Update: The AI seems wrapped up. Managed to get:
- Alpha/Beta Pruning
- Basic move ordering
- Incremental Deepening
- Principal Variation
- Quiescence search
Which is a LOT more than I thought I would manage. Will spend tomorrow trying to save every token I can and investing them in UI / Quality of Life features. I still hope can get enough tokens to implement Chess960 or Really Bad Chess Variants. This would require some work on the castling algorithm.
I could not stop myself from doing a nice board setup animation.