“Performance hovered at around 75% in the best case, when the robots had a limited number of options to consider.”
Our reading, the sentence that was checked:
PaLM-SayCan's performance was around 75% at best, when the robot had a limited number of options.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, What’s wrong with Google’s new robot project, paragraph 45.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-whats-wrong-with-googles-new-robot-56bf7a-a1
We recorded 15 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 2022-08-18. 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
- best case with limited options (qualifier, stated by the speaker)
From the speaker
- Performance hovered at around 75%
- in the best case
- a limited number of options to consider
Our interpretation
- Performance means overall end-to-end execution success in the evaluation discussed, rather than planning accuracy.
- Best case means the better overall evaluated setting, not the highest-scoring individual task or task family.
- Around 75% allows ordinary rounding of a reported result close to three quarters.
- The target is the evaluation released in August 2022, not subsequent versions.
Counts as supported if: The best overall evaluated setting reports approximately 75% execution success using a finite, predefined skill repertoire.
Counts as contradicted if: The comparable best overall execution result is materially different from approximately 75%, or the evaluation did not restrict execution to an available skill repertoire.
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
The best overall execution result was 74%, which reasonably rounds to about 75%, and execution used a predefined skill repertoire. This supports the ordinary contextual reading. It does not establish a 75% ceiling for every individual task.
Limitations
- Best case is interpreted as the best overall setting.
- These are developer-reported benchmark results, not independent household reliability measurements.
- Event dates identify the August evaluation release; actual trial dates are unspecified.
Evidence (2 items)
Supports the claim: arXiv; Michael Ahn and colleagues,
101 tasks.
Why it matters here: The reported overall execution rates are 74% in the mock kitchen and 60% in the real kitchen. Thus the better overall setting is approximately 75%; planning is a separate measure.
- Publisher
- arXiv; Michael Ahn and colleagues
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Supports the claim: Robotics at Google and Everyday Robots,
Once the skill is selected
Why it matters here: The system selects among available robot skills. This supports the limited-options qualification rather than unrestricted action generation.
- Publisher
- Robotics at Google and Everyday Robots
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, What’s wrong with Google’s new robot project
Performance hovered at around 75% in the best case, when the robots had a limited number of options to consider.
- Where
- paragraph 45
- Published
- 2022-08-18
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-whats-wrong-with-googles-new-robot-56bf7a-a1: 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-whats-wrong-with-googles-new-robot-56bf7a-a1.