Sam Earle
@smearle.bsky.social
PhD at NYU in AI and Games. RL, environment generation, open-ended learning.
Now, we can begin see how AI players respond to these challenges. The picture may be starkly surprising to some: simple tree search finds most solutions rapidly, while RL falls prey to obvious local minima, and LLMs spin their wheels when faced with unfamiliar semantics.
August 27, 2025 at 11:34 PM
Now, we can begin see how AI players respond to these challenges. The picture may be starkly surprising to some: simple tree search finds most solutions rapidly, while RL falls prey to obvious local minima, and LLMs spin their wheels when faced with unfamiliar semantics.
PuzzleScript games make for a great benchmark. Often, despite their mechanical simplicity, they elicit moments of insight in human players. Since 2013, casual and professional designers have brought considerable ingenuity to the language, generating a plethora of diverse games.
August 27, 2025 at 11:34 PM
PuzzleScript games make for a great benchmark. Often, despite their mechanical simplicity, they elicit moments of insight in human players. Since 2013, casual and professional designers have brought considerable ingenuity to the language, generating a plethora of diverse games.
Paper: arxiv.org/abs/2508.16821
Code: github.com/smearle/scri...
PuzzleJAX is a faithful re-implementation of PuzzleScript (puzzlescript.net) capturing all of the engine's major features. It leverages the convolutional nature of rewrite rules to achieve major speedups in JAX.
Code: github.com/smearle/scri...
PuzzleJAX is a faithful re-implementation of PuzzleScript (puzzlescript.net) capturing all of the engine's major features. It leverages the convolutional nature of rewrite rules to achieve major speedups in JAX.
August 27, 2025 at 11:34 PM
Paper: arxiv.org/abs/2508.16821
Code: github.com/smearle/scri...
PuzzleJAX is a faithful re-implementation of PuzzleScript (puzzlescript.net) capturing all of the engine's major features. It leverages the convolutional nature of rewrite rules to achieve major speedups in JAX.
Code: github.com/smearle/scri...
PuzzleJAX is a faithful re-implementation of PuzzleScript (puzzlescript.net) capturing all of the engine's major features. It leverages the convolutional nature of rewrite rules to achieve major speedups in JAX.
We introduce PuzzleJAX, a benchmark for reasoning and learning. 🧩💡🦎
PuzzleJAX compiles hundreds of existing grid-based PuzzleScript games to hardware-accelerated JAX environments, and allows researchers to define new tasks via PuzzleScript's concise rewrite rule-based DSL.
PuzzleJAX compiles hundreds of existing grid-based PuzzleScript games to hardware-accelerated JAX environments, and allows researchers to define new tasks via PuzzleScript's concise rewrite rule-based DSL.
August 27, 2025 at 11:34 PM
We introduce PuzzleJAX, a benchmark for reasoning and learning. 🧩💡🦎
PuzzleJAX compiles hundreds of existing grid-based PuzzleScript games to hardware-accelerated JAX environments, and allows researchers to define new tasks via PuzzleScript's concise rewrite rule-based DSL.
PuzzleJAX compiles hundreds of existing grid-based PuzzleScript games to hardware-accelerated JAX environments, and allows researchers to define new tasks via PuzzleScript's concise rewrite rule-based DSL.