I'm not sure how many times this has to be restated.
It's car manufacturing. Everything that could be done by a purpose specific robot arm bolted down to the factory floor is already done by a purpose specific robot arm bolted down to the factory floor.
What remains unautomated, then?
The long tail of tasks that are too minor, too finicky, too open-ended or too reliant on manual dexterity to be offloaded onto traditional robots.
This is where this new generation of robotics comes in. This is the kind of task they're designed to do: "a task that's still done by a human in a high automation environment". Universal robots are angling for the tasks that are impossible or uneconomical to automate with traditional industrial robots.
> Everything that could be done by a purpose specific robot arm bolted down to the factory floor is already done by a purpose specific robot arm bolted down to the factory floor.
Hah! Hardly. I say this as someone whose first "real job" was in applying robotics research to automotive assembly - there are still a ton of assembly tasks that could be performed by a fixed-base robot arm, or a robot arm on a linear rail/fixed gantry. Wheeled mobile manipulators are only needed in a few cases, and humanoid form-factor is only "necessary" in very few cases (and I don't think the current crop of humanoids is particularly suited to these tasks).
In my opinion/experience, the impediments are that (1) the system integrators that are usually responsible for assembly-line robotics are too stupid to figure out how to apply robots to the problem, (2) the automakers themselves are often too short-sighted/stupid/unwilling to invest in increased automation (and particularly in building the in-house competency that they really need), (3) the hostile/exploitative relationship between (most) automakers and their main suppliers means that low-hanging improvements to parts/assemblies are a non-starter, and (4) the automaker C-suite (and investors) are too drawn to silver-bullet solutions (e.g. humanoids) than practical automation improvements.
"Could be in principle" and "could be in practice, under technical and economical considerations in play" are two very, very different beasts.
Everyone in the industry learned that the hard way.
At a certain point, the tasks that remain stop being "dexterity" problems and start being "AI" problems. That is: a robot could do the task - if you either spent big $$$ on redesigning the entire task around the robot's intellectual limitations (uneconomical), or if you had an incredibly advanced AI capable of problem solving driving that robot (impossible with 00s AI).
The "universal robot" bet is the "incredibly advanced AI capable of problem solving" bet. That in 2020s, AI is finally capable. The body only has to be "good enough to make most tasks possible".
> "Could be in principle" and "could be in practice, under technical and economical considerations in play" are two very, very different beasts.
> Everyone in the industry learned that the hard way.
The auto industry is notorious for making incredibly myopic choices to save money/make money in the near term versus long-term investments. The relationship between automakers and their suppliers/vendors is basically a century-plus of the automakers trying to (1) outsource anything they can for a quick buck, and (2) grind the supplier/vendor margins down to nothing. (This is part of why the newer Chinese automakers with much greater vertical integration are such a threat to the traditional automakers; vertical integration has a high up-front investment but the payoff in flexibility and speed is significant).
How long ago was your robotics experience?
An Amazon warehouse or Tesla factory tour would likely change your mind.
I had to do both of these in the last year and not a lot of humans around…
I have visited factories for work and my experience is the same. There is so much stuff that could easily be automated but is not because it is too expansive for too little value to make a custom one off machine. The big high volume things will be automated but these machines will have 90% success rate and lot's of stuff that needs to be done by hand. You can search for factory tours on youtube to get an idea. Here are two videos, an Amazon warehouse and a Tesla factory. the big heavy stuff is automated but lot's of work is still done by hand.
https://www.youtube.com/shorts/-R6cBkza17k
https://www.youtube.com/watch?v=45slYC99uUg
> How long ago was your robotics experience?
This is over the last decade at one of the largest automakers in the world. Naturally there is significant variation between individual lines and plants; some are newer and more automated, some are older and much less automated. Are some cars being built on more automated lines? Yes. But a great many, probably the vast majority, are being built with fairly low assembly* automation.
* There is a significant split in automation between "body weld" stages and "assembly" stages. Body weld is very heavily automated basically everywhere (although there are some surprising exceptions in places), while assembly is much less automated.
