“On some problems that capable students might well find hard, the AI did perfectly; on other problems that lot of mathematically-inclined college students people would find easy, the AI failed miserably.”
Our reading, the sentence that was checked:
In the authors' tests, GPT-4 with plug-ins did perfectly on some problems that capable students might well find hard, and failed miserably on other problems that many mathematically inclined college students would find easy.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Getting GPT to work with external tools is harder than you think, paragraph 10.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-getting-gpt-to-work-with-external-c94be9-a2
We recorded 16 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-29. The source may have been edited after 2023-10-23. 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)
Conditions
- on the authors' 105-problem test sets (scope, stated by the speaker)
- difficulty judged relative to capable or mathematically inclined college students (qualifier, stated by the speaker)
From the speaker
- On some problems that capable students might well find hard
- the AI did perfectly
- on other problems
- mathematically-inclined college students
- the AI failed miserably
Our interpretation
- This describes contrasting examples within the authors' experiment, not measured average performance against students.
- Hard and easy are ordinary educational judgments based on the knowledge and reasoning required.
- Did perfectly means a correct solution to the particular problem. Failed miserably means a substantive wrong answer or explanation.
- The referenced experiment occurred during June–August 2023.
Counts as supported if: The records contain correct solutions to demanding mathematical problems and substantive failures on elementary problems.
Counts as contradicted if: Checking the problems and responses removes either side of the claimed contrast.
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
The June–August 2023 records show both systems correctly answering problems involving high-dimensional probability and finite-field linear algebra [e1]. Both also incorrectly answered that two people could not each live 10 miles from Lake Michigan while living 100 miles apart; GPT4+WA's explanation treated the lake as a point [e2]. These examples support the qualitative contrast between success on demanding mathematics and substantive failure on elementary spatial reasoning [e1, e2].
Limitations
- No student comparison group was tested, so the evidence does not establish student success rates or a measured ranking against students [e1].
- The small, selected collection does not support statistically generalizable performance conclusions [e1].
Evidence (2 items)
Supports the claim: New York University; Ernest Davis and Scott Aaronson,
correct answer
Why it matters here: The report identifies correct high-dimensional probability and finite-field linear algebra answers alongside elementary failures. Their mathematical prerequisites support the qualitative difficulty contrast, but the authors explicitly report no human-subject testing.
- Publisher
- New York University; Ernest Davis and Scott Aaronson
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Supports the claim: New York University; Ernest Davis and Scott Aaronson,
not possible.
Why it matters here: Both configurations incorrectly deny that two people can each live near Lake Michigan while living far apart; the explanation treats the lake as a point. Other recorded failures concern elementary eclipse geometry and distances between landmarks.
- Publisher
- New York University; Ernest Davis and Scott Aaronson
- Kind of source
- Benchmark result
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Getting GPT to work with external tools is harder than you think
On some problems that capable students might well find hard, the AI did perfectly; on other problems that lot of mathematically-inclined college students people would find easy, the AI failed miserably.
- Where
- paragraph 10
- Published
- 2023-10-23
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-getting-gpt-to-work-with-external-c94be9-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-getting-gpt-to-work-with-external-c94be9-a2.