The Situational Awareness timeline, stage by stage
The ladder, with live verdicts
| When | Predicted stage | Verdict, mid-2026 |
|---|---|---|
| 2025/26 | Models outpace college graduates on knowledge work | On track (~83% GDPval, ~80% SWE-Bench Pro) |
| 2027 | AGI — AI does the work of an AI researcher | Open — autonomy undemonstrated; resolves Jan 2028 |
| 2027/28 | "The Project" — US government takes over AGI development | Open — involvement growing, no formal project |
| 2027–29 | Intelligence explosion — a decade of progress in ~a year | Pending — trigger (AGI) hasn't fired |
| 2030s | Superintelligence; decisive strategic advantage | Pending — two rungs away |
Why the sequence is strict
Each rung depends on the one below. No college-grad capability → no AGI. No AGI → nothing to explode. No explosion → no superintelligence in the 2030s. That structure is what makes the essay checkable: the whole tower resolves or falls with the 2027 AGI claim, which is why it is the scorecard's headline verdict. The parallel input predictions — compute ~0.5 OOM/yr (on track), trillion-dollar capex (exceeded), open source fading (wrong) — feed the ladder but don't reorder it.
Where we are on the clock
Mid-2026 sits at the boundary between stage one and stage two: the capability stage has largely arrived on schedule, and the AGI stage has about 18 months left on its clock. The next 18 months therefore carry all the information — either autonomous AI research appears and the ladder holds, or January 2028 passes and the timeline's spine breaks. Follow it live on the scorecard homepage.
Frequently asked questions
A five-stage sequence: college-grad-level AI by 2025/26, AGI in 2027, a US government Project by 2027/28, an intelligence explosion 2027–29, and superintelligence in the 2030s.
The first is on track — models perform near skilled-human level on knowledge work (~83% GDPval) and agentic coding (~80% SWE-Bench Pro). AGI (2027) is open, and every later stage is pending.
The whole ladder loses its spine: the intelligence explosion and 2030s superintelligence stages depend on AGI arriving first. The claim resolves by January 1, 2028.
Roughly on schedule through stage one, with the decisive stage unresolved: input trends (compute, capex, capability) have held or exceeded, but the AGI stage's defining evidence — autonomous AI research — hasn't appeared.
The live scorecard updates as models ship and verdicts change.
View the live scorecard →Get the weekly AGI progress briefing
Verdict changes, lab milestones, and what they mean for the 2027 clock. Free — no hype, just signal.
Subscribe free →