“A new study on math, supporting what Davis and I wrote yesterday re LLMs struggling with mathematical reasoning from Mahdavi et al, converges on similar conclusions”
Our reading, the sentence that was checked:
A new study by Mahdavi et al. supports and reaches similar conclusions to what Marcus and Davis wrote about LLMs struggling with mathematical reasoning.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Scaling is over, the bubble may be deflating, LLMs still can’t reason, and you can’t trust Sam, paragraph 3.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-scaling-is-over-the-bubble-may-be-e26f32-a2
We recorded 11 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 2025-04-06. 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
- A new study on math
- supporting what Davis and I wrote yesterday
- LLMs struggling with mathematical reasoning
- from Mahdavi et al
- converges on similar conclusions
Our interpretation
- Assess whether the study's reported findings align with the preceding essay's central concern: difficulty producing sound mathematical proofs and recognizing faulty reasoning.
- Agreement does not require identical experiments or support for every assertion in the essay.
- The target is the study and essay available in early April 2025, not subsequent models.
Counts as supported if: The identified study reports weaknesses in rigorous Olympiad mathematics and in identifying invalid reasoning, consistent with the preceding essay.
Counts as contradicted if: The identified study instead reports reliable proof generation and verification, or its conclusions do not materially resemble the essay's central concern.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Supported
The dated evidence matches what the claim committed to, under the criteria frozen before grading.
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
Mahdavi et al.'s April 1, 2025 preprint reports difficulty solving challenging Olympiad problems and frequent failures to distinguish correct reasoning from flawed solutions [e1]. The April 5 Marcus–Davis essay emphasizes proofs presented confidently despite gaps or errors [e2]. The study's reported findings therefore support the essay's central concern about flawed mathematical reasoning and failure to recognize errors [e1, e2].
Limitations
- This comparison does not establish every broader assertion in the essay.
- The study was a preprint, and its findings concern its tested models, tasks and evaluation procedures.
- The assessment concerns sources available in early April 2025, not later models.
Evidence (2 items)
Supports the claim: Mahdavi et al., arXiv,
Our study reveals that current LLMs fall significantly short of solving challenging Olympiad-level problems and frequently fail to distinguish correct mathematical reasoning from clearly flawed solutions.
Why it matters here: The original abstract contains the conclusions Marcus attributes to the study. They align directly with concerns about invalid proofs and unreliable recognition of mathematical errors.
- Publisher
- Mahdavi et al., arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: Ernest Davis and Gary Marcus, Marcus on AI,
outright error
Why it matters here: This establishes the preceding essay's emphasis on flawed proofs presented confidently. It is not independent evidence of model performance.
- Where
- Second paragraph
- Publisher
- Ernest Davis and Gary Marcus, Marcus on AI
- Kind of source
- Secondary summary
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Scaling is over, the bubble may be deflating, LLMs still can’t reason, and you can’t trust Sam
A new study on math, supporting what Davis and I wrote yesterday re LLMs struggling with mathematical reasoning from Mahdavi et al, converges on similar conclusions
- Where
- paragraph 3
- Published
- 2025-04-06
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-scaling-is-over-the-bubble-may-be-e26f32-a2: Current. Outcome: Supported. 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.
About assessment marcus-scaling-is-over-the-bubble-may-be-e26f32-a2.