Building AI agents for 2048, learning stuff, sharing the journey.

We’re exploring everything from traditional heuristics to cutting-edge neural networks, implementing research-backed strategies that actually work. This is where we document the discoveries, share the code, and maybe occasionally complain about getting stuck on a particularly tricky algorithm.

Whether you’re interested in game AI, expectimax search, or just want to see some TypeScript that makes squares slide around intelligently, you’re in the right place.

Check out the live demo to see our agents in action, or dive into the posts below for the technical details.