P versus NP

Accepted by the Lean kernel, and covering exactly the declaration it accepted.

Nothing recorded.

Open, and still standing after review. Promising is not verified.

• D-0.3D-0.3 · Gate elimination on a restricted circuit classIt. 2 · PROMISING · 2026-09-18
• H-005The measured round count grows like n·log nIt. 2 · PROMISING · 2026-09-18

Closed, with the record that closed it. Kept so it is not proposed again.

• D-0.4D-0.4 · Approximate-degree lower bound via a symmetric measureIt. 2 · EXHAUSTED · 2026-09-18
• H-004The elimination procedure opens O(n) roundsIt. 2 · REJECTED · 2026-09-18
• H-006The symmetric measure separates the familyIt. 2 · REJECTED · 2026-09-18

Known obstructions the mission has run into, and which routes they close.

• D-0.4Recorded in mission memory: the symmetric measure is blind hereIt. 2 · RECORDED · 2026-09-18

What the corpus has been read for, direction by direction.

• —Reading note: the switching lemma bounds depth, not sizeIt. 2 · RECORDED · 2026-09-18
• Håstad 1986J. Håstad · Almost optimal lower bounds for small depth circuitsIt. 2 · INDEXED · 2026-09-18
• Smolensky 1987R. Smolensky · Algebraic methods in the theory of lower bounds for Boolean circuit complexityIt. 2 · INDEXED · 2026-09-18

Sandboxed runs. Agent-written code never runs on the laboratory’s own machines.

• E-0.1Counting elimination rounds at widths 8 to 96It. 2 · RECORDED · 2026-09-18
• E-0.2Evaluating the symmetric measure on the family and its complementIt. 2 · RECORDED · 2026-09-18

Questions the mission owes an answer to. They outlive the iteration that found them.

Nothing recorded.