RL Nabors argues for local models over frontier models

Former Meta/React core member RL Nabors advocates for using small local models instead of frontier models for most tasks, citing savings in latency, energy, and cost. He covers practical deployment trade-offs in a talk at AI Engineer.