“LLMs are not AGI, and (on their own) never will be; scaling alone was never going to be enough.”
Our reading:
LLMs are not AGI and will never become AGI on their own.
Assessment: Unresolved
What happened
Did the forecast come true?
Outcome: Unresolved
We checked, and the evidence so far doesn’t settle it.
Judged on a plain reading. They named no measure, so we used the reading a typical informed reader would take. It is spelled out under “How we judged it”.
Evidence up to . Checked .
Why
On September 3, 2026, ARC Prize reported progress toward generalization but explicitly declined to call Astra AGI [e1]. In its version clarified September 14, 2026, the AGI Society found the published evidence insufficient to establish that Astra had achieved AGI [e2]. Neither assessment demonstrates LLM-only AGI or establishes that it is impossible, so these findings leave the forecast unresolved [e1] [e2].
Caveats
- Insufficient evidence of AGI does not establish its absence.
- The packet contains neither an independently established LLM-only AGI demonstration nor a validated general impossibility result.
- The evidence does not fully separate model capabilities from supporting machinery.
- Present limitations cannot settle a forecast that says 'never' and has no deadline.
Evidence (2 items)
Context only: ARC Prize,
progress towards generalization
Why it matters: The evaluator reports substantial progress but explicitly declines to establish AGI. This does not meet the early contradiction rule or establish permanent impossibility.
- Publisher
- ARC Prize
- Kind of source
- Independent measurement
- About events on
- Published
Context only: AGI Society,
published evidence
Why it matters: This review finds the published evidence insufficient to establish a prominent recent AGI claim. Its judgment concerns available evidence, not whether future LLMs could reach AGI.
- Publisher
- AGI Society
- Kind of source
- Secondary summary
- About events on
- Published
How we judged it
- Kind of claim
- Forecast
- Deadline
- None given. We don’t add one.
Conditions
- on their own (their words)
Where we had to interpret
- AGI means broad human-level cognitive competence, including learning unfamiliar tasks, rather than success on one benchmark.
- On their own restricts the system being evaluated: an LLM cannot rely on separate tools or cognitive machinery to supply the capabilities needed to qualify.
- This is a scope restriction, not a conditional forecast awaiting an antecedent.
- The present-tense clause and the indefinite future claim are assessed together. Evidence of present limitations cannot establish never.
Right if: A validated general limitation establishes that LLMs within the stated scope cannot attain broad human-level cognitive competence.
Wrong if: Independent evaluations establish broad human-level cognitive competence in an LLM alone, with sufficient technical disclosure to establish that the qualifying capabilities do not depend on excluded machinery.
We wrote this test after the outcome was known.
What would settle it
Independent evaluations establishing broad human-level cognitive competence, transfer and learning in an LLM alone, with human comparisons and technical disclosures showing no dependence on excluded machinery; or a validated general impossibility result covering LLMs within that scope.
We found 6 claims in this post. See them all.
Full record
Source and dates
Marcus on AI, The Great AI Retrenchment has begun, paragraph 4.
- Said
- We read the source
- live page, full text, public
- Evidence up to
- Checked
We read this wording on 2026-10-28. The post may have been edited after 2024-06-15.
Where it was said (1)
Said once in the posts we read.
: Marcus on AI, The Great AI Retrenchment has begun
LLMs are not AGI, and (on their own) never will be; scaling alone was never going to be enough.
- Where
- paragraph 4
- Published
- 2024-06-15
- We read it
- , live page
- Access
- full text, public
History
Versions
- Current: Unresolved. Evidence up to 2026-10-28.
What happened to this record
PublishedBy us
Published after both AI reviews and the final check passed.
Held backBy us
The final check failed: a cited source didn’t say what the write-up said.