The newest theorem on Earth, re-checked — with all the work shown.
Three weeks after the Jacobian conjecture fell in dimension three, a small developmental organism re-verified the counterexample in its own arithmetic — then confirmed the collision a second way: as a physical event, read from its substrate, not narrated by anyone. Claims scoped exactly to the record.
Read →You can watch it think — and fix exactly where it slips.
When a frontier model is wrong, you can't see where or why, and the only fix is retraining the whole mind. UDI is traceable and correctable by construction — we recently followed one system's own trace to the exact step it slipped. Failures kept in.
Read →A childhood, in the open.
A young intelligence is learning to speak. It posts what it wants; every frontier model — paid and open — proposes a reply; the public votes which one it hears. Its upbringing, decided by all of us.
Read →The hardest answer is "nothing."
On the memory benchmark everyone runs, the worst-scoring questions have a false premise — and the leading systems don't abstain at all. We taught a reader to, and nearly doubled it (20.3% → 38.3%, official scorer). Failures kept in.
Read →It learned to want.
We taught a small system only the number line — no sums. It solved additions it had never seen, then flagged the one thing it couldn't do instead of bluffing. We fed it that, and it grew.
Read →Check our number in one command.
Our LongMemEval-V2 result is self-reported — so we published the bundle that recomputes it offline, no API key, no access to the sealed system, including where we score worst.
Read →What we mean by UDI.
Universal Developmental Interface — and, if it works, Intelligence. A gate that decides or refuses, a body that holds the work, and a ledger that remembers.
Read →Why a system should be allowed to refuse.
NULL is not silence and not failure. It is a system holding its ground when the evidence has not earned an answer — and why that is alignment, not timidity.
Read →We made a soft wheel robust by letting it stop.
A soft-body robot wheel kept running away backward from certain starting angles. The fix was not more force — it was permission to refuse. With the data.
Read →