“The Humane AI pin/virtual assistant just doesn’t work well, and a lot of the problem is the underlying AI, which I believe is powered by OpenAI.”
Our reading, the sentence that was checked:
The Humane AI pin just doesn't work well, and a lot of the problem is its underlying AI, which he believes is powered by OpenAI.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, TED, AI, and the test of time, paragraph 9.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-ted-ai-and-the-test-of-time-a9d5a7-a1
We recorded 11 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 2024-04-15. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Causal claim
- 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
- which I believe is powered by OpenAI (qualifier, stated by the speaker)
From the speaker
- The Humane AI pin/virtual assistant
- just doesn’t work well
- a lot of the problem is the underlying AI
- which I believe is powered by OpenAI
- Reviews last week
Our interpretation
- Assess the product available around its April 2024 launch, using April 1–15 as the observation window indicated by the present-tense statement and reference to the previous week's reviews.
- The ordinary reading is that recurring AI failures materially impaired practical use. 'A lot' does not specify a majority or a numerical share of failures.
- The reference to OpenAI is a qualified provider attribution, not a condition that must occur before the claim applies.
- Underlying AI includes answer generation and visual recognition. Hardware, connectivity and service integration failures must be considered separately.
Counts as supported if: Independent launch-period testing documents poor practical performance and recurring wrong or unreliable AI responses that materially contribute to it. Provider disclosures corroborate the qualified OpenAI attribution.
Counts as contradicted if: Representative launch-period testing finds dependable practical performance, or failure analyses show that the supposed AI problems instead arose predominantly from unrelated components without a material AI contribution.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Does the evidence support the cause?
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
Independent launch reviews found poor practical performance and recurring incorrect AI responses that undermined central functions. Humane also disclosed an OpenAI partnership. This supports the ordinary qualitative reading that AI unreliability was a substantial contributor. It does not establish that foundation models caused most failures or that OpenAI specifically caused each error.
Limitations
- The causal conclusion is qualitative and rests on documented product behavior, not a controlled component analysis.
- Successful tasks and substantial hardware, connectivity and integration problems limit the breadth of the conclusion.
- A stronger reading of 'a lot' as a measured majority would remain unresolved.
Evidence (4 items)
Supports the claim: WIRED,
Accuracy of answers is mixed
Why it matters here: The launch review documents a false answer about a California food ban and a mistaken temple identification. These failures directly impair the question-answering function, although the review also identifies missing features and hardware problems.
- Publisher
- WIRED
- Kind of source
- Independent measurement
- Describes events of
- Public from
Supports the claim: The Verge,
But the AI Pin doesn’t work.
Why it matters here: First-hand testing supports poor practical performance, including incorrect visual identification and unsuccessful translation. The review also reports server failures, overheating and missing functions, so its overall failure rate cannot be attributed entirely to models.
- Publisher
- The Verge
- Kind of source
- Independent measurement
- Describes events of
- Public from
Context only: Inverse,
when it hits
Why it matters here: The reviewer found useful answers and successful visual recognition alongside hallucinations. Humane also attributed some failures to server bugs and weak wireless connections, limiting any claim that foundation models alone caused most problems.
- Publisher
- Inverse
- Kind of source
- Independent measurement
- Describes events of
- Public from
Context only: Humane Inc., distributed by PR Newswire,
partnering
Why it matters here: Humane's announcement confirms an OpenAI partnership and access to its models. This supports the provider attribution as background, but establishes neither exclusive OpenAI use nor responsibility for particular April failures.
- Publisher
- Humane Inc., distributed by PR Newswire
- 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, TED, AI, and the test of time
The Humane AI pin/virtual assistant just doesn’t work well, and a lot of the problem is the underlying AI, which I believe is powered by OpenAI.
- Where
- paragraph 9
- Published
- 2024-04-15
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-ted-ai-and-the-test-of-time-a9d5a7-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-ted-ai-and-the-test-of-time-a9d5a7-a1.