“You simply cannot fix them. Hallucinations are not going away.”
Our reading, the sentence that was checked:
Hallucinations in generative AI models cannot be fixed and will not go away.
Assessment: Unresolved No deadline stated on 2026-10-29 Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, Have We Reached Peak AI?, paragraph 33.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-peakai-5b807b-a1
We recorded 16 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 2024-03-18. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Forecast
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No deadline stated
- Deadline state on 2026-10-29
- No deadline stated (the release’s reference date)
From the speaker
- You simply cannot fix them
- Hallucinations are not going away
Our interpretation
- This is an unconditional, open-ended claim that hallucinations cannot be eliminated.
- The ordinary reading concerns persistent fabrication in useful general-purpose generative systems. It does not require answering every possible question correctly; appropriate abstention may be part of a solution.
- Lower error rates, perfect scores on a finite narrow test, and systems that always abstain do not by themselves establish elimination of the general problem.
Counts as supported if: A decisive impossibility result establishes that hallucinations cannot be eliminated within the claimed scope. Continued hallucinations in a finite set of models are insufficient to prove 'cannot' or 'will not'.
Counts as contradicted if: A publicly available, useful general-purpose generative system is independently and reproducibly shown to eliminate hallucinations across its intended broad uses, including appropriate handling of uncertainty.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Did the forecast come true?
Outcome: Unresolved
The record was checked, and the evidence available by the cutoff does not settle it.
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 deadline stated
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
The forecast has no deadline. Research documents persistent hallucinations, but that cannot prove they will never disappear. Work on abstention and bounded solutions also prevents treating inevitability as established. No cited event meets the early-resolution rule for either a general solution or a decisive impossibility proof.
Limitations
- An open-ended impossibility claim cannot be supported merely by waiting.
- The result depends on distinguishing a general solution from narrow task-specific guarantees.
- The search does not establish that no qualifying solution could exist.
Evidence (3 items)
Context only: Nature,
Nonetheless, the problem persists even in state-of-the-art language models6,7.
Why it matters here: The paper documents continued hallucinations. Persistence by this date does not establish that eliminating them is impossible.
- Where
- Abstract
- Publisher
- Nature
- Kind of source
- Peer-reviewed paper
- Describes events of
- Public from
Context only: OpenAI,
Finding: They are not, because language models can abstain when uncertain.
Why it matters here: This is material counterevidence to a categorical inevitability argument. It identifies abstention as a possible remedy but does not independently demonstrate a broadly useful system that has eliminated hallucinations.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Context only: arXiv; Yuke Li and Xuehan Hou,
translating abstract user instructions into precise element coordinates
Why it matters here: Despite its hallucination-free title, the proposed method addresses GUI coordinate grounding. That bounded scope does not satisfy the rule for eliminating hallucinations in general-purpose generation.
- Where
- Abstract
- Publisher
- arXiv; Yuke Li and Xuehan Hou
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
What would settle it
Independent, reproducible evidence of a useful general-purpose system that eliminates hallucinations under broad deployment conditions, or an applicable impossibility proof covering that scope.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, Have We Reached Peak AI?
You simply cannot fix them. Hallucinations are not going away.
- Where
- paragraph 33
- Published
- 2024-03-18
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-peakai-5b807b-a1: Current. Outcome: Unresolved. 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.