ANALYSIS
Dittert-5 emerged as the main signal: a candidate mathematical result produced through a highly autonomous pipeline from problem selection through discovery, certificate generation, Lean formalization, verification and manuscript preparation. Human review was still missing, so it did not justify a trajectory change. Inverse Generator and Crouzeix were weaker autonomy signals. No robotics or frontier-intelligence event justified a score change. ASI probability remained 2% / 7% / 19% / 44% / 82% across the 1 / 2 / 3 / 5 / 10-year horizons.