AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
How to Maximize the Planning Ability of LLMs While Reducing the Computation Cost
It's always interesting to see how quickly we are having these additional insights that in this case are more theoretical for instance and like that optimization paper as well lots left to discover as we can tell. It kind of makes me feel hopeful there's been a lot of discussion of if you're in academia and you don't have compute you know can you even do anything useful right? I guess all of this is showing that you can yeah so it just shows how much low hang fruit there still is with the architectures we currently have, he says.