“It doesn't solve hallucinations or problems with reliability.”
Our reading, the sentence that was checked:
DeepSeek r1 does not solve hallucinations or reliability problems.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Five things most people don't seem to understand about DeepSeek, paragraph 2.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-five-things-most-people-dont-seem-570c01-a1
We recorded 9 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 2025-01-28. 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)
From the speaker
- It doesn't solve hallucinations
- or problems with reliability
Our interpretation
- It refers to the original full DeepSeek-R1 model.
- The ordinary reading is that factual fabrication and unreliable behavior remain practical problems. The claim does not say R1 never improves accuracy.
- Use the original release-to-post period as the latest publicly assessable period for this ongoing observation.
Counts as supported if: Documented release-era behavior shows factual fabrication and continuing failures of dependable task performance.
Counts as contradicted if: Broad, reproducible release-era evaluations establish that factual fabrication and ordinary reliability failures have been effectively resolved.
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
Release-era evidence documents both unreliable prompt behavior and a fabricated self-identification. These support the narrow claim that hallucination and reliability problems remained. The later independent study corroborates that reading but is not needed to place an outcome inside the target period.
Limitations
- The hallucination example comes from a reported reasoning trace, rather than a systematic release-period benchmark.
- This does not establish that R1 was less reliable than every predecessor or that it made no accuracy improvements.
Evidence (3 items)
Supports the claim: DeepSeek-AI / arXiv,
Few-shot prompting consistently degrades its performance.
Why it matters here: DeepSeek reports prompt sensitivity in the released R1 model. This establishes a continuing reliability limitation, though it does not itself measure hallucinations.
- Publisher
- DeepSeek-AI / arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Supports the claim: The Register,
My official name is Claude, created by Anthropic.
Why it matters here: The report quotes R1 fabricating its identity in its reasoning trace. The original issue's screenshot was not available as plain text, so this item relies on the contemporaneous news transcription; it establishes an example, not a population failure rate.
- Publisher
- The Register
- Kind of source
- News report
- Describes events of
- Public from
Context only: Vectara,
The results were surprising: Deepseek-R1’s hallucination rate was 14.3% - much higher than its non-reasoning predecessor: Deepseek-V3.
Why it matters here: A subsequent summarization evaluation found continuing hallucinations. The precise test dates are unspecified, so this is corroborating context rather than decisive evidence for the January 20–28 period.
- Publisher
- Vectara
- Kind of source
- Independent measurement
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Five things most people don't seem to understand about DeepSeek
It doesn't solve hallucinations or problems with reliability.
- Where
- paragraph 2
- Published
- 2025-01-28
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-five-things-most-people-dont-seem-570c01-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-five-things-most-people-dont-seem-570c01-a1.