Guanghan Ning(@quietnning):Jonathan Blow’s latest puzzle game Order of the Sinking Star is finally released! A decade ago, The Witness came out; as a tribute to this legendary puzzle game, we introduce our work today: Witness-Inspired Testbed of Novel Environments with Symbolic Structure (WITNESS). We show that interactive discovery of unseen rules in unfamiliar environments is a major difficulty for frontier models like Opus-5 and GPT-6, while smaller models further struggle with rule-based execution. Our experiments also indicate that RL on hidden-rule puzzles transfers to broader rules and real-world tasks beyond training. 🧵 Long thread [0/n].

2026.10.08 17:22

Jonathan Blow’s latest puzzle game Order of the Sinking Star is finally released! A decade ago, The Witness came out; as a tribute to this legendary puzzle game, we introduce our work today: Witness-Inspired Testbed of Novel Environments with Symbolic Structure (WITNESS). We show that interactive discovery of unseen rules in unfamiliar environments is a major difficulty for frontier models like Opus-5 and GPT-6, while smaller models further struggle with rule-based execution. Our experiments also indicate that RL on hidden-rule puzzles transfers to broader rules and real-world tasks beyond training. 🧵 Long thread [0/n].