Olive Song discusses MiniMax's open-weight models and RL stack

Song, who leads reinforcement learning at MiniMax, explains the stack behind its open-weight models and argues for an open-source bet: release weights, let builders optimize, and share the gains.

Featured · Olive Song