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

“In particular, DALL-E2 was very poor at following specifications about parts of objects.”

Don’t Ride This Bike! Generative AI’s persistent trouble with compositionality and parts, paragraph 6. The full statement is at the source.

Our reading, the sentence that was checked:

As of October 2022, DALL-E2 was very poor at following prompt specifications about parts of objects.

Assessment: Unresolved Evidence as of

Where and when Gary Marcus wrote it

Marcus on AI, Don’t Ride This Bike! Generative AI’s persistent trouble with compositionality and parts, paragraph 6.

Said
Source observed
live page, full text, public
Evidence as of
the cutoff for this assessment
Assessment recorded
assessment marcus-dont-ride-this-bike-generative-ais-7975a8-a1

We recorded 10 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 2024-12-08. 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)

Conditions

  • as of October 2022 (qualifier, stated by the speaker)

From the speaker

  • DALL-E2
  • very poor at following specifications about parts of objects
  • October 2022

Our interpretation

  • The ordinary reading is that explicit requirements about object parts frequently failed across varied prompts.
  • Success requires satisfying the specified part attributes, counts, substitutions, or omissions. General image quality does not establish success.
  • October 2022 is the observation period supplied by the surrounding passage.

Counts as supported if: Contemporaneous, independently assessed results show frequent failures across varied prompts specifying object parts.

Counts as contradicted if: Comparable October 2022 results show that DALL-E2 usually satisfied explicit object-part requirements.

Retrospective: these criteria were written after the outcome was public.

Current assessment

Was it true for the period described?

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 horizon
Assessment recorded
Policy version
2026-10-28.7

Rationale

Independent research documents related compositional weaknesses, but it concerns different tasks and an earlier period. The original October record was located, but its text and the authors' conclusions do not independently establish the claimed performance. The evidence therefore does not settle this specific observation.

Limitations

  • This is an evidence gap, not evidence that DALL-E2 performed well.
  • The assessment does not generalize results about object relations to object parts.

Evidence (1 item)

  • Context only: arXiv; Colin Conwell and Tomer Ullman,

    ~22% of images matched basic relation prompts.

    Why it matters here: This documents difficulty with relations between objects. It does not measure October 2022 performance on specifications about object parts.

    Publisher
    arXiv; Colin Conwell and Tomer Ullman
    Kind of source
    Preprint, not peer reviewed
    Describes events of
    Public from

What would settle it

An independent audit of the dated October outputs, or a contemporaneous evaluation covering varied object-part specifications.

Where it was said (1)

Said once in the sources this register covers.

  • : Marcus on AI, Don’t Ride This Bike! Generative AI’s persistent trouble with compositionality and parts

    In particular, DALL-E2 was very poor at following specifications about parts of objects.

    Published
    2024-12-08
    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-dont-ride-this-bike-generative-ais-7975a8-a1.