“I just don’t happen to think large language models like ChatGPT will get us there.”
Our reading, the sentence that was checked:
Large language models like ChatGPT will not get us to AGI.
Assessment: Unresolved No deadline stated on 2026-10-29 Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Three years on, ChatGPT still isn't what it was cracked up to be – and it probably never will be, paragraph 5.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-three-years-on-chatgpt-still-isnt-f0fb14-a1
We recorded 30 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 2025-11-29. 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
- large language models like ChatGPT
- will get us there
- artificial general intelligence (AGI)
Our interpretation
- The plain reading is that LLM-based systems will not achieve the breadth and resourcefulness of human intelligence.
- Ordinary tools and additional training do not automatically put a system outside the claim. A fundamentally different source of its general intelligence could.
- Human-level general intelligence does not require infallibility. Independent evidence of broad, adaptable competence is required; a product announcement or one benchmark is insufficient.
Counts as supported if: A validated result establishing that the relevant class of LLM-based systems cannot achieve human-level general intelligence. Continued failure over a finite period cannot establish this indefinite negative.
Counts as contradicted if: An LLM-based system demonstrably achieves human-level breadth and resourcefulness, with independent confirmation and sufficient architectural information to establish that it falls within the claim.
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 includes substantial progress, including human-surpassing action efficiency on one benchmark. That benchmark's authors expressly decline to equate its results with AGI. Neither these gains nor remaining limitations settle whether LLM-based systems can eventually achieve human-level general intelligence. Marcus supplied no deadline.
Limitations
- AGI and the architectural boundary are not precisely defined.
- Benchmark success is narrower than general intelligence; present shortcomings do not establish permanent inability.
Evidence (2 items)
Context only: ARC Prize Foundation,
human baseline
Why it matters here: This is material evidence of progress against human performance. The evaluator nevertheless says the benchmark's bounded environments do not establish AGI, so it does not meet the early-resolution rule.
- Publisher
- ARC Prize Foundation
- Kind of source
- Benchmark result
- Describes events of
- Public from
Context only: OpenAI,
professional work
Why it matters here: OpenAI reports advances across several domains. This is relevant counterevidence to pessimism about the approach, but the announcement does not independently establish human-level general intelligence.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
What would settle it
Independent, broad validation of AGI in an in-scope LLM-based system, or a validated impossibility result that actually covers the claimed systems and capability.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Three years on, ChatGPT still isn't what it was cracked up to be – and it probably never will be
I just don’t happen to think large language models like ChatGPT will get us there.
- Where
- paragraph 5
- Published
- 2025-11-29
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-three-years-on-chatgpt-still-isnt-f0fb14-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-three-years-on-chatgpt-still-isnt-f0fb14-a1.