“The dirty little secret that pretty much everyone in the AI community knows is that doing well on a benchmark is not actually tantamount to being intelligent.”
Our reading, the sentence that was checked:
Doing well on a benchmark does not amount to being intelligent.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, No, Virginia, AGI is not imminent, paragraph 11.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-no-virginia-agi-is-not-imminent-e8ff52-a1
We recorded 19 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 2023-08-20. 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
- doing well on a benchmark
- not actually tantamount to being intelligent
- pretty much everyone in the AI community knows
Our interpretation
- In the AGI discussion, the plain reading is that a high benchmark score alone does not establish broad, reliable cognitive competence.
- This does not assert that benchmark performance provides no evidence of ability or that benchmark-tested systems possess no intelligence.
- The surrounding 2019-to-2023 discussion supplies the period for the empirical assessment.
- The community-awareness language is treated as emphasis; no literal near-unanimity estimate is inferred.
Counts as supported if: Systems with strong benchmark scores fail basic novel tests of abilities that those scores might reasonably be taken to establish.
Counts as contradicted if: Evidence establishes that the benchmark scores in question are sufficient to demonstrate the broad, reliable competence being inferred, including performance on novel tests.
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 behavioral tests show that strong benchmark results can coexist with failures on basic capabilities relevant to the same tasks. That supports the ordinary contextual claim that benchmark success alone does not establish general, reliable intelligence. It does not settle a philosophical definition of intelligence or prove near-universal agreement within the community.
Limitations
- The assessment depends on reading intelligent as broad, reliable competence in the surrounding AGI discussion.
- The evidence concerns language models and does not imply that benchmarks are useless.
Evidence (1 item)
Supports the claim: arXiv,
While useful, accuracy on benchmarks is not sufficient for evaluating NLP models.
Why it matters here: CheckList identifies basic linguistic failures in commercial and leading models across three tasks despite strong benchmark results. These empirical failures support the inference that benchmark success alone does not establish broad, reliable competence.
- Publisher
- arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, No, Virginia, AGI is not imminent
The dirty little secret that pretty much everyone in the AI community knows is that doing well on a benchmark is not actually tantamount to being intelligent.
- Where
- paragraph 11
- Published
- 2023-08-20
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-no-virginia-agi-is-not-imminent-e8ff52-a1: 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-no-virginia-agi-is-not-imminent-e8ff52-a1.