“Any fool can simply eyeball these data and see that in virtually every condition, people who have access both to the internet and LLMs (darker bars) are more accurate than people without (lighter bars).”
Our reading, the sentence that was checked:
In OpenAI's study, participants with both internet and LLM access were more accurate than those without in virtually every condition.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Diving deep into OpenAI’s new study on LLM’s and bioweapons, paragraph 10.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-when-looked-at-carefully-openais-596625-a1
We recorded 22 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 2024-02-04. 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
- in virtually every condition of the study (qualifier, stated by the speaker)
From the speaker
- virtually every condition
- access both to the internet and LLMs
- more accurate
Our interpretation
- Conditions means the five accuracy tasks in each of the two cohorts: ten comparisons.
- The comparison group had internet access without the LLM.
- More accurate means a higher observed group mean, without requiring statistical significance.
- Virtually every means all or all but one of these ten comparisons.
Counts as supported if: The LLM-plus-internet mean exceeds the internet-only mean in at least nine of the ten cohort-by-task comparisons.
Counts as contradicted if: Several comparisons fail to show a higher mean, leaving fewer than nine positive differences.
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 released results show higher mean accuracy with LLM access in nine of ten comparisons. That matches the ordinary meaning of virtually every condition. The single negative comparison is already allowed by Marcus's wording.
Limitations
- Higher observed means do not establish significance or practical danger.
- The numerical source contains CSV records rather than prose sentences; the quotation preserves a complete record.
Evidence (1 item)
Supports the claim: OpenAI,
,Mean difference,0.2400,-1.0800,0.9600,0.8800,0.2400,1.2400,0.7200,1.2400,1.1200,1.0800,0.6800,4.8400
Why it matters here: Nine task differences are positive. Student acquisition is the sole exception, with a difference of −1.08 points.
- Where
- Mean difference row; five Student task columns and five Expert task columns, excluding Total columns
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Arithmetic
- Count differences greater than zero: 4 student comparisons + 5 expert comparisons = 9 of 10. (inputs: Student differences: 0.24, -1.08, 0.96, 0.88, 0.24; Expert differences: 0.72, 1.24, 1.12, 1.08, 0.68; No rounding in the count.)
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Diving deep into OpenAI’s new study on LLM’s and bioweapons
Any fool can simply eyeball these data and see that in virtually every condition, people who have access both to the internet and LLMs (darker bars) are more accurate than people without (lighter bars).
- Where
- paragraph 10
- Published
- 2024-02-04
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-when-looked-at-carefully-openais-596625-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-when-looked-at-carefully-openais-596625-a1.