“More data won’t solve [Sora’s] problem [with physics].”
Our reading, the sentence that was checked:
More training data will not solve Sora's problems with physics.
Assessment: Unresolved No deadline stated on 2026-10-29 Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Sora still appears to have trouble with physics, paragraph 9.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-sora-still-appears-to-have-trouble-969209-a1
We recorded 6 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 2024-12-10. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Forecast
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No deadline stated
- Deadline state on 2026-10-29
- No deadline stated (the release’s reference date)
From the speaker
- More data
- won’t solve
- [Sora’s] problem [with physics]
Our interpretation
- The plain reading is that increasing training data alone will not make Sora reliably physically consistent across varied scenes.
- 'Solve' means reliable physical consistency, not merely better average performance or a few successful demonstrations.
- A causal comparison must distinguish additional data from changes to architecture, training objectives, post-training, or external physical constraints.
- The February prediction is reaffirmed in the December post without a deadline.
Counts as supported if: A decisive technical result establishes that increasing data alone cannot resolve the relevant physics failures within the specified Sora model family. Repeated failures at finite scales are suggestive but cannot settle an unlimited forecast.
Counts as contradicted if: A controlled, reproducible Sora comparison shows that additional training data alone produces reliable physical consistency across varied, previously unseen situations.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Did the forecast come true?
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 deadline stated
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
The available research documents scaling limits in other models, while OpenAI reports improved physics in Sora 2. Neither establishes whether additional data alone can solve Sora's physics problems. Persistent errors cannot settle a forecast with no deadline, and the evidence does not meet the early-resolution rule.
Limitations
- The controlled research concerns different models.
- The Sora 2 evidence is a company disclosure rather than a controlled causal evaluation.
- Improvement and complete resolution are different claims.
Evidence (2 items)
Context only: Kang and colleagues, arXiv,
failure
Why it matters here: The experiments show both benefits and limits of scaling in a controlled physical simulation setting. They do not test Sora or establish what all future data increases can achieve.
- Where
- Abstract
- Publisher
- Kang and colleagues, arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: OpenAI,
rebound off the backboard.
Why it matters here: OpenAI reports an improvement directly relevant to physical consistency. The announcement also acknowledges remaining imperfections and does not isolate additional training data as the cause.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
What would settle it
A reproducible Sora data-only scaling study demonstrating reliable physical consistency across broad unseen tests, or a technical impossibility result that covers the specified model family.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Sora still appears to have trouble with physics
More data won’t solve [Sora’s] problem [with physics].
- Where
- paragraph 9
- Published
- 2024-12-10
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-sora-still-appears-to-have-trouble-969209-a1: 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-sora-still-appears-to-have-trouble-969209-a1.