>>510421528
LLMs are like an incomplete brain
that's why my approach is that robots will be the last part of the puzzle
robotics is making great progress in terms of the mechanics and engineering of the robot itself, like balance and being able to manipulate things, move around, get up after falling and things like that
think of it like attaching mocap data to a robot, and that it actually works, that's the impressive thing
but now we have to build a better, more robust AI brain, a multi-model AI with self-learning capabilities, and that's exactly what everyone is working towards
right now we need better hardware, and I think that's coming somewhat soon
then, you'll get another AI explosion moment like the one we've been living for the past 3 years, when the open source world is able to run and train more complex models, combined with self-train capabilities
