“Machines will just get more and more reliable and more general.”
Our reading, the sentence that was checked:
AI systems will gradually become more reliable and more general.
Assessment: Unresolved No deadline stated on 2026-10-29 Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, AGI will not happen in your lifetime. Or will it?, paragraph 65.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-agi-will-not-happen-in-your-lifetime-143ce1-a1
We recorded 28 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 2023-01-22. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Forecast
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No deadline stated
- Deadline state on 2026-10-29
- No deadline stated (the release’s reference date)
From the speaker
- Machines
- more and more reliable
- more general
Our interpretation
- The plain reading is a continuing trend toward more dependable performance across a broader range of tasks.
- The surrounding discussion describes uneven, gradual progress. It does not promise improvement in every model or every release.
- No deadline, minimum improvement, or evaluation measure is supplied.
Counts as supported if: Comparable evaluations over time show sustained improvements in both dependable performance and breadth of competence. Such evidence supports the direction of the forecast without conclusively settling its indefinite continuation.
Counts as contradicted if: Broad, sustained evidence establishes a lasting reversal or limit in reliability or generality. A temporary setback or weakness in one model family is insufficient.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Did the forecast come true?
Outcome: Unresolved
The record was checked, and the evidence available by the cutoff does not settle it.
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 deadline stated
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
The evidence is consistent with broader capabilities and improving performance on some tasks. It also documents reliability problems that capability gains do not automatically solve. Neither finding conclusively settles this undated forecast of continuing improvement.
Limitations
- Benchmark performance is an incomplete measure of generality.
- The reliability study and software-task study use different evaluations and model populations.
Evidence (3 items)
Context only: Stanford Institute for Human-Centered Artificial Intelligence,
AI models improved rapidly
Why it matters here: The report documents gains across several domains, consistent with broader competence. Benchmark gains alone do not establish dependable general performance or indefinite future improvement.
- Publisher
- Stanford Institute for Human-Centered Artificial Intelligence
- Kind of source
- Secondary summary
- Describes events of
- Public from
Context only: METR researchers via arXiv,
doubling
Why it matters here: This study reports improving task-completion capability in a bounded software evaluation. Its narrower task distribution cannot establish generality across arbitrary domains; the availability date identifies the version read.
- Publisher
- METR researchers via arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: Nature,
may have become less reliable.
Why it matters here: The study challenges the assumption that increasing model capability automatically improves reliability. Its findings concern particular models and reliability tests, so they do not disprove Marcus's broader long-run forecast.
- Publisher
- Nature
- Kind of source
- Peer-reviewed paper
- Describes events of
- Public from
What would settle it
Further comparable evaluations could strengthen or weaken the forecast. Conclusive grading would require a specified evaluation period and measures, or a decisive event meeting the early-resolution rule.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, AGI will not happen in your lifetime. Or will it?
Machines will just get more and more reliable and more general.
- Where
- paragraph 65
- Published
- 2023-01-22
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-agi-will-not-happen-in-your-lifetime-143ce1-a1: Current. Outcome: Unresolved. 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-agi-will-not-happen-in-your-lifetime-143ce1-a1.