“Furthermore, OpenAI apparently hasn’t even fulfilled their own promises to devote 20% resources to AI safety.”
Our reading:
OpenAI apparently has not fulfilled its promise to devote 20% of resources to AI safety.
Assessment: Can’t be checked
What happened
Not graded on evidence
Outcome: Can’t be checked
An opinion, a recommendation, or too vague to test. Listed, not scored.
Evidence up to . Checked .
Why
Marcus says “apparently” and leaves both resources and the timing of fulfilment unspecified. The likely underlying pledge covered compute over four years. Contemporary testimony about compute shortages provides a basis for his concern, but does not establish that a defined allocation was already overdue. Calling the statement supported or contradicted would require supplying terms he did not state. This is not a finding that his concern was unfounded.
Caveats
- The evidence documents a pledge and reported shortages, not a complete allocation audit.
- The four-year pledge is contextual evidence, not a deadline attributed to Marcus.
Evidence (2 items)
Claims like this aren’t scored, so there is no evidence to show.
How we judged it
- Kind of claim
- Observation
- Deadline
- None given. We don’t add one.
Conditions
- apparently (their words)
Where we had to interpret
- Marcus makes a qualified allegation about the situation at publication, not a forecast or an unqualified finding of breach.
- The Superalignment compute pledge is a plausible referent, but identifying resources exclusively with that compute is an interpretation.
- Neither the quotation nor its context specifies an allocation schedule or an amount already due by May 23, 2024.
- No reporting period or fulfilment deadline is imposed because the statement cannot be given a binary test without resolving these ambiguities for the speaker.
We don’t grade claims like this. We only check the quote and who said it.
What would settle it
A clarified claim identifying the resources, allocation schedule, and obligation allegedly unmet by May 23, 2024, followed by comparable allocation records.
We found 4 claims in this post. See them all.
Full record
Source and dates
Marcus on AI, ScarJo is just the tip of the iceberg, paragraph 4.
- Said
- We read the source
- live page, full text, public
- Evidence up to
- Checked
We read this wording on 2026-10-29. The post may have been edited after 2024-05-23.
Where it was said (1)
Said once in the posts we read.
: Marcus on AI, ScarJo is just the tip of the iceberg
Furthermore, OpenAI apparently hasn’t even fulfilled their own promises to devote 20% resources to AI safety.
- Where
- paragraph 4
- Published
- 2024-05-23
- We read it
- , live page
- Access
- full text, public
History
Versions
- Current: Can’t be checked. Evidence up to 2026-10-28.
What happened to this record
PublishedBy us
Published after both AI reviews and the final check passed.
Held backBy us
A reviewer said our test wasn’t fair to what was actually said.