Situational Awareness predictions, tracked
Every prediction and its current verdict
| Prediction | Evidence (mid-2026) | Verdict |
|---|---|---|
| Outpace college grads on knowledge work | ~83% GDPval; ~80% SWE-Bench Pro | On track |
| Compute scales ~0.5 OOM/yr | Pace roughly held (Epoch) | On track |
| Trillion-dollar capex | Exceeded projections | Exceeded |
| Open source fades; moat holds | DeepSeek/Qwen ~3–6 mo behind | Wrong |
| AGI by 2027 | Agentic strong; autonomy unproven | Open |
| US govt AGI project by 27/28 | No formal project yet | Open |
| Intelligence explosion 2027–29 | Too early | Pending |
| Superintelligence 2030s | Too early | Pending |
How the grading works
Each prediction is graded against the reading a careful 2024 reader would have taken — not the most defensible retreat available in hindsight. Verdicts are: On track (evidence consistent with the trajectory), Wrong (evidence contradicts the claim as stated), Open (deadline not reached, evidence mixed), and Pending (too early to bear on it). Every verdict has a pre-registered condition that would flip it, so changes aren't post-hoc.
The two-year headline
The input curves Aschenbrenner bet on — compute, capex, raw capability — have largely held or exceeded expectations. His biggest miss is on diffusion: open source did not fade. The defining claim, AGI by 2027, resolves by January 2028.
Every row in this ledger is also a position.
The exposure tool maps 17 listed names to exactly these eight verdicts, with the same weights the Thesis Tracker uses. Score the rows that held, the one that broke, or the still-open ones.
Instant result, no sign-up. Not investment advice.
Frequently asked questions
Aschenbrenner's 2024 essay predicted AGI by 2027, an intelligence explosion 2027–29, superintelligence in the 2030s, continued compute scaling, trillion-dollar capex, open source fading, and a US government AGI project.
As of mid-2026: 3 are on track, 1 is wrong (open source fading), 2 are open, and 2 are too early to grade.
Partly. Its compute, capex, and capability predictions have largely held; its open-source prediction is wrong; and its defining AGI-2027 claim is still open, resolving by January 2028.
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 →