“It’s also, as ever, stuck in the liminal space between superintelligence and super-lost.”
Our reading, the sentence that was checked:
Make a Video, as with AI systems before it, is caught between very impressive output and basic errors, such as giving a cat a human hand or a painting that never changes.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, DeepFakes versus Deep Understanding, paragraph 3.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-deepfakes-versus-deep-understanding-dbe10a-a2
We recorded 4 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-29. The source may have been edited after 2022-10-02. 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 ever (qualifier, stated by the speaker)
From the speaker
- MetaAI’s new text-to-movie software Make a Video
- straight-up amazing
- between superintelligence and super-lost
- A cat holding a remote control shouldn’t have a human hand.
- the painting never changed
Our interpretation
- The ordinary reading is that recognizable, impressive generated scenes also contain basic semantic or physical errors. The reference to superintelligence is figurative.
- The target is the release demonstrations available when Marcus wrote, from September 29 through October 2, 2022.
- The examples illustrate the broader observation; support need not establish that every demonstration fails. Nevertheless, a general technical limitation alone does not verify a visible error.
- The phrase 'as ever' supplies rhetorical continuity, not a separately measurable claim about every earlier AI system.
Counts as supported if: The original demonstrations, or reliable independent descriptions of them, establish recognizable generated scenes alongside basic semantic or physical errors.
Counts as contradicted if: Reliable inspection establishes that the alleged defects are absent or misdescribed and that the cited demonstrations do not exhibit comparable basic errors.
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
Meta announced Make-A-Video on September 29, 2022, describing its output as brief, high-quality video clips [e1]. Its September 2022 paper acknowledged that the approach could not learn associations between text and phenomena inferable only from video [e2]. Neither item establishes whether the release demonstrations showed a human-handed cat, an unchanged portrait during painting, or comparable basic errors, so the observation remains unresolved [e1, e2].
Limitations
- The original moving clips were not independently inspected in this assessment.
- The evidence contains no independent confirmation or refutation of the alleged visual defects.
- Meta's description of quality and its acknowledged technical limitation do not establish the quality or errors of the specific demonstrations.
Evidence (2 items)
Context only: Meta AI,
brief, high-quality video clips.
Why it matters here: Establishes the release date and the company's description of the capability. It does not independently establish quality or the alleged errors.
- Publisher
- Meta AI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Context only: arXiv; Meta research authors,
can not learn associations
Why it matters here: The authors acknowledge a gap in learning relationships specific to moving scenes. That makes Marcus's interpretation plausible but does not establish what happened in the cat or portrait clips.
- Publisher
- arXiv; Meta research authors
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
What would settle it
Inspection across frames of the original September 29–October 2, 2022 release demonstrations, or reliable independent descriptions, establishing whether recognizable scenes contain the alleged or comparable basic errors. The assessment must distinguish intentional surreal content from unintended inconsistency.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, DeepFakes versus Deep Understanding
It’s also, as ever, stuck in the liminal space between superintelligence and super-lost.
- Where
- paragraph 3
- Published
- 2022-10-02
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-deepfakes-versus-deep-understanding-dbe10a-a2: 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.
About assessment marcus-deepfakes-versus-deep-understanding-dbe10a-a2.