“One of the largest automakers in the world” makes me think of very low-tech companies like Ford or whatever. I can’t imagine this would bring much actual experience with this new generation of robotics.
What an unnecessary comment dismissing the expertise of an actual expert - what’s your robotic experience to dismiss him so contemptuously?
automotive workers unions started around 1918 and became major political players in the 30s -- a fact that i'm sure is wholly unrelated to why there are so many un-automated tasks in that industry.
Hyundai’s manufacturing facilities in the U.S. are not unionized.
Amazon warehouses still have a huge number of ununionized workers doing manual labor
Pretty different tasks, environments, outcomes, metrics, goals, and other things in a warehouse vs. a factory… really have no clue how this is supposed to be relevant. Why not mention farms or libraries?
Also, a general purpose robot vs a custom purpose robot represent very different capital investment profile for the factories?
But if these tasks are too minor, too finicky, too open-ended or too reliant on manual dexterity for a purpose-built robot, how can a general purpose robot perform them better? If anything, they should be doing worse.
The only thing I can think of are tasks that are so rarely done, it's not economical to build a robot for. But I then I also don't see how another robot solves this problem.
A) the idea is that these robots do have dexterity capabilities a lot closer to human hands
B) there’s a long tail of individual tasks it’s uneconomical to build purpose-built robots for each individual task. But it’s economical to have 1 robot that can do all of them.
These industrial robots have much better dexterity than any human alive.
The point is, human shape plus general purpose intelligence is an amazing combination to resolve the “long tail”.
Without the intelligence part, the body is useless.
Perhaps Boston Dynamics has that part resolved now too.
> much better dexterity than any human
Do they? A human can both chuck kilograms of stuff across a room or kick in a door, but then pick up a single hair off the ground, or feel and manipulate (things even lighter than) a literal feather.
Robots can certainly do things more repeatably, if not more precisely.
Intelligence is absolutely a valuable addition to dexterity, but no, current industrial robots have nowhere near the dexterity of a human hand.
The human arms and hands are very versatile, and imitating them is a good choice for a universal robot, though 3 or 4 arms are definitely better than 2, and the hardest to imitate are the sensors, not the actuators.
But the rest of the human body is not useful in a factory environment, so the arms could be mounted on a mobile base that does not have any resemblance to a human.
And C) they don't always have to be at parity with human hands to be good enough because humans are flat out expensive. Humans need health accommodations, have sick days, vacation, and make mistakes too. The bar is much lower and the incentives are much higher than many people probably think.
and humans collude via unions
These robots operate on completely different principles.
One can lift insane weights, has insane torque, and absurd precision, and can do the same movement millions of times with virtually no deviation. You program these with an exact movement plan, just like you would programm a CnC with a tool path. They are basically cnc machines.
The other one is a inacurate, unstable, dynamic system controlled by neural networks and heuristics. It has massive deviation over each run, but that means that the programming must be able to account for it. Which makes it suitable to operate on problems that are messy, unrepeatable and human-shaped.
Pretty much. It's a total paradigm shift from how industrial robots normally work. A pre-planned motion executed carefully and precisely vs open-ended "do this thing" powered by a very large bag of opaque neural network heuristics.
A robot that has to be carefully adapted and set up for the task vs a robot that you can point at a task and have it figure out how to do it. A robot that doesn't deviate vs a robot that absorbs all kinds of deviations.
It's a bet that The Bitter Lesson will win over Moravec's Paradox, in the end.
It's also worth noting that when e.g. inputs to a stage might have unpredictable defects or alignment, a robot arm utilizing neural networks for planning and analysis might still be the best way to handle that - without the extra degree of freedom of movement-relative-to-floor, planning can be done more rapidly, and movement can be executed more aggressively and quickly.
If I were Hyundai, I'd be looking at this as buying a significant amount of vision, dynamics, and integration systems expertise, not necessarily the dream of self-motive walking systems.
Well, humans obviously do those jobs, so a clearly a general purpose robot (in this case, a biorobot) has been found to do the job better. Don't overthink it.
