PERSON DIRECTORY
Eric Jang
Eric Jang appears in 1 indexed conversation across Dwarkesh Podcast. This directory brings every appearance, source, TL;DR, digest, and transcript into one searchable feed.
1 EPISODE1 SHOW
Language
What rebuilding AlphaGo teaches us about self-play, RL, and future of LLMs - Eric Jang
Rebuilding AlphaGo has cut a project that once required DeepMind and “millions of dollars” to roughly $10K, using rented compute and LLM coding agents.MCTS supplies improved labels with far lower learning-signal variance than policy-gradient LLM RL, while forward search for language may make a comeback despite PUCT’s uncertain fit.