🆕 Training Agentic Reasoners today's feature is @willccbb's triumphant return to the AIE stage RL track - now as part of @PrimeIntellect! A lot of agent builders are basically doing "RL by hand". He concisely explains current RL algorithms in one slide (!) but then argues that RL - particularly for open models - is stuck in math and code Q&A land the new hotness is multi-turn agentic RL, and the new verifiers library is the ultimate toolkit for building an agent and turning it into an RL loop. More people should be exploring building better agent models and Will + PI is enabling that for everyone!