“The technical problems are not new.”
Our reading, the sentence that was checked:
The core technical problems of LLMs are not new.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, The AI bubble is all over now, baby blue, paragraph 8.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-the-ai-bubble-is-all-over-now-baby-d891b6-a1
We recorded 7 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 2025-12-26. 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
- The technical problems
- are not new
- since 2019
Our interpretation
- The ordinary reading is that factual unreliability and failures of reasoning were documented years before this post.
- The historical comparison includes the period beginning in 2019 mentioned in the surrounding text. It does not require every current failure to have been identified in 2019.
- This claim concerns the age of the problems, not whether they are permanent or whether performance has improved.
Counts as supported if: Dated research from years before publication documents hallucinations and reasoning failures in large language models or their direct predecessors.
Counts as contradicted if: Historical records establish that the relevant problems first appeared near publication and were materially different from earlier documented failures.
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
Research published in 2020 documented both hallucinated content and basic reasoning failures. That supports the ordinary historical claim that these reliability problems were not new in December 2025. It does not establish that they are inevitable or unchanged.
Limitations
- The evidence establishes continuity in broad failure categories, not identical causes across model generations.
- It does not assess Marcus's separate claim that world models are necessary for reliability.
Evidence (2 items)
Supports the claim: arXiv; Joshua Maynez and colleagues,
hallucinated content
Why it matters here: The authors documented fabricated or unsupported content in neural summarization systems more than five years before the post. These systems are predecessors of current LLM applications.
- Where
- Abstract
- Publisher
- arXiv; Joshua Maynez and colleagues
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Supports the claim: arXiv; Tom B. Brown and colleagues,
GPT-3 has difficulty
Why it matters here: The GPT-3 paper describes failures involving ordinary physical reasoning, alongside inconsistent and incoherent generation. This establishes that reliability problems were already documented in a large autoregressive language model.
- Publisher
- arXiv; Tom B. Brown and colleagues
- 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 AI bubble is all over now, baby blue
The technical problems are not new.
- Where
- paragraph 8
- Published
- 2025-12-26
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-the-ai-bubble-is-all-over-now-baby-d891b6-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-ai-bubble-is-all-over-now-baby-d891b6-a1.