That's the same question a lot of people had during the Fukushima disaster in 2011. People were trying to contact Honda to convince them to send their ASIMO robot to shut off valves that could not be accessed by humans. That was well beyond it's capabilities. ASIMO was designed for the stage, not for a disaster zone. No one had the know-how to build such a robot.
Flash forward to today, we are still in quite the same position where robots can do fancy, flashy tech demos, but when it comes to doing something useful that is also unpredictable, the know-now is still not there. Even teleoperation is not a robust answer to this yet, it still has some maturing to do.
So that means we just stop innovating and not try to go beyond what we already have? This is not regression since this has no effect on the robots we already have, industrial robots are on their own innovation branch. These are prototypes, a glimpse into the future at what's possible with more experimentation and technological progression.
Why not be impressed? The reasons you gave are enough for the majority of people following along. The exciting parts are the steps along the way. Once we have something that is actually usable and integrated into our everyday lives, it will lose its novelty fast.
“Atlas is autonomously moving engine covers between supplier containers and a mobile sequencing dolly, using ML to detect and localize the environment fixtures and individual bin. There are no prescribed or teleoperated movements.“
Not exactly. If you actually look at those hands on Disney's A1000 animatronics, they are actually pretty basic, usually 5 to 6 degrees of freedom at most. They are designed to be simple, because as you mentioned they are in use for like 16 hours a day. The more complicated you make it, the more likely it is to break down. Each hand is designed to give the illusion of life with the least amount of moving parts possible.
A more accurate comparison would be the Shadow Robot dexterous hand used by openai to rotate a block one handed about 6 years ago. That one actually has 2 more DOF than the new tesla one. Like tesla's, it's also cable driven, but it's pretty slow and jerky.
What I think the new tesla bot hand now achieves is the lifelike A1000 movement with a vastly more dexterous range of motion like the Shadow. It's kind of the best of both worlds.
Flash forward to today, we are still in quite the same position where robots can do fancy, flashy tech demos, but when it comes to doing something useful that is also unpredictable, the know-now is still not there. Even teleoperation is not a robust answer to this yet, it still has some maturing to do.