← Live scorecard
Explainer

AI “orders of magnitude” (OOMs), explained

Last updated: August 16, 2026 · Updated as verdicts change
By the AGI Scorecard team · methodology & independence
The counting method behind the 2027 forecast. An order of magnitude (OOM) is a factor of 10. Aschenbrenner’s whole AGI case is built on counting OOMs of effective compute — raw compute + algorithmic efficiency + “unhobbling” — and betting they compound at ~0.5 OOM/yr toward AGI by 2027. Two years on, the pace has roughly held.
The part a chat answer goes stale on
62.5/100 how much of the AGI-2027 thesis is still standing — recomputed from eight graded predictions, as of 2026-08-08
See all eight verdicts, each with its pre-registered flip condition →

The OOM forecast is now a graded bet, not an argument.

Aschenbrenner’s compute arithmetic resolves prediction by prediction — the 62.5/100 above moves only when one of the eight graded verdicts actually flips. Subscribe and you hear the day the score moves.

Alert me when the score moves →

What an OOM is

One order of magnitude = 10×. Two OOMs = 100×. Aschenbrenner reasons in OOMs because AI progress is exponential: it's easier to track "how many 10×s per year" than absolute numbers. His forecast is essentially an addition problem — stack enough OOMs of effective compute and you cross the AGI threshold.

The three sources of OOMs

SourceWhat it adds
Raw computeBigger training runs
Algorithmic efficiencyMore capability per FLOP
UnhobblingUnlocking latent capability (reasoning, tools, agents)

The bet, and how it's tracking

He projected roughly 0.5 OOM/yr of effective compute, sustained. As of mid-2026 an independent audit calls the pace "roughly supported," with launches scattered within about ±0.5 OOM of the trend — graded On track. This OOM engine is what sits under every downstream claim: capability, AGI timing, and the intelligence explosion. If the OOMs stop stacking, the whole 2027 case slips.

Frequently asked questions

What are orders of magnitude in AI?

An order of magnitude (OOM) is a factor of 10. Aschenbrenner counts OOMs of 'effective compute' — raw compute plus algorithmic efficiency plus unhobbling — because AI progress is exponential and easier to track as '10×s per year.'

How many OOMs per year did Aschenbrenner predict?

Roughly 0.5 orders of magnitude of effective compute per year, sustained over the decade. As of mid-2026 the pace has roughly held, graded On track.

Why do OOMs matter for AGI?

Because his AGI forecast is essentially an addition problem: stack enough OOMs of effective compute and you cross the AGI threshold. If the OOMs stop stacking, the 2027 timeline slips.

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 →