Remix.run Logo
computably an hour ago

Narrow pre-LLM models that spit out content and ad recommendations have never been capable of also suggesting, let alone implementing, self-improvements.