“GenAI certainly sounds amazing – but it hasn’t always worked well enough to justify the costs.”
Our reading, the sentence that was checked:
Generative AI has not always worked well enough to justify its costs.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, OpenAI’s future, foretold?, paragraph 10.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-openais-future-foretold-d8e08f-a1
We recorded 2 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 2025-09-07. 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
- not always (qualifier, stated by the speaker)
From the speaker
- GenAI
- hasn’t always worked well enough
- to justify the costs
- the recent MIT study
Our interpretation
- The ordinary reading is existential: some real uses or deployments had delivered insufficient benefits to justify their costs.
- The claim does not say all deployments failed, or specify a failure percentage.
- Use January–June 2025, the research period of the recent study invoked in the surrounding passage.
- Credible deployment evaluations reporting investment without returns can support this modest claim. Failure to measure a benefit is weaker than proof that no benefit exists.
Counts as supported if: Credible evidence documents some GenAI deployments with incurred costs and reported benefits insufficient to justify those costs during the target period.
Counts as contradicted if: Evidence establishes that the purported shortfall cases actually justified their costs and that the claimed exceptions lack a factual basis. Successful deployments elsewhere alone cannot contradict 'not always'.
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 preliminary NANDA research reports organizations investing in GenAI without obtaining returns during its observation window. That supports Marcus's limited statement that results had not always justified costs. The report also describes successful uses; those are compatible with 'not always'. Its findings do not justify treating the headline failure percentage as a reliable estimate for all businesses.
Limitations
- Support rests on preliminary, largely self-reported and anonymized deployment research.
- Unmeasured benefits and later returns remain possible.
- METR provides productivity context, not a financial return calculation.
Evidence (2 items)
Supports the claim: Project NANDA report authors; copy hosted by MLQ,
zero return
Why it matters here: The January–June research reports invested resources without returns in some organizations, supporting the modest existence claim. August 20 is a conservatively confirmed public-availability date; this assessment does not endorse the report's population-wide percentage.
- Publisher
- Project NANDA report authors; copy hosted by MLQ
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: METR,
19% longer
Why it matters here: The primary paper says most tasks occurred in February and March 2025. The observed time penalty corroborates practical shortcomings, but it does not directly measure enterprise financial returns.
- Where
- Opening summary
- Publisher
- METR
- Kind of source
- Independent measurement
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, OpenAI’s future, foretold?
GenAI certainly sounds amazing – but it hasn’t always worked well enough to justify the costs.
- Where
- paragraph 10
- Published
- 2025-09-07
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-openais-future-foretold-d8e08f-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.