“The progress that Alexander is vaunting shows that the system did better on a handful of items than DALL-E2 (which was not even state of the art at compositionality when it appeared), but not that the newer system (Imagen) systematically groks compositionality”
Our reading, the sentence that was checked:
Alexander's test shows only that Imagen beat DALL-E 2 on a handful of items, not that Imagen systematically understands compositionality.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Did GoogleAI Just Snooker One of Silicon Valley’s Sharpest Minds?, paragraph 50.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-did-googleai-just-snooker-one-of-79d6ee-a2
We recorded 12 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 2022-09-14. 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
- did better on a handful of items than DALL-E2
- not that the newer system (Imagen) systematically groks compositionality
- none of which Alexander tested
Our interpretation
- Assess the normalized proposition about what Alexander's September 2022 comparison establishes.
- The plain reading of systematic compositionality is reliable handling of new combinations of objects, attributes and relations across varied prompts.
- A small comparison selected from earlier failures can establish local improvement without establishing that broader ability.
- The suggestion of possible regressions is a statement about an untested possibility, not a claim that regressions occurred.
Counts as supported if: The published comparison uses a small set of earlier failure cases and lacks the coverage and controls needed to establish reliable compositional generalization.
Counts as contradicted if: The cited comparison includes a broad, systematically selected evaluation demonstrating reliable compositional generalization, including cases the earlier model handled successfully.
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
Alexander's September 12, 2022 comparison used five prompts and ten images per prompt, reporting DALL-E 2 at 0/5 and Imagen at 3/5 under his scoring [e1]. Imagen's prompts replaced humans with robots because of its restrictions [e1]. The comparison did not systematically retest earlier successes or cover varied new compositions [e1]. It supports improvement on those selected cases, but does not establish systematic compositional ability or exclude regressions elsewhere [e1].
Limitations
- The scores reflect Alexander's judgments, not independent verification [e1].
- This assessment concerns what the comparison establishes; it does not establish that Alexander claimed full compositional mastery.
- The comparison does not establish actual regressions, Imagen's overall compositional ability, or DALL-E 2's ranking at launch [e1].
Evidence (1 item)
Supports the claim: Astral Codex Ten,
We generate 10 images for each prompt, just like DALL-E2 does.
Why it matters here: The page specifies five prompts, reports DALL-E 2 at 0/5 and Imagen at 3/5 under Alexander's scoring, and contains no systematic retest of earlier successes. Alexander also expressly declines to claim full mastery, so the methodological limitation does not establish that he claimed otherwise.
- Publisher
- Astral Codex Ten
- Kind of source
- Demonstration
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Did GoogleAI Just Snooker One of Silicon Valley’s Sharpest Minds?
The progress that Alexander is vaunting shows that the system did better on a handful of items than DALL-E2 (which was not even state of the art at compositionality when it appeared), but not that the newer system (Imagen) systematically groks compositionality
- Where
- paragraph 50
- Published
- 2022-09-14
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-did-googleai-just-snooker-one-of-79d6ee-a2: Current. Outcome: Supported. 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-did-googleai-just-snooker-one-of-79d6ee-a2.