We accept the threshold

1. I

   Every unsolved problem is a thinking problem. Cheap intelligence touches all of them at once.

   The scarcity is intelligence, not matter
2. II

   There is no door. There is a curve, and it is steepening.

   The threshold is not a moment, it is a curve
3. III

   Serious forecasts span 2026 to 2047. We commit to a method, not a date.

   We commit to a method, not a date
4. IV

   Scaling is physically possible through 2030. Whether it happens is a decision.

   Acceleration is a choice, not a fate
5. V

   The risk of accelerating is always counted. The cost of waiting never is.

   Waiting has a price too
6. VI

   Every paradigm drops 2–3× from ARC-AGI-1 to ARC-AGI-2. Generalisation is unsolved.

   Generalisation is not solved yet
7. VII

   A system that cannot be controlled is not powerful. It is only dangerous.

   Safety is not a brake on speed, it is what carries it
8. VIII

   Same data, two models, opposite signs. We say so.

   We do not hide the uncertainty
9. IX

   17 of 25 researchers expect the best systems to stay behind closed doors. We measure anyway.

   We will not be spectators
10. X

    The human is the end, not the instrument. If that stops being true, so does this manifesto.

    The human is the end, not the instrument

If the threshold can be crossed, it will be. Our responsibility is not to stop it or to applaud it — but to cross it with measurements in hand.

Read the full manifesto, with evidence, objections and sources.