“Hallucinations and boneheaded errors of reasoning and what I like to call discomprehensions are, as far as I can tell, an inevitable byproduct of LLMs; certainly they have yet to disappear, despite many promises to the contrary.”
Our reading, the sentence that was checked:
Marcus says that, as far as he can tell, hallucinations, boneheaded reasoning errors and 'discomprehensions' are an inevitable byproduct of LLMs, and that they certainly have not yet disappeared despite many promises to the contrary.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, AlphaProof, AlphaGeometry, ChatGPT, and why the future of AI is neurosymbolic, paragraph 39.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-alphaproof-alphageometry-chatgpt-a8a13f-a1
We recorded 17 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 2024-07-28. 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
- as far as I can tell (qualifier, stated by the speaker)
From the speaker
- Hallucinations and boneheaded errors of reasoning
- as far as I can tell, an inevitable byproduct of LLMs
- certainly they have yet to disappear
- despite many promises to the contrary
Our interpretation
- The checkable firm part is 'they have yet to disappear': as of the post date, leading LLM systems still hallucinated and made reasoning errors.
- The 'inevitable byproduct' part is hedged ('as far as I can tell') and is treated as opinion, not scored.
- Period: 2024-01-01 to the post date (2024-07-28), the most recent stretch before the post.
Counts as supported if: Evaluations or documented incidents from 2024 show leading LLM products still hallucinating, including products that claimed to avoid hallucinations.
Counts as contradicted if: Evidence that leading LLM systems in mid-2024 had essentially eliminated hallucinations and reasoning errors.
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 firm part of the claim is that hallucinations had not disappeared by mid-2024, and it holds. In May 2024, Stanford researchers found that commercial legal AI tools sold as avoiding hallucinations still gave wrong information on more than 17% to 34% of queries. In late July 2024, OpenAI's SearchGPT gave wrong festival dates in its own launch demo. The 'inevitable byproduct' part is hedged ('as far as I can tell') and is treated as opinion, not scored.
Limitations
- Only the 'yet to disappear' part is scored.
- The evidence covers selected products and domains.
Evidence (2 items)
Supports the claim: Stanford HAI / RegLab,
But even these bespoke legal AI tools still hallucinate an alarming amount of the time: the Lexis+ AI and Ask Practical Law AI systems produced incorrect information more than 17% of the time, while Westlaw’s AI-Assisted Research hallucinated more than 34% of the time.
Why it matters here: In mid-2024, LLM products whose providers claimed they 'avoid' hallucinations still hallucinated on 17% to 34% or more of test queries.
- Publisher
- Stanford HAI / RegLab
- Kind of source
- Independent measurement
- Describes events of
- Public from
Supports the claim: TechRadar,
In a prerecorded video showcasing SearchGPT, the tool provided incorrect information about the dates of the Appalachian Summer Festival in Boone, North Carolina.
Why it matters here: OpenAI's newest system made a factual error in its launch demo days before the post.
- Where
- paragraph 2
- Publisher
- TechRadar
- Kind of source
- News report
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, AlphaProof, AlphaGeometry, ChatGPT, and why the future of AI is neurosymbolic
Hallucinations and boneheaded errors of reasoning and what I like to call discomprehensions are, as far as I can tell, an inevitable byproduct of LLMs; certainly they have yet to disappear, despite many promises to the contrary.
- Where
- paragraph 39
- Published
- 2024-07-28
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-alphaproof-alphageometry-chatgpt-a8a13f-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-alphaproof-alphageometry-chatgpt-a8a13f-a1.