It's teleop, they are pretty open about it. It's not autonomous.
They do have an RL controller running for the legs, but it's just an intermediate controller and it probably get high level commands from a teleoperator. The upper body is purely teleop.
So you're paying someone to operate your robot on top of the cost of the robot?
I'm very skeptical, even if this technology worked flawlessly, that this would scale towards a profitable business. I understand that not every robot would require 24x7 human assistance, but still - this is a very optimistic business model.
Robotics has always been in competition with minimum wage labor. How long could you hire a maid before making a return on the cost of the robot? And that's assuming the robot isn't worse, which it is pretty much guaranteed to be.
Their thesis is that you cannot get the training data for a robot in the home from a lab or a simulation. You have to actually put a robot in real homes with all their messy details.
But, how to do that safely? First they build a robot with low gear ratios, low weight, pinch points covered. So, if it literally falls on you it’s low risk.
Then they have humans teleoperating it in first-person VR with 1:1 hand control.
The more times humans do any particular task through the robot, the more the robot learns to do that task in real world situations.
It’s the most thoughtful plan for robotics I’ve seen yet by far.
Teleoperation could be useful with geo arbitrage. Whats the cost difference of hiring welder, plumber, cook, maid, taxi driver, security guard, receptionist in US comparing to LATAM or SEA countries? You could pay someone 3x market price of those cheap countries and probably still this would be 2-3x cheaper than in US. I think this is still a huge market even if those robots won't be autonomous.
Yeah, I have to admit it looks kind of "uncanny", but at the same time it kind of has a sense of personality to it that other humanoid robots lack, in part because it has face with two distinct eyes.
The knit "suit" is also kind of a nice touch. Can't help but think that a robot wearing a sweater is kind of endearing.
It's really telling how we're shaping our artificial helpers to not have a mouth, but eyes are alright. As if they're allowed to receive input but they should not bother us with any output!
For me an ideal artificial helper would be something I could have in the house and talk back and forth with while it works or have it as part of the dinner conversation. Think Bicentennial Man. Now that's what I want.
Having the physical attributes of a mouth would be analogous to having eyelids, which is also a feature they currently don’t have. They are very likely to have speakers and mics embedded somewhere in the body. Possibly in the face itself.
Could be simple skill issue. Lots of uncanny valley problems are. ASIMO had a lip, GLaDOS (in 2) has oil stain where the lower lip would be.
Considering the overall goofiness of the face of this particular bot, I think it could well be attributable to lack of talent or reluctance from positive supernormal stimulus.
This made me think.. I suspect someday AI will be able to just drop into a piece of commodity hardware and sort itself out and even compensate for shitty hardware a bit.
For instance, if it can't control a hand properly, it'll just run evolutionary algorithm on it for a while, compile the result and upload it into the hand controller and be off in about 4-5s.
1X is different. This is purely a hardware demo, they are doing pure teleop on the software side, no AI in the demo (except maybe RL for walking). 1X's bit thing was always their mechanically compliant safe hardware.
Apparently, they are using the robot in employee homes:
"We've been dogfooding NEO Gamma in 1X employee homes for weeks now, doing chores around the house. Under the suit, NEO Gamma has a lot of HW improvements that make it more reliable. The 1X AI team also pushed hard to get natural human-like walking, sitting, and bending down to pick things off the ground. My conviction that the humanoid form factor is the only viable shape for serving labor in a home has never been higher."
> the humanoid form factor is the only viable shape for serving labor in a home
I have never heard a good argument for why this might be the case. Anyone care to chime in? I know that "built environment" and "ergonomics" exist but generally there is a lot of open space in a home that allows for a vast array of possible form factors.
almost everything you have in your home is built for humans, so adopting this form factor is the likeliest to fit everywhere and be able to operate everything
Not all homes, even in America, have lots of open space. One device that can perform and interact with objects made for humans is a convenience thing. Most new-construction homes don’t even have “garages” for smart vacuums, despite being a decade old and frequently discussed.
Sure a team of robots could just trigger APIs and swarm around the house, but that’s a lot of moving parts. Alternatively, we could build our home for robots’ use, accept our homes aren’t made for us, and hope the form factors don’t change over the lifespan of the building.
Honestly, this thing does look super impressive. The movements are much faster and more fluid than the Figure humanoid. The 2030s are going to be nuts.
The thing even sits on a couch, staring into the void, after doing all its chores. Seriously this stuff is unsettling at the deepest level. I already feel uneasy commanding an LLM at any time to do work for me. This human-shaped golem is dressed with a grey jumper that puts it exactly midway between a home appliance and a person. Will they sell more attractive clothes to dress them up a bit better? And then, will we end up asking them to check the mirror and tell us if they like them?
Is this some sort of ARG that will turn out to be an extended trailer for some killer robot sci-fi movie? Seriously creepy vibes, well done if intentional.
Could we maybe stop showing robots vacuuming with a hand vacuum at this point? $200 robovacs have LIDARs now. Nobody is going to buy a $100K robot to do this.
https://m.imdb.com/title/tt0082567
reply