raphael lee@raphsblJun 30ArticleVerifying Long-Horizon Tasks with RLMsDuring my time at @mercor_ai, I made a couple verifier variants to address problems with our evals: Browser-Use Verification Verifying browser and computer-use trajectories has become more difficult,...13111.3K
raphael lee@raphsblSep 19, 2025did poke mention @marsxiang_ to every waterloo student who signed up2162.2K
raphael lee@raphsblJul 21, 2025apparently doomscrolling improves my sleep quality. do I trust my whoop or @bryan_johnson?2141.8K
funny how LLM generated seeds are often 42whos going to hackmit
