Do You Think We're Going to Be Using Metah Rl Algorithms?

i wouldn't be surprised if every rl algorithm used had some automatically tuned component whether that is directly using metah rl or some other form of auto rl. i would also be surprised if it turned out that the long horizon multitask setting will end up giving us something that could beat state-of-the-art methods we're hand designing. There's definitely a lot of people who believe that uh the the you can do better like you know learned optimizers and those kinds of things are very relevant here.