"Robotics hasn't yet had its GPT moment, but we think it’s close"
From what I have been reading from those who run top-tier research labs in the field most agree that we are really far away. Why would people casually quote 30y away? https://techcrunch.com/2023/12/02/robotics-qa-with-metas-dhr...
It is because core issues remain unsolved, has nothing to do with some old engineering process that can be disrupted and more so with the fact that science isn't there yet. Most of the GPT wrappers that are used as reasoning still sit on top of classical low-level control. Personally, as someone who works in AI specifically around robotics I believe there is headway to make here.
replies(1):