benchmarksopinionbearishMuch of what appears to be reasoning in models may be priors leaking back when a model recognizes familiar patterns like a mazeMachine Learning Street Talk27 Jul 2026https://podcasters.spotify.com/pod/show/machinelearningstreettalk/episodes/The-Benchmark-With-No-Instructions--ARC-AGI-3-winning-team-e3lh6i9