In Praise of Useless Robots
thereader.mitpress.mit.edu
thereader.mitpress.mit.edu
This article comes just in time to silence the heretics who ask what the use of an Optimus is. Promise a factory robot, deliver a Tamagotchi.
On the factory floor of a car plant, every job that was easy to do with an industrial robot arm bolted down to the floor next to the assembly line and running the same sequence on repeat 24/7? It's already done by an industrial robot arm bolted down to the floor next to the assembly line and running the same sequence on repeat 24/7.
What remains is: the many, many gaps. All the tricky little hard-to-automate tasks that humans can do, but industrial robots would struggle with. Tasks important enough that they have to be done, but small enough and tricky enough that it's not worth redesigning the entire process to make them amenable to automation or building a special snowflake robotic system to do them.
Which is where the general purpose worker robots come in. If the same robot can be trained and retrained for humanlike performance on multiple of those tasks, then it becomes a worthwhile addition to any high automation factory.
Of course, that requires a general purpose worker robot! Which is very much an AI problem. Tesla can build the hardware, but they're banking on being able to back it by software too, and that clearly isn't there yet. Whether it will be remains to be seen - but Tesla isn't the only company banking on being able to solve robot AI.
I was with you for the first sentence, then you started diverging. Obviously the purpose is to strike fear into the minds of Tesla factory workers, otherwise they'd have built a completely different robot.
Do robots by robots?
Maybe you've confused generative AI with humanoid robots (powered by similar Machine Learning models to generative AI). And to anyone looking to inundate this thread with edgy nonsense, vague moody negativity stops looking clever to anyone with mental age above 14.