Tag
A user discusses the high heat generated when running large LLMs on MacBook Pro with unified memory, questioning if others run such models continuously or only for benchmarks, and seeks recommendations for a small model for constant use with agent tasks.