Low volume industrial type of manufacturing and assembly. Not something like automotive where the sheer volume of production allows for heavy investment in fixed robotics, but things that sell in the few thousands per year. Or at a cheap enough price point straight fabrication work. Lots and lots of things are produced on manual production lines by small fab shops, which in turn feed your larger production facilities. Outside of automotive there just isn’t the volume to support spending hundreds of millions or even a tenth of that on a vast majority of things out there.
What’s one product that can be manufactured more efficiently by humanoid robots than by some other kind of process? And if there really is a demand for such robots then why haven’t any manufactures started using them yet?
Don’t get me wrong, these robots are pretty cool. But “cool” and “useful” are very different things. Humanoid robots strike me as an invention in search of a need more than an invention created to fill a need.
5 Likes
It’s not one product, it’s a company that makes similar products but at lower volume. Something like a side by side or a skid steer isn’t produced on a fully robotic line like a Camry. It’s job shops and international suppliers welding components up and people assembling them. Someone like Bobcat make several different models of skid steers. They are all similar enough to be produced on a single assembly line, but they are all different enough that you need a “person” who can reach inside, outside, over, under, basically everywhere. Leaving manufacturing, what about a robotic automotive tech?
Price. People are cheap and easily replaced. Why doesn’t a $10k EV with 500 mile range exist yet, surely there is demand? Because the tech and cost to manufacture isn’t there yet. At some people robots will build robots (I’m speaking in capitalist terms, not AI taking over the world aspect). Once you have vertically integrated the entire production with a general purpose bot the cost to manufacture is pretty much irrelevant. At this point they are all proof of concept. Those with deep pockets can experiment with them, but they are just stepping stone to more advanced units.
1 Like
Assuming you could build a robot smart enough to repair a car, why would a humanoid form factor make more sense than (say) an R2-D2 style droid with tools for limbs? Or even a team of specialized robots for specific automotive tasks? You could have a low-bodied wheeled bot that could easily inspect an undercarriage without even having to get the car on a lift, or a robot arm with a pneumatic socket wrench that could do tire changes as fast as a NASCAR pit crew. None of those tasks require a humanoid form factor.
It doesn’t usually take decades for people to dream up a practical use case scenario for a genuinely useful invention.
1 Like
I’m genuinely surprised that some filmmaker hasn’t already hired Boston Dynamics to build a bespoke robot character for a movie project. How much do these prototypes cost to design and build?
1 Like
So you’re saying long after we are gone the robots will be stuck with our arbitrary railroad tracks widths, and staircase steps heights.
6 Likes
Ah, Space Truckers. Highly underrated movie, A list names, D list plot, and B list gags. Worth it for the cheese factor and the Square Pigs.
At some point, we are all Secret Dumbasses.
2 Likes
The first bicycles were comically impractical. The first automobiles were inferior in every way to a horse and cart/buggy, other than the fact that it didn’t poop in the street.
Right now for anything automated, it is more economical to have a specific task design, and isn’t a humanoid. Like putting on a car door and welding it, all you need is a robot arm.
IMO, in the future when they get advanced enough, and humanoid robots are both cheaper and work better, they will do a multitude of tasks, a general purpose robot. (ETA) Though point taken, maybe a general purpose robot in the future will look different than humanoid. But, like, our world is build to be navigated by humanoids, including turning wrenches in cars. So it makes some sense future tools will retain a humanoid shape - OR - our world will change to accommodate the future robot shape.(/ETA)
There will still be job specific automation, especially at bigger companies, with the robot arms or other specialized form factors. But a small or medium sized company could buy one robot that does many tasks, vs 3 or 4 separate robots to do specific ones.
That’s what I see, at least. It may turn out the only thing they are good for is use in movies/entertainment, killbots, and movies/entertainment about killbots.
2 Likes
In an interview the Boston Dynamics CEO explains their business strategy:
This part of the interview was telling:
So it doesn’t look like they really have a very different strategy this time around, and they admitted elsewhere in the interview that they don’t have many customers that buy more than just one Spot robot for experimentation, so that doesn’t seem like people have found them to be useful.
The fact that they’re part of Hyundai now, and that the Hyundai factories are available to provide friendly testing grounds, means that they should have plenty of opportunity to find out how practical these things can be in industrial applications. Guess we’ll see how that plays out in the next few years.
2 Likes
It absolutely did not take more than a quarter century between the invention of the bicycle or the automobile and the first people making practical use of bicycles and automobiles.
Turning a bolt is a perfect example of something that a machine that has a built-in driver attachment could do far, far more efficiently than a robot holding a wrench in a humanoid hand could do.
Why not buy several simpler machines that are good at specific tasks instead of trying to make one machine that’s good at everything? Like I said, a humanoid robot is not going to be able to vacuum your floor as well as a Roomba or wash your dishes as well as a dishwashing machine or crash your car as efficiently as a Tesla autopilot even though it’s probably going to cost more than all those things put together.
2 Likes
How about home health care support.
Moving from bed to a chair.
Chair to shower.
Bedroom to kitchen.
Help in the kitchen preparing meals.
Moving large items around the house.
