“Time and money haven’t robustly solved these problems.”
Our reading:
Between February and December 2024, time and money did not robustly solve Sora's physics problems.
Assessment: Supported
What happened
Was it true for the period described?
Outcome: Supported
The evidence matches what they said.
Judged on a plain reading. They named no measure, so we used the reading a typical informed reader would take. It is spelled out under “How we judged it”.
Evidence up to . Checked .
Why
OpenAI's December 9, 2024 launch announcement said the deployed Sora version often generated unrealistic physics and struggled with complex actions over long durations [e1]. It also reported that Sora Turbo was significantly faster than the February preview [e1]. The acknowledged recurring physics failures support the claim that these problems had not been robustly solved by December 2024 [e1].
Caveats
- The evidence concerns the December 2024 release, not later versions.
- It does not quantify changes in physics reliability or separately establish the frequency of object-continuity failures.
- It establishes neither spending totals nor the causal effect of additional time or money.
Evidence (1 item)
Supports the claim: OpenAI,
often generates unrealistic physics
Why it matters: OpenAI describes physical errors as frequent in the deployed version, which supports a lack of reliable resolution. The announcement also identifies development since February and improved speed, so it supports continuing limitations without implying that development achieved nothing.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- About events on
- Published
How we judged it
- Kind of claim
- Observation
- Deadline
- None given. We don’t add one.
Where we had to interpret
- The context compares the February preview with the December release, ending at the December 10 post.
- The ordinary reading of 'robustly solved' is reliable handling of physical interactions and object continuity, rather than occasional successful outputs.
- This is an observation about the result of development to date. It does not claim that progress was absent, identify a spending amount, or estimate the causal effect of additional spending.
- Because no financial quantity is asserted, metric_scope is null.
Right if: Evidence about the December release establishes that physical inconsistencies or continuity failures remain recurring limitations after the development interval.
Wrong if: Evidence about the December release establishes reliable handling of the relevant physical interactions and object continuity, with the earlier failure modes substantially resolved rather than remaining recurring limitations.
We wrote this test after the outcome was known.
We found 8 claims in this post. See them all.
Full record
Source and dates
Marcus on AI, Sora still appears to have trouble with physics, paragraph 10.
- Said
- We read the source
- live page, full text, public
- Evidence up to
- Checked
We read this wording on 2026-10-28. The post may have been edited after 2024-12-10.
Where it was said (1)
Said once in the posts we read.
: Marcus on AI, Sora still appears to have trouble with physics
Time and money haven’t robustly solved these problems.
- Where
- paragraph 10
- Published
- 2024-12-10
- We read it
- , live page
- Access
- full text, public
History
Versions
- Current: Supported. Evidence up to 2026-10-28.
What happened to this record
PublishedBy us
Published after both AI reviews and the final check passed.