The Minutes of Dylan Patel on rl https://minutesof.com/dylan-patel/on/rl/ Everything Dylan Patel has said on rl: 1 verbatim quote between February 2026 and May 2026, each with a timestamp and a link to the recording it came from. en Sun, 30 Aug 2026 16:22:24 +0000 Patel explains agentic training only needs to sync tokens every few minutes versus weights every seconds in pr https://minutesof.com/q/86748cbb-055a-4578-92be-361a647d16ac/ https://minutesof.com/q/86748cbb-055a-4578-92be-361a647d16ac/ “When you're doing these rollouts and especially as things get more and more agentic and training, you might not only need to send not the entire weights but just the tokens that are relevant.” — TBPN Tue, 03 Feb 2026 23:08:03 +0000 rl