Announcement: we're going to be launching ARC 2, a larger & more diverse successor to the ARC challenge -- an exciting new benchmark for AI reasoning abilities. You're invited to take part in creating it: ARC 2 content will be crowdsourced. Sign up now: arc-editor.lab42.global

Today at POPL 2026, Maddy Bowers presented our latest work on abstraction/library learning, dubbed "Top-Down Synthesis for Library Learning". Thread 🧵!

Awesome National Museum of Mathematics presentation on the discovery of the Hat! A summary 🧵: This is Dave Smith, a mathematical artist. He spends *a lot* of time just messing around, seeing what shapes he can tile in usual ways. Nov 20, 2022, he emails @cs_kaplan to say: he can't figure out...

New ARC-AGI paper ARC Prize w/ fantastic collaborators Wen-Ding Li @ ICLR'25 Keya Hu Zenna Tavares evanthebouncy Basis For few-shot learning: better to construct a symbolic hypothesis/program, or have a neural net do it all, ala in-context learning? cs.cornell.edu/~ellisk/docume…

We reproduced DeepSeek R1-Zero in the CountDown game, and it just works Through RL, the 3B base LM develops self-verification and search abilities all on its own You can experience the Ahah moment yourself for < $30 Code: github.com/Jiayi-Pan/Tiny… Here's what we learned 🧵