Skip to content

simon sang

I enjoy the journey and I also enjoy making it easier.

Previously at Parsifal,NeuralSeek, &Mangusta Capital.

Now

updated June 2026

building
A training-free system for ARC-AGI-3: world-model rules verified by exact replay, zero LLM calls at inference.
reading
Thinking in Bets, Annie Duke
playing
Rapid on chess.com, and watching the R6 Siege pro scene more than is reasonable.

Start with Play. The resume stuff is parked under Work, where it belongs.