I've been eyeing robotics for a long time and sincerely believe that a GPT moment in robotics is near. If it happens, the world will change much more significantly than with the advent of LLM.
I've had a SO-101 robotic arm gathering dust for a while: at one of the hackathons, I managed to assemble and calibrate it only 95%, and that's where it ended. It's clear that with neural networks, learning and building prototypes has become faster, so I want to return to it. In simulation, I sketched out what I would like to implement in reality (it's clear that in reality it's much more complicated due to backlash and the fundamental differences between simulation and the real world).
Speaking of robotics: what is the top-1 task you would like to assign to a robot, so much so that you are willing to pay to buy it?
#robotics@rvnikita_blog #so101@rvnikita_blog #simulation@rvnikita_blog