Newsletters & Bylines
Nathan Lambert's technical deep-dives on frontier model training, RLHF, and open models, from inside the research frontier.