“The problem of hallucinations is not solved; reliability is not solved; planning on complex tasks is (as the authors themselves acknowledge) not solved.”
Our reading, the sentence that was checked:
GPT-4 has not solved hallucinations, reliability, or planning on complex tasks.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, The Sparks of AGI? Or the End of Science?, paragraph 18.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-the-sparks-of-agi-or-the-end-of-science-16c15a-a1
We recorded 6 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 2023-03-24. 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
- hallucinations is not solved
- reliability is not solved
- planning on complex tasks
- not solved
Our interpretation
- Read as saying material factual, reasoning, and multistep-planning failures remained, rather than denying improvements.
- Hallucinations and invalid plans are evidence of unreliability; these are overlapping assertions.
- Use the March 2023 launch-period findings available before March 24, including the early model studied in Sparks.
Counts as supported if: Documented hallucinations and planning failures show that GPT-4 remained materially unreliable in these areas.
Counts as contradicted if: Evidence establishes that the contemporaneous model had overcome the identified factual and planning failure modes across suitably varied tests.
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 remaining hallucinations at launch. The original Sparks paper also documented a failed multistep plan. Together these establish that factual reliability and planning remained unsolved in the ordinary sense Marcus used. They do not imply that GPT-4 failed every planning task or had made no progress.
Limitations
- The planning example concerns the early research model examined in Sparks.
- These items establish persistent failures, not their frequency across all uses.
Evidence (2 items)
Supports the claim: OpenAI,
hallucinations, and adversarial prompts.
Why it matters here: OpenAI acknowledged that hallucinations remained at launch. Persisting fabricated answers support the related claim that factual reliability was not solved.
- Where
- Limitations
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Supports the claim: Microsoft Research researchers via arXiv,
the model gets it wrong.
Why it matters here: The paper documents an invalid plan for a problem requiring dependent moves. This supports the persistence of planning failures and the broader reliability concern.
- Publisher
- Microsoft Research researchers via arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, The Sparks of AGI? Or the End of Science?
The problem of hallucinations is not solved; reliability is not solved; planning on complex tasks is (as the authors themselves acknowledge) not solved.
- Where
- paragraph 18
- Published
- 2023-03-24
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-the-sparks-of-agi-or-the-end-of-science-16c15a-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-sparks-of-agi-or-the-end-of-science-16c15a-a1.