Minotaur AI and Earstrip Triangulator

Minotaur AI can reason like a language model in a quarter of a second on a single CPU core. It performs deduction, induction, abduction, diagnosis, and belief revision using just one network operation. Meanwhile, the new earstrip library processes 3.7 million map polygons in 489 milliseconds, beating earcut by up to 2.1×. In C, earstrip is 14% faster than earcut.hpp, and in browsers it runs 1.4 to 2.1 times faster. Minotaur’s reasoning was tested blind against a language model and matched its performance on police interviews and data‑driven diagnostics. These breakthroughs show that powerful AI can arise from simple, verifiable logic rather than massive neural networks. Minotaur’s architecture, called ELC, measures regulatory compliance by comparing real rules to recorded applications, avoiding GPU use. Earstrip’s ear‑clipping method, based on Meisters’ theorem, cuts 33‑to‑256 vertex polygons from 327 ms to 216 ms. It reduces total time to 6.0 times less than Delaunay on the same polygons.