Sam Altman's AGI prediction: what he's actually said
The public record
Altman's stated position has been consistent in direction: AGI is coming soon, and OpenAI knows the path. In The Intelligence Age (September 2024) he wrote that superintelligence might be “a few thousand days” away. In early 2025 he wrote that OpenAI was “confident we know how to build AGI as we have traditionally understood it,” and predicted AI agents would begin materially joining the workforce. He has also repeatedly softened the significance of the milestone itself — suggesting AGI will arrive and matter less on day one than people expect, with the world continuing to look surprisingly normal at first.
Altman’s timeline is a claim; the scorecard is what actually resolved.
Every quote above ages the moment a graded verdict moves — the 62.5/100 Thesis Tracker recomputes from eight predictions, each with a pre-registered flip condition. Subscribe and you hear on the day one flips, not when a headline does.
Tell me when a verdict flips →How it compares
| Forecaster | Position | Falsifiable? |
|---|---|---|
| Elon Musk | By end of 2026 | Yes — dated |
| Leopold Aschenbrenner | 2027 "strikingly plausible" | Yes — resolves Jan 2028 |
| Sam Altman | Soon; ASI "a few thousand days" | Partly — no hard date |
| Demis Hassabis | ~50% by 2030 | Yes — probabilistic |
| Andrej Karpathy | About a decade | Yes — loosely dated |
“A few thousand days” spans roughly the late 2020s through early 2030s — compatible with Aschenbrenner's aggressive path and with Hassabis's more cautious one. That width is why this scorecard tracks Aschenbrenner's dated claim rather than Altman's: a prediction you can't miss isn't a prediction you can grade.
Altman won't name a year. You can — where do you land versus “a few thousand days”?
One tap — your pick lands on the real forecaster spectrum next to Altman, Musk, Hassabis and Metaculus.
The incentive question
Altman runs the company whose valuation most benefits from near-term AGI expectations, which cuts both ways: he has better visibility than almost anyone, and stronger incentives than almost anyone. The scorecard's approach is to note the position, weight the falsifiable versions of it, and grade against public evidence — where the story is the same as everywhere else: capability strong, autonomy undemonstrated.
The dated claims that can be graded are tracked on one auditable number: the AGI-2027 Thesis Tracker, currently 62.5/100 across eight pre-registered predictions. Altman's own position sits alongside every other public forecast on the prediction receipts page, each with its deadline running.
A timeline with no deadline still has a price tag.
“A few thousand days” can’t cleanly miss — but the compute buildout that timeline rides on is very falsifiable, and it is graded on this site. Score the names doing the spending against the eight verdicts.
Instant result, no sign-up. Not investment advice.
Frequently asked questions
Altman has said OpenAI is “confident we know how to build AGI as we have traditionally understood it,” and in The Intelligence Age (2024) suggested superintelligence may be “a few thousand days” away. He deliberately avoids hard dates.
No. He has predicted AI agents joining the workforce and described AGI as near, while also arguing the milestone will feel less dramatic than expected when it arrives.
They're directionally aligned — both very near-term relative to expert medians. The difference is falsifiability: Aschenbrenner's 2027 resolves by January 2028, while “a few thousand days” spans roughly the late 2020s to early 2030s and can't cleanly miss.
Because it has no deadline. This scorecard grades pre-registered, dated claims; Altman's public position is tracked here as context for the graded forecasts.
The live scorecard updates as models ship and verdicts change.
View the live scorecard →Be first to know when a verdict flips
One email when one of the eight verdicts changes and the AGI-2027 Thesis Tracker moves — the single auditable score no other tracker has. Not a weekly newsletter; nothing arrives until the score actually moves.
Subscribe free →