“I expect for example that physical reasoning will improve (a radical boon for domestic robots) before psychological reasoning; there won’t be a magic day in which we say “aha AGI landed on January 27, 2062” or whatever it is.”
Our reading:
AI physical reasoning will improve before psychological reasoning, and AGI will not arrive on a single identifiable day.
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
The sources document progress relevant to both domains, but do not establish which improved first after the post. RT-2's gains and later theory-of-mind robustness gains cannot settle that order from publication dates alone. Contemporary AGI declarations also do not establish the abrupt transition needed to contradict the second clause. Marcus supplied no deadline, and the illustrative 2062 date is not one. No decisive event found in the reviewed record settles the compound forecast.
Caveats
- Robot-task success is an imperfect measure of physical reasoning.
- The theory-of-mind robustness study is a preprint.
- The AGI Society source is an assessment of published evidence, not a comprehensive independent test.
- The search does not establish an exhaustive chronology of every relevant advance.
Evidence (3 items)
Context only: Google,
performance on novel, unseen scenarios to 62%
Why it matters: The announcement reports improved robot performance in unfamiliar scenarios, relevant to physical reasoning. It neither isolates reasoning from other improvements nor establishes that psychological reasoning had not improved earlier.
- Publisher
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- About events on
- Published
Context only: arXiv; Ian B. de Haan, Peter van der Putten, and Max van Duijn,
increased robustness to prompt variations and task perturbations
Why it matters: This reports improvement on theory-of-mind tasks and guards against assuming psychological reasoning remained static. It does not identify when the first such improvement occurred relative to physical reasoning.
- Publisher
- arXiv; Ian B. de Haan, Peter van der Putten, and Max van Duijn
- Kind of source
- Preprint, not peer reviewed
- About events on
- Published
Context only: AGI Society,
Our current evaluation
Why it matters: This assessment examines contemporary declarations of AGI and finds their published evidence insufficient. It does not establish either an abrupt arrival or a completed gradual arrival, so those declarations do not supply a conclusive event.
- 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.
Where we had to interpret
- The ordinary reading predicts meaningful advances in everyday physical reasoning before corresponding advances in reasoning about minds, relative to capabilities at publication.
- Improvement means demonstrated gains over earlier systems on comparable tasks. It does not mean first attaining reliable or human-level competence.
- Physical reasoning includes reasoning about objects, their interactions, and physical consequences. Psychological reasoning includes reasoning about beliefs, intentions, and other mental states.
- The domestic-robot parenthesis illustrates the usefulness of physical reasoning; it does not restrict the forecast to commercial household robots.
- The second clause predicts gradual development across dimensions of intelligence rather than a single abrupt arrival of general intelligence. A launch date or promotional declaration alone would not contradict it.
- January 27, 2062 is explicitly an illustrative date, not a deadline.
Right if: Dated, comparable evaluations establish that meaningful physical-reasoning improvements occurred before psychological-reasoning improvements after publication. Supporting the entire compound claim also requires a documented emergence of general intelligence through staggered capability advances rather than a single abrupt transition.
Wrong if: Comparable dated evidence establishes psychological-reasoning improvement before physical-reasoning improvement, or establishes simultaneous improvements at sufficient temporal resolution to reject the proposed ordering. Separately, a well-established abrupt transition to general intelligence would contradict the gradual-arrival clause.
We wrote this test after the outcome was known.
What would settle it
A dated comparison using consistent tests and publication-time baselines could settle the ordering of improvements. Evidence establishing how broadly general intelligence emerged would separately settle the gradual-arrival clause. Gains need not reach human-level competence to count.
We found 29 claims in this post. See them all.
Full record
Source and dates
Marcus on AI, AGI will not happen in your lifetime. Or will it?, paragraph 64.
- Said
- We read the source
- live page, full text, public
- Evidence up to
- Checked
We read this wording on 2026-10-29. The post may have been edited after 2023-01-22.
Where it was said (1)
Said once in the posts we read.
: Marcus on AI, AGI will not happen in your lifetime. Or will it?
I expect for example that physical reasoning will improve (a radical boon for domestic robots) before psychological reasoning; there won’t be a magic day in which we say “aha AGI landed on January 27, 2062” or whatever it is.
- Where
- paragraph 64
- Published
- 2023-01-22
- 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
A reviewer said our test wasn’t fair to what was actually said.