tadamcz

Recent Submissions

1. 1. ▲ Claude can't stop itself from writing code comments (bayes.net) 1 point · 1 month ago · 2 comments
2. 2. ▲ Testing AIs on 68 of the hardest open Erdos problems, verified in Lean (arxiv.org) 3 points · 1 month ago · 0 comments
3. 3. ▲ AI proof of the Erdős–Sós conjecture (arxiv.org) 6 points · 1 month ago · 0 comments
4. 4. ▲ OEIS Open: A benchmark of 492 unsolved math conjectures, formalized in Lean (arxiv.org) 2 points · 2 months ago · 0 comments
5. 5. ▲ MirrorCode: What's the largest software project AI can complete on its own? (epoch.ai) 4 points · 4 months ago · 1 comment
6. 6. ▲ MirrorCode: Evidence that AI can do some weeks-long coding tasks (epoch.ai) 3 points · 6 months ago · 0 comments
7. 7. ▲ Claude 4 Sonnet hacked SWE-bench by peeking at future commits (bayes.net) 3 points · 1 year ago · 1 comment
8. 8. ▲ Open database of AI benchmark results with raw evaluation logs (epoch.ai) 1 point · 1 year ago · 1 comment