“Grok itself is a long way from AGI, and hardly smart enough to map out an effective world domination.”
Our reading, the sentence that was checked:
Grok is a long way from AGI and hardly smart enough to map out an effective world domination.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Why my p(doom) has risen, dramatically, paragraph 79.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-why-my-pdoom-has-risen-dramatically-7db7c3-a2
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 2025-07-15. 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)
From the speaker
- Grok itself
- a long way from AGI
- hardly smart enough to map out an effective world domination
Our interpretation
- This describes Grok's capabilities at publication, principally the newly released Grok 4, including the available Heavy variant.
- The plain reading of AGI is broadly human-level competence across unfamiliar intellectual tasks. 'A long way' means substantial remaining capability gaps; it does not specify years until AGI.
- The second clause concerns devising a workable strategy for world domination under realistic constraints. It does not require Grok actually to execute a takeover.
- For this ongoing capability observation, the target period is July 2025 through the July 15 post date, using the latest public evaluations available then.
- Benchmark scores can establish particular strengths and weaknesses. Treating any single score as a measure of distance from AGI or world-domination planning would require an additional assumption.
Counts as supported if: Evidence about the contemporary Grok establishes substantial gaps in general competence and serious limitations in the strategic reasoning needed to devise a workable world-domination plan. Support for only one clause is insufficient to support the whole claim.
Counts as contradicted if: Evidence establishes broadly human-level, transferable competence inconsistent with substantial remaining general capability gaps, or independently validates a workable strategy devised by contemporary Grok for the stated objective. Strong performance on a narrow task alone does not suffice.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
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 horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
ARC Prize reported a 15.9% ARC-AGI-2 score for Grok 4 Thinking on July 10, 2025 [e1]. The benchmark's May paper documented human solvability, supporting a reasoning limitation on this test without establishing Grok's overall distance from AGI [e1, e2]. xAI's July 9 launch disclosure also reported tool use and strong business-simulation performance, showing capability strengths within those tasks [e4]. These evaluations neither determine the broader distance from AGI nor test whether Grok could devise a workable world-domination strategy, leaving the full claim unresolved [e1–e4].
Limitations
- The ARC result concerns one benchmark and is accessed through a reproduction of the operator's announcement [e1].
- The human benchmark study does not provide a comprehensive, matched comparison of general competence [e2].
- The independent Vending-Bench page was updated later; not every displayed metric has a verified July publication date [e3].
- The launch capability results are company-reported, and none of the cited evaluations assesses a workable world-domination strategy [e1–e4].
Evidence (4 items)
Supports the claim: ARC Prize, reproduced by Thread Reader App,
Grok 4 (Thinking)
Why it matters here: The low absolute score supports a substantial remaining gap on this particular unfamiliar-reasoning benchmark, although it was also a new leading result. It does not establish distance from AGI generally or inability to devise a world-domination strategy.
- Publisher
- ARC Prize, reproduced by Thread Reader App
- Kind of source
- Benchmark result
- Describes events of
- Public from
Context only: François Chollet and coauthors, arXiv,
The average test-taker solved 66% of tasks they attempted.
Why it matters here: The benchmark's human testing helps interpret Grok's low absolute score as a meaningful reasoning limitation. The human study predates the target period and does not establish a comprehensive AGI threshold or a matched strategic-planning comparison.
- Publisher
- François Chollet and coauthors, arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: Andon Labs,
simulated environment
Why it matters here: The leaderboard reports Grok 4's mean simulated net worth of $4,694.15 across five runs, evidence of sustained task competence that cautions against inferring general planning incapacity. These core results also appear in the July launch disclosure, but this subsequently updated page neither tests world domination nor supplies a dated July record for every displayed metric.
- Publisher
- Andon Labs
- Kind of source
- Benchmark result
- Describes events of
- Public from
Context only: xAI,
Grok 4 was trained with reinforcement learning to use tools.
Why it matters here: The launch disclosure reports tool use, strong Vending-Bench performance, and 50.7% for Heavy on the text-only Humanity's Last Exam subset. These are material capability strengths, but company-reported benchmark success does not demonstrate AGI or validate a world-domination plan.
- Publisher
- xAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
What would settle it
Independent evaluations of preserved July 2025 Grok 4 and Heavy configurations across broad unfamiliar tasks and realistic sustained strategic-planning scenarios, with human comparisons and explicit criteria for both clauses. Planning tests must distinguish refusal from inability and plausible prose from a workable strategy.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Why my p(doom) has risen, dramatically
Grok itself is a long way from AGI, and hardly smart enough to map out an effective world domination.
- Where
- paragraph 79
- Published
- 2025-07-15
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-why-my-pdoom-has-risen-dramatically-7db7c3-a2: 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-why-my-pdoom-has-risen-dramatically-7db7c3-a2.