Comment by cle

I'm struggling to find where these authors are treating LLMs and transformers like they're the solution to everything, or where they are just following hype around? These are long-time AI researchers at Berkeley who've developed a framework for dealing with limited context windows, drawing inspiration from operating systems.