Because it is general purpose. We did not have the ability to create a single robot form which could do all of these minor, finicky, and opened ended tasks. Now that seems within reach. The nice property of humanoid robots is that the world is already made for human form, and so if you're trying to replace people naturally this is what you'd want.
>> how can a general purpose robot perform them better
Better than what? It seems that as long as they perform the tasks "better" (cheaper / faster / lower-error) than the humans that are currently performing them, that is an improvement for the factory owner.
It's not a "general purpose" robot, it is a "human replacement" robot, with similar skills and shortcomings to a human. Humans are not general purpose.
All you need to do is look at a recent video of car manufacturing process, and watch what the humans are doing.
>It's car manufacturing. Everything that could be done by a purpose specific robot arm bolted down to the factory floor is already done by a purpose specific robot arm bolted down to the factory floor.
>What remains unautomated, then?
Stuff that can be done by purpose specific robot arms on wheeled platforms, which is very difficult, but will be much more feasible than a humanoid robot doing anything.
What you need then is a better arm (or even just hand), not a human.
Or a new take on car design with automated production in mind regarding all the wiring and what not (easier said than done, I'm sure many have tried and failed, but eventually someone will succeed).
> I'm not sure how many times this has to be restated.
This strike me more as a repeated internet myth more than anything else. There is near endless opportunity for purpose-specific robot forms.
> What remains unautomated, then?
And the tasks that change from day to day.
Naw, the real answer is that factories have been built around human labor - they weren't built to be forward-compatible with purpose-built robots, so during the transitional period where we build these purpose-built robots, you need humanoid robots to fill in the parts where the factories were geared for human labor.
According to this widely cited comic strip, if you are over 30 and didn't know it, you should be ashamed.
'Everyone over 30 knows this' is a prior assumption (it is not necessarily correct; and nothing is said about shame).
The comic strip is saying if above is true, then people still have to learn at some point so on average it would be around 10k people per day.
I think the math is this:
For people born in a given year: 4000000/365/30 = 365 people per day
but you have 30 sets of those people (those born this year, those born last year, those born two years ago, etc.) So 365 * 30 = 10950. 10k is easier to say for viral purposes.
I don't think that's what it's saying at all. It explicitly says we should not be negative towards people for not knowing things.
The math on the xkcd is just wrong tbh. In multiple ways.
I took a tour of the BMW Group Plant Spartanburg body shop. It's heavily automated with industrial robots inside safety cages. But they still have human workers pick up parts from rolling carts and place them into templates for the robot arms. BMW has been running a trial with Figure humanoid robots to automate that remaining piece. Apparently those robots haven't worked very well, but presumably Hyundai thinks they can do it better?
https://www.bmwgroup.com/en/news/general/2024/humanoid-robot...
Some colleague of mine visited the Mini plant earlier this week. Apparently they had a Boston Dynamics dog patrolling simply to spot stuff that had been left where it wasn't supposed to be left
Can't you just, you know, stick another robot arm on the damn carts and have it offload itself? Surely there's a simpler way.
Maybe what they're actually acquiring is Handle, not Atlas.
Humanoidal robots make sense when they need to operate in spaces designed for humans bodies. Cars are designed and built to be used and serviced by humans, especially their interiors so you need humanoid robot to automate building them. Car exteriors are not built for humans to interact with so they are already being built by specialized robots.
Custom built robots are expensive (basically an R&D project in themselves) and inflexible, if you want to update the process you have to redesign your automated system. The dream of humanoid robots is they can adapt to new manufacturing processes like human workers.
Two things: 1) we have abundant training data for humanoid embodiments (watch humans do things), and 2) the world is already designed for humans.
Training data of task completion. See, e.g., robots doing backflips. Presumably there’s an optimal robot for gymnastics but if you start with humanoid form you can train based on many videos of human movement. The alternative - world model sim with physics and loss functions- perhaps ends up being too unconstrained when you add in optimization of the robot form…
AI watching videos of humans seems an incredibly inefficient way to solve acrobatic balance. It is just physics and engineering. The hard part is in the materials and assemblages for enabling complex and subtle movements and the fine motor control, not knowing what the motions are.