You can probably build purpose built robots optimized for each individual task, but at some point you’re using 6, 7, a dozen different things and a single humanoid one would be more useful. In this use case, the robot is navigating and interacting with the humanoid environment.
Home health aids, using a variety of tools can accomplish these tasks too. The tooling and the labor costs are both expensive. It’s a tough job and not attractive to many people either. All things that make it a good target for a robot solution.
I look forward to my robot caregiver assistant of the future allowing me to age in place longer.
3 Likes
And helping you to pull off one last heist?
3 Likes
Even assuming you would want one robot to do all of those tasks a humanoid form factor doesn’t necessarily make the most sense. If you wanted a machine to lift and transport a person then designing it to balance on two legs would be overly complicated and needlessly dangerous when you could just design it to move around on treads.
Who do you think would be easier to knock over: C3PO or Johnny 5?
2 Likes
If that’s an option, it would certainly help with retirement income planning.
I was picturing a time when I would have less mobility than was displayed in that movie trailer. Maybe the robot can do the heist on it’s own?
1 Like
Perhaps. But, having done this years ago for a family member, you learn very quickly that houses are designed for people first. Even getting a wheelchair around many houses is extremely difficult.
From bed to a shower. First there’s the lift out of bed into a wheelchair, using the lift machine… Navigate the chair into the bathroom, the one with the special door hinges to make the opening 2 inches wider and barely fit the chair. Then, the remodeled shower with no threshold to roll into.
Johnny 5 is going to cause some serious structural damage along the way. While C3PO can barely lift a foot. Perhaps more of an IG-88.
Japan’s been working on this solution for a long time.
1 Like
Yes, with wheeled robots.
We’ve had treaded robots that can navigate stairwells and doorways in practical deployment for police, military, search & rescue etc. since at least the 1980s. There’s not really any need to build a bipedal humanoid robot beyond the “cool” factor.
4 Likes
It feels worth noting that nature does not have many species built like humans, and except for birds nearly everything walks on more than two legs. It’s clearly not a very robust design.
4 Likes
Exactly. If you’re starting with a four-limbed organism it might make evolutionary sense to sacrifice the use of two legs in favor of winged flight or dextrous, tool-using forelimbs but it would hardly make sense to design something that way if you were starting from scratch. A general use robot could have as few or as many limbs as its tasks required.
And of course evolution didn’t really have wheels as an option at all.
1 Like
It took about that before cars were practical enough for them to be more than a novelty. But cars are very simple machines (an automated horse drown carriage). compared to an autonomous robot. See below for more*.
Sure, the humanoid robot could have a socket wrench attachment for a finger or something. It doesn’t need to physically turn a wrench. The humanoid form factor is still useful to walk over there, hunch over the engine bay, and stick their arm in there like a human would. Then use a different finger to plug into the little socket thingy where it reads the cars diagnostics for errors. Another finger that tests if a spark plug is good or not. Then it gives Jimmy a back rub with a vibrating palm because he played too much pickle ball yesterday. And finally cleans and sanitizes the toilets with a spinning scrub brush and a combination UV light and flame thrower before plugging in to recharge.
Maybe that will be the way to go? But if a device CAN do more than one thing and it does it well, there will be a market for it. And again, maybe it won’t look humanoid. I guess it will come down to what works well and/or what people end up liking - do we like humanoid looking robots, or prefer it to look more like R2-D2 or Eve from Wall-E.
Where is the robot working? A Roomba doesn’t need anything more than little wheels because its working on one level of a house. A robot in a warehouse might be wheeled or tracked. But if you have to go outside and navigate in a city made for humans, with stairs and hills and uneven side walks, you may need something that can walk (and yes, we have ADA compliance, but taking my mom in a mobiltiy scooter around San Antonio showed that the world is still made for humans with legs. 
).
*We aren’t even there yet to where if we had a job for it to do, it could do as well or better than a human. So the whole “we don’t have an application for this form factor” is partly because the technology isn’t there to do the tasks well enough.
It may end up being where we use them for service jobs. Yes, a Barista Bot 3000 with a dozen octopus-like arms that can squirt out any concoction in seconds and instantly heat or chill it would be more economical. But maybe people will prefer a humanoid looking robot making the drink on a traditional cappuccino machine. Maybe we find out if we make them look like people, people treat the robots better because of anthropomorphism. Maybe we treat them worse and want robots to look like robots.
You’re right it might go the way where robots stay specialized gadgets that are task specific. It could be a mix of both (look at Star Wars, you don’t need your walking power source to have a face, but it helps with your teaching droids. Or a humanoid top, but a wheel or tread bottom). I think the future will have a wide variety, IMO.
I think man is compelled to make things in their own image, and honestly I am OK with making something that doesn’t have (currently) a real world practical application. Can we just - make stuff because we can? An impractical novelty? If it progresses far enough, we may later find uses for it that we aren’t even thinking of now. I think this is all very far off future-y, but possibly in my life time, and it will require a paradigm shift in the economy.
From an engineering stand point, that is very practical.
From an interacting with humans standpoint, unless you make them look like lady bugs or butterflies, most people are going to be icked out by them.
2 Likes
