178K Parameter Neural Net That Wins Poke(rogue)like
178K parameter model beats a Pokemon roguelike after the author rage-quit hundreds of times.

178K neural net beats Pokémon roguelike with clever 1386-dim state encoding.
ML engineers, RL enthusiasts, game developers
OpenAI Gym environments · Stable Baselines3 · RLlib
178K parameter model beats a Pokemon roguelike after the author rage-quit hundreds of times.
Two RL clones duel in-browser with epiplexity selection — no human tuning the explore-exploit knob.
One formula predicts qubit failure 20 days early, but cross-domain claims lack independent peer review.
Academic neural cryptography with error correction—interesting research, niche application.
Direct ANE access bypasses CoreML to enable training—genuinely novel Apple Silicon unlock.
65K lines of C++20 making blockchain produce useful AI models instead of wasting electricity.