I understand the sentiment, but you assume they are planning to build humanoid robots to walk around the human-oriented space in the factory.
Perhaps they want to put some of the sensing and control features in, so a humanoid-like dexterity or adaptability for the business end of a floor-mounted robot arm?
Maybe it's like Formula 1: it's a purposefully-extreme goal, which drives new development and makes "lesser" goals feel more achievable.
Because a humanoid robot can replace (theoretically) any human worker, whereas a purpose-built robot can only replace one kind of worker. Or at least that's what Asimov said in Caves of Steel.
I'm spitballing here as I don't actually have a concrete answer for you. But from my understanding automobile manufacturing is one of if not THE most advanced 'purpose-built' robotics sectors. While I agree with you that having a purpose-built thing usually wins out for assembly line manufacturing, I wonder if this isn't an attempt to branch out away from single-purpose robotics into more general or multi-faceted manufacturing.
I'm looking for a better video of it (from one of the engineers), but look at the NASA robot hand. https://www.youtube.com/watch?v=vfDXzkFHnz0
https://www.google.com/search?q=nasa+robonaut+video+hand+why...
The gist of it is that all tools on the spacecraft (eg: space-drill, space-coffee-maker, space-airlock) are all designed to fit a gloved human astronaut hand. Waaaaay more complicated to make a robo-hand than a robo-suction cup or robo-claw, but then you are matching the environment, and guarantee tool compatibility against all extant tasks!
We already have specialized robots on earth... paper slicer, lawn mower, bazooka, whatever. They're all machines that are specialized for the task at hand, we're not making a humanoid robot that gets down on all fours and individually plucks blades of glass.
The car factories already have specialized robots... they're not mimicking a human hand holding a can of spray paint, shaking it up, and painting the car that way... it's a 6-axis arm, or a whole "grab the car and flip it while spraying paint" system.
It's not about inventing purpose-specific robots, it's about handling that long-tail of "stuff with tools that a human is designed to be able to use." Go over there, push that button. Go move that box from table1 to table2. Etc.
For well defined tasks in the factory domain, make a "real robot". For ad-hoc tasks in the interim... strap an LLM to a camera, battery, robo-legs and arms, cross your fingers, and hope for the best?
I think the rationale is that they are already using typical car factory automation, but they see a huge potential market for general purpose robotics in the coming decades, they don't need the humanoids, they are simply dogfooding a future product.
I think this is smart and not very risky. Tesla is playing a similar game with Optimus, for now Hyundai/Boston Dynamics is at least 5 years ahead.
Our'legacy' world is built for human shsped operation. So a generic operator will hsve to mimic human shape.
The human form IS the most optimal way to do most tasks
Most human tasks. Earth-movers and mining-trucks aren't human form.
The first somewhat general purpose humanoid robot will sell like gangbusters to the rich, even just as a parlor toy/trick.
That's like saying there's no need for a generic CPU. The only way forward is a ASIC. Once a generic CPU does everything well enough, it's extremely versatile.
They have god knows how manny bajillion dollars tied up in machinery designed for human use, a human can step in when the robot breaks, and they can buy more robots if the humans get uppity. Those seem like a bunch of good reasons to me.
The humanoid form factor is certainly may not be ideal but I guess they think the flexibility is worth it
Patient complex moves first.
You don’t understand because you’re not an expert? First off they have numerous non-humanoid robots if you follow them at all. Second, Clearly they’ve got strong reasoning, they’ve just been bought. Third, out of thousands of attempts at humanoids, their robots are seemingly the best we’ve yet seen in this class. They must be doing it right when so few others got any traction.
This same nonsensical question again. Because the world is built for humans, because these are general purpose to replace a human laborer. It can immediately go from picking up parts in a factory to mowing the lawn in the same day.
> they need to be useful
They would be. When everything what could be done would be done by a robot. 24/7. Even without the lights on the floor.
Why do you need software when FPGA can solve everything.
It's much cheaper to create one general-purpose robot for everything than many different robots, each optimized to each task.
Robot Chicken had a fairly cynical take on this. I won't link it here.