If we're moving to a world with humanoid robots, we might need less automated and smart appliances and devices. What if instead of having full self-driving cars you have robot chauffeurs that can drive any car.

The open web is dead, long live the open web. ⌐◨-◨

Building /warpcast and /farcaster

— gabrielayuso.com

Isn't it less efficient? I mean building all the stuff required to have a robot that controls muscles like a human, in order to drive a car, and then the human-car interface (steering wheel, pedals, knobs and switches and levers). It seems more efficient to connect the AI directly to the car, has sub-millisecond input and reaction time, gets direct input from cameras on the visible, UV, IR spectrum, direct input from proximity sensors and radars, and engine status?

Dad, geek, conversation liquidity provider. vrypan.net (home), /weatherxm (work), @l--o--l (fun). Also https://paragraph.xyz/@purplesubmarine

If the robots are going to be built anyway for other purposes, at some point they might be able to drive a car.

It's like LLMs that are much more general purpose than traditional ML models. Then you can tune them to specialize them in something like coding.

If robotics reaches a point where fine motor control, reaction time, vision, etc. is good enough for many use cases, then they'll be able to drive a car.