PERSON DIRECTORY
Will Brown
Will Brown appears in 1 indexed conversation across Latent Space. This directory brings every appearance, source, TL;DR, digest, and transcript into one searchable feed.
1 EPISODE1 SHOW
Language
⚡️Multi-Turn RL for Multi-Hour Agents — with Will Brown, Prime Intellect
Claude 4’s differentiated signal is a move from standalone reasoning toward tool-using agents that can sustain longer work, with reported over-eager coding behavior falling from roughly 45% to 15%.Multi-turn RL must reward useful tool results rather than tool calls alone, making intermediate states and LLM judges important for improving agent reliability beyond brittle deterministic benchmarks.