Pundit ClaimsAI claims by Gary Marcus and Ed Zitron, with the evidence

“Hallucinations haven’t been magically solved, etc.”

Grok 3 Hot Take, paragraph 9. The full statement is at the source.

Our reading, the sentence that was checked:

Grok 3 has not solved hallucinations.

Assessment: Supported Evidence as of

Where and when Gary Marcus wrote it

Marcus on AI, Grok 3 Hot Take, paragraph 9.

Said
Source observed
live page, full text, public
Evidence as of
the cutoff for this assessment
Assessment recorded
assessment marcus-grok-3-hot-take-29a51f-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 2025-02-18. 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 haven’t been magically solved
  • Grok 3

Our interpretation

  • The plain reading is that Grok 3 still produces fabricated factual content, including invented citations or URLs.
  • Magically is rhetorical emphasis, not a separate technical condition.
  • The product includes the DeepSearch capability discussed in the post.
  • For this ongoing-state claim, the target is the latest publicly observable version at launch on February 17–18, rather than later revisions.

Counts as supported if: Credible launch-period testing documents fabricated factual assertions, sources or URLs from Grok 3.

Counts as contradicted if: Comprehensive, independently validated launch-version testing establishes that factual fabrication has been eliminated within the ordinary product scope.

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

Launch-day hands-on testing reports that Grok 3's DeepSearch invented nonexistent URLs. That directly supports the modest assertion that hallucinations persisted. The evidence does not establish their frequency or whether Grok 3 hallucinated more or less than competing systems.

Limitations

  • The decisive evidence is an identified researcher's informal testing, not a controlled benchmark.
  • It concerns the DeepSearch product mode and does not isolate the underlying model from its tools.

Evidence (1 item)

  • Supports the claim: Andrej Karpathy,

    URLs that don't exist.

    Why it matters here: Karpathy directly reports fabricated URLs during early-access testing of Grok 3's DeepSearch. Such fabrication establishes that hallucinations persisted in the launch product.

    Publisher
    Andrej Karpathy
    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, Grok 3 Hot Take

    Hallucinations haven’t been magically solved, etc.

    Published
    2025-02-18
    Source observed
    , live page
    Access
    full text, public

History

Assessment versions

Recorded events

  1. Assessment publishedBy this site

    First assessment published after both reviews and the final-text check passed.

    About assessment marcus-grok-3-hot-take-29a51f-a1.