated ensemble. That makes Marcus's criticism testable, but it does not itself prove either graph omitted those results or implied a comparison with the entire field. Public and semi-private scores also differ substantially. Without verified original graph contents and framing, the claim about both graphs remains unresolved.
Limitations
- An image's absence from extracted text cannot establish an omission within that image.
- Numerically higher scores from a different evaluation split are not interchangeable comparators.
Evidence (1 item)
Context only: ARC Prize,
The SOTA on ARC-AGI-1 increased from 33% to 55.5% during the 2024 Kaggle competition.
Why it matters here: The table separately reports MARA(BARC) + MIT at 47.5% on semi-private evaluation and 62.8% on public evaluation, alongside other entrants. These prior results identify relevant comparisons but do not establish what either later graph omitted.
- Publisher
- ARC Prize
- Kind of source
- Benchmark result
- Describes events of
- Public from
What would settle it
Readable, dated copies of both original graphs and accompanying explanations, compared with contemporaneous results under matching evaluation conditions.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, ๐ผ๐ฏ โ๐๐ฅ๐ ๐๐๐โ ๐ฝ๐ผ๐๐๐บ๐ผ๐ฟ๐๐ฒ๐บ ๐บ๐ฒ๐ด๐ฎ๐๐ต๐ฟ๐ฒ๐ฎ๐ฑ: ๐๐ต๐ ๐๐ต๐ถ๐ป๐ด๐ ๐ด๐ผ๐ ๐ต๐ฒ๐ฎ๐๐ฒ๐ฑ, ๐๐ต๐ฎ๐ ๐๐ฒ๐ป๐ ๐๐ฟ๐ผ๐ป๐ด, ๐ฎ๐ป๐ฑ ๐๐ต๐ฎ๐ ๐ถ๐ ๐ฎ๐น๐น ๐บ๐ฒ๐ฎ๐ป๐
Two graphs, one presented by OpenAI and one by Chollet were misleading.
- Where
- paragraph 8
- Published
- 2024-12-22
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-c39-aff2ef-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.