$ ls ./blog --sort=date

// notes & write-ups

-rw-r--r--2026-07-07finding-a-73-move-solution.md

Finding a 73-Move Solution When Every Correct Move Looks Wrong

How I built an AI solver for Google's Tic-Tac-Go puzzle — from breadth-first search, to reinforcement learning, to a hybrid beam search + CNN approach that hit a 93% solve rate.

  • AI
  • Search Algorithms
-rw-r--r--2026-07-06hello-world.md

Hello World

The first post on this blog

  • meta