“The company has not demo’d a working, reliable agent, and does not seem to have GPT-5 level model thus far.”
Our reading:
OpenAI has not demonstrated a working, reliable agent and does not appear to have a GPT-5-level model as of May 2024.
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
The sentence should not receive a single factual grade as written. Its agent-demonstration clause is checkable, but the GPT-5 clause explicitly says does not seem and supplies no capability threshold. Treating it as a firm assertion about possession or public demonstration would change Marcus's commitment. The primary sources document assistant functionality and caution about future agents, but do not independently settle the agent clause's reliability requirement.
Caveats
- This does not declare the agent clause uncheckable or supported.
- A product announcement alone neither proves reliability nor proves that no reliable demonstration existed.
- Later releases cannot turn the hedged May 2024 impression into a firm historical assertion.
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.
Where we had to interpret
- The company is OpenAI, and thus far ends on May 5, 2024.
- The first clause is a checkable historical assertion about public demonstrations. Working and reliable would ordinarily require repeatable completion of the demonstrated class of tasks, rather than one selected successful run.
- The second clause explicitly describes an impression: does not seem to have. It neither asserts that OpenAI possessed no successor model nor merely asserts that none had been publicly demonstrated.
- No factual success or failure criterion is assigned to the hedged successor-model clause. The plain-reading exception does not remove that hedge.
- Because this record combines the firm demonstration claim with the explicitly hedged assessment, no single factual outcome is assigned to the whole sentence. The agent clause remains independently checkable.
We don’t grade claims like this. We only check the quote and who said it.
What would settle it
Separate the firm agent-demonstration clause for historical assessment and obtain dated demonstration and reliability records. Leave the hedged GPT-5 clause unscored unless the speaker supplies a definite factual commitment.
We found 19 claims in this post. See them all.
Full record
Source and dates
Marcus on AI, The Sam Altman Playbook, paragraph 52.
- 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 2024-05-05.
Where it was said (1)
Said once in the posts we read.
: Marcus on AI, The Sam Altman Playbook
The company has not demo’d a working, reliable agent, and does not seem to have GPT-5 level model thus far.
- Where
- paragraph 52
- Published
- 2024-05-05
- 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.