So much alpha in this paper by @PrimeIntellect Finally had a chance to read it, impressive work combining decentralized training, async RL, and permissionless compute Truly a masterpiece in engineering and research, worth the reading
19,4K