Our mission
Measure what AI can do.
AI capability is rising measurably. Timelines are not reliable. The Foundation measures the first, refuses to fake the second, and publishes the method with every result.
Why a foundation
Frontier labs evaluate their own models on their own benchmarks and audit their own commitments. That is not bad faith; it is missing capacity. Independent measurement fits on one hand, and none of it works in Turkish.
Most of the leading researchers we rely on expect the most capable systems to be held inside companies and never shown publicly. If they are right, external measurement will be the only remaining instrument of scrutiny.
Our position
Capability is rising, and accelerating.
Timelines are unreliable.
We argue for crossing the threshold, with measurements in hand.
Seven pledges
- 1Open measurement
- 2Dated forecasts
- 3Room for the other side
- 4Independent funding
- 5Public record in Turkish
- 6Generation-time advocacy
- 7Error log
The reasoning, in full: the manifesto and the charter.