“I also warned that LLMs would continue to have a truthiness problem and they have.”
Our reading, the sentence that was checked:
LLMs have continued to have a truthfulness problem.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, The rapid rise and slow decline of Sam Altman, paragraph 19.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-the-rapid-rise-and-slow-decline-of-7121a9-a1
We recorded 18 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2026-01-13. 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
- LLMs
- continue to have a truthiness problem
- and they have
Our interpretation
- The ordinary reading is that LLMs still produce plausible but false factual answers.
- The retrospective period runs from the referenced January 2024 warning through the post date.
- The claim does not require hallucination rates to have remained constant or increased. The accuracy of Marcus's account of his earlier warning is outside the normalized proposition.
Counts as supported if: Evaluations or documented outputs show that contemporary, widely used LLMs continued to generate confident factual errors.
Counts as contradicted if: Representative evaluations establish that plausible factual fabrication had ceased to be a material problem in contemporary LLMs.
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
OpenAI acknowledged that GPT-5 still hallucinated in September 2025. That directly supports the ordinary reading that plausible false answers remained a problem. The source also reports improvement; Marcus's sentence does not claim that accuracy failed to improve.
Limitations
- This establishes persistence, not a uniform error rate across models or tasks.
- The decisive item is a developer disclosure rather than an independent replication.
Evidence (1 item)
Supports the claim: OpenAI,
but they still occur.
Why it matters here: OpenAI explicitly acknowledges persistent hallucinations in GPT-5. The same sentence reports improvement, which is compatible with Marcus's claim that the problem continued.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, The rapid rise and slow decline of Sam Altman
I also warned that LLMs would continue to have a truthiness problem and they have.
- Where
- paragraph 19
- Published
- 2026-01-13
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-the-rapid-rise-and-slow-decline-of-7121a9-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-the-rapid-rise-and-slow-decline-of-7121a9-a1.