“It is shocking that neither OpenAI nor the media noted this (they also did not discuss in any detail one other paper that reached a somewhat similar conclusion, by Kevin Esvelt’s lab).”
Our reading, the sentence that was checked:
Neither OpenAI nor the media noted the fourfold difference in expert success on pathogen formulation, nor discussed in detail a similar paper from Kevin Esvelt's lab.
Assessment: Unresolved 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 29.
- 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-b51513-a2
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)
From the speaker
- neither OpenAI nor the media noted this
- did not discuss in any detail
- one other paper
- Kevin Esvelt’s lab
Our interpretation
- This refers to the fourfold formulation threshold result.
- Read media as contemporary coverage of the OpenAI study through February 4, 2024, rather than all earlier reporting about Esvelt.
- The post links the Esvelt reference to Soice et al., Can large language models democratize access to dual-use biotechnology?
- Read noted as explicitly drawing attention to the fourfold comparison. Merely publishing recoverable counts is a different, plausible reading.
- Detailed discussion means substantive treatment of the Esvelt study's methods or findings, beyond a citation.
- The word shocking is an opinion and is not assessed.
Counts as supported if: A sufficiently comprehensive contemporary record shows neither an explicit fourfold comparison nor substantive discussion of the linked Esvelt study by OpenAI or the relevant media.
Counts as contradicted if: A qualifying contemporary OpenAI publication or media report explicitly highlights the fourfold comparison or substantively discusses the linked Esvelt study.
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 January 31, 2024 Bloomberg report neither highlights the fourfold formulation comparison nor discusses the linked Esvelt-lab paper [e3]. The February 1 Verge report does not identify the fourfold comparison, although it discusses other benefits from model access and questions OpenAI's framing [e4]. These articles establish omissions in those reports, but cannot establish the broader claim about OpenAI and contemporary media through February 4, 2024 [e3][e4].
Limitations
- The supplied evidence items omit OpenAI's publication and do not provide a sufficiently comprehensive contemporary media record.
- The assessment interprets 'noted' as explicitly highlighting the fourfold comparison and 'in any detail' as substantive discussion of the linked study's methods or findings.
Evidence (2 items)
Context only: Bloomberg, syndicated by Yahoo Finance,
The other group was just given internet access to complete the exercise.
Why it matters here: Inspection of the complete article found neither the fourfold formulation comparison nor discussion of Soice et al. This establishes only what this particular report omitted.
- Publisher
- Bloomberg, syndicated by Yahoo Finance
- Kind of source
- News report
- Describes events of
- Public from
Context only: The Verge,
Researchers also found that participants who relied on GPT-4 had more detailed answers.
Why it matters here: The full article discusses benefits and challenges OpenAI's framing, but does not identify the specific fourfold comparison. It cautions against extending Marcus's narrow omission claim into a claim that media ignored all concerning findings.
- Publisher
- The Verge
- Kind of source
- News report
- Describes events of
- Public from
What would settle it
A qualifying OpenAI publication or media report available before the February 4, 2024 post that explicitly highlights the fourfold comparison or substantively discusses the linked Esvelt study would contradict the claim. Support would require complete relevant OpenAI materials and a sufficiently comprehensive, defined contemporary media record showing neither treatment.
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
It is shocking that neither OpenAI nor the media noted this (they also did not discuss in any detail one other paper that reached a somewhat similar conclusion, by Kevin Esvelt’s lab).
- Where
- paragraph 29
- Published
- 2024-02-04
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-when-looked-at-carefully-openais-b51513-a2: 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.
About assessment marcus-when-looked-at-carefully-openais-b51513-a2.