MapleBench is progressing well, the e2e game environment is working. Im fixing a few other minor issues and then will start eval'ing different models across multiple metrics!
"Announcing" a project means very little these days... therefore this isn't an announcement of a project... instead it's a thread about a project I'm actively vibe coding to enable dark mode for my agent after noticing two major agent x crypto privacy gaps 🧵 ⬇️ (gh link @ end)
@RonTuretzky@colludingnode@mhaydenb and I are live, come hang out!
Hayden is super good at magic the gathering, Ron and I are vibe coding, and cnode is 3d printing! Live from Brooklyn!
Using Codex and Claude computer use at the same time is hilarious. Codex is orchestrating 7 different projects while Claude is fighting to figure out which screen it should click on