“Generative AI is hitting a wall, and any future improvements will be marginal, and cost billions to achieve.”
Our reading, the sentence that was checked:
Generative AI is hitting a wall; any future improvements will be marginal and cost billions.
Assessment: Contradicted Settled before its deadline on 2026-10-29 Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, The Shareholder Supremacy, paragraph 121.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-tss-ba8f03-a1
We recorded 22 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 2024-07-01. 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
- Settled before its deadline
- Deadline state on 2026-10-29
- Settled before its deadline (the release’s reference date)
From the speaker
- Generative AI is hitting a wall
- any future improvements will be marginal
- cost billions to achieve
Our interpretation
- The ordinary reading is that improvements after July 1, 2024 will all be small, while requiring expenditure measured in billions.
- A substantial, independently measured advance on demanding tasks counts against marginality. This does not require AGI, universal reliability, or profitable deployment.
- Generative AI includes systems using language models with reasoning, tools, or additional inference computation; the sentence does not restrict its forecast to unchanged training methods.
- The cost assertion specifies neither a company nor a cost boundary. Training-run cost alone cannot settle comprehensive development cost.
- The forecast is conjunctive: a clear counterexample to universal marginality can contradict it without resolving the cost assertion.
Counts as supported if: Evidence would need to establish that all subsequent improvements remain small and require billions in comparable comprehensive costs. A finite observation period cannot conclusively establish this unlimited forecast.
Counts as contradicted if: A dated post-publication advance demonstrates a clearly substantial capability gain under comparable evaluation, independently corroborated; alternatively, a meaningful improvement has documented comprehensive costs below one billion dollars.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Did the forecast come true?
Outcome: Contradicted
The dated evidence goes against 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
- Settled before its deadline
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
The universal marginal-improvement forecast has a substantial counterexample. ARC Prize independently measured a large advance in o3-preview, and METR's subsequent task-length research corroborates meaningful capability progress. These findings contradict the ordinary reading that any future improvement would be marginal. They do not establish AGI, universal reliability, or cheap development. The separate billions-of-dollars assertion remains unverified.
Limitations
- The conclusion uses an ordinary capability-based reading of marginal.
- ARC performance depended on benchmark-specific training and substantial inference computation.
- METR's results concern selected tasks and a 50% success threshold.
- The cost component is unresolved and is not counted as contradicted.
Evidence (3 items)
Goes against the claim: ARC Prize,
87.5%
Why it matters here: ARC Prize reports 75.7% and 87.5% for the o3 preview configurations and contrasts these with GPT-4o's 5% performance in 2024. The evaluator describes a substantial advance, while disclosing benchmark-specific training and increased inference computation.
- Publisher
- ARC Prize
- Kind of source
- Independent measurement
- Describes events of
- Public from
Goes against the claim: METR researchers / arXiv,
doubling
Why it matters here: Independent comparisons extending through early 2025 find substantial growth in the human-duration tasks agents can complete at 50% success. This corroborates progress beyond a single ARC evaluation, although the task sample and success threshold limit generalization.
- Where
- Abstract
- Publisher
- METR researchers / arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: ARC Prize,
fails
Why it matters here: Material qualification: the same evaluator identifies continuing basic failures. Substantial measured improvement does not establish AGI or eliminate reliability limitations.
- Where
- So is it AGI?
- Publisher
- ARC Prize
- Kind of source
- Independent measurement
- Describes events of
- Public from
What would settle it
Comparable comprehensive development-cost disclosures would settle the remaining cost component.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, The Shareholder Supremacy
Generative AI is hitting a wall, and any future improvements will be marginal, and cost billions to achieve.
- Where
- paragraph 121
- Published
- 2024-07-01
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-tss-ba8f03-a1: Current. Outcome: Contradicted. 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.