The World Humanoid Robot Games took place in Beijing a few days ago. They’re fascinating and baffling to watch.
That “Humanoid” in the title is clearly a significant constraint that artificially forces the human creators into making some very bad design choices. But why? No practical devices will be built with that limitation: in the races, for instance, if you’re going to mimic biology, you ought to go for four or six legs, and why bother to imitate an animal, just put wheels on the goddamn thing. What are the advantages of requiring your running robot to clumsily use a bipedal form? I suppose it’s an exercise in building limbs to a difficult and ultimately stupid specification, but why?
It is kind of menacing and intimidating to see a field of soldier-like robots, but only because we carry along associations with human armies. When I’m someday hiding in the post-apocalyptic rubble waiting for the Chinese Robot Army to kill me, I’m not going to worry about the bipedal ersatz human tripping awkwardly through broken terrain, I’m going to be scared of the flying drone with lasers, or the rat robot scuttling through wreckage to drop a grenade at my feet. The robots at the games are just comical.
In more benign situations, I’m not bothered by a machine being able to run faster than I can — I’m used to that. An autonomous e-bike could beat those robots easily.
The video focuses on the failures, which is a little unfair. I was actually impressed most with how some of them compensated for failure. They fell down, and then smoothly pivoted and twisted and stood right back up, unless they caught on fire, of course. There’s some really good and clever design work going on there. There needs to be, because people are bad models for high-performance physical activity.
I do have questions. Are these robots required to operate autonomously? If there is some Chinese nerd hunkered down behind walls of monitors with a joystick and keyboard, triggering series of pre-programmed routines, I’m much less impressed. If they’re run by AIs with massive datacenters trying to model human behavior, I’m also unimpressed, other than by the raw wastefulness of the exercise. I’m also curious about maintenance. Sure, you can fire off actuators to move a limb at a high rate of speed, but for how long before one burns out? And if losing one little electronic piston sends your whole elaborate robot with its sensors and cameras and expensive circuitry crumpling against a wall, what’s the cost-effectiveness or reliability of the robot?
I suppose there will be some long-term return on investment trying to solve those little problems, but I think the first solution is to ditch the stupid “humanoid” restriction and design machines that do specific tasks well.
1943, Anthony Boucher: “Q.U.R”
Quinby’s Usuform Robots
fusilier, surfacing from lurk mode
James 2:24
Because the whole thing is based on the tropes of mainstream science fiction. It’s not about solving real problems, it’s about selling familiar fantasies to dumbass VC investors.
I saw some of those videos and had the same response. As an engineer I’m fascinated by the achievement, assuming they are really completely autonomous. But I’d be a lot more impressed if they would show me one of those things standing still in the garage by the bikes and lawn mower. Then when I give the command it calmly trundles into the house, gathers up the laundry and starts a load in the machine. Or cleans the cat box, then trundles back out into the garage and shuts itself off.
“Talking of horseless carriage suggests to my mind that the horse is doomed. The bicycle, which, 10 years ago, was a curiosity, is now a necessity. It is found everywhere. Ten years from now you will be able to buy a horseless vehicle for what you would pay today for a wagon and a pair of horses. The money spent in the keep of the horses will be saved and the danger to life will be much reduced.”
“It is only a question of a short time when the carriages and trucks of every large city will be run by motors. The expense of keeping and feeding horses in a great city like New York is very heavy, and all this will be done away with. You must remember that every invention of this kind which is made adds to the general wealth by introducing a new system of greater economy of force. A great invention which facilitates commerce, enriches a country just as much as the discovery of vast hoards of gold.”
Thomas A. Edison
New York World of November 17, 1895
From this article that I found to be quite fascinating and apropos…
Get A Horse – Saturday Evening Post February 8, 1930
Just sayin :)
I see humanoid robots as a difficult problem to solve, not for its own sake, but for the “side effects,” like how we got some cool “space age” materials and machines from going to space. While I’m a fan of mechs, I don’t think we’ll ever have much reason for bipedal robots when the alternatives are simpler.
The humanoid design is good for versatility, but is not optimized for any specific function. Battlebots shows us robots designed to destroy robots, and they trend toward looking like lawnmowers with the blades exposed.
Well.. I am kind of with Recursive Rabbit on this. Its an interesting engineering problem. I am sort of with PZ on this, but the reality is that some of the same problems would exist with bipeds, or the like. Wheels or treads are easy. They tend to be stable, but also limited. Anything that walks, presumably, could also learn to climb, and presumably could thus climb things, and maneuver, in ways that a) only a drone could duplicated, but b) unlike a drone, could presumably do so while also carrying a load, or performing a task, which requires bracing, strength, etc., that the drone simply wouldn’t have. So, solving the bi-pedal problem translates, in principle, to solving problems with all walkers.
That said, one thing PZ comments on I would object to a bit. You don’t need a “data center” for the sort of small scale, and this really is an example of “purpose built, non-general”, neural network that lets these things learn to do what they do. You could fit it on a home computer, and that computer might be small enough to fit “in” the chassis of the robot. Now, making it do more than just that… might start edging into the data center thing. But, again, there exists a more sensible in between – one that can do all the complicated stuff a biped can, but is remote controlled – letting a user work, remotely, through it, in conditions that a) other designs literally can’t be as effective for, and b) are not safe for humans.
So, no, I don’t think the whole thing is silly. I do think the obsession with a lot of these people to try to make one that does everything, including complex problem solving, and would thus need to be connected to a freaking data center to do that, and badly, is just… seriously people, solve the useful problems, like getting them to actually move properly, first, then maybe rethink whether you really need the other one at all.
The original argument for humanoid robots was simple.
Our entire world is designed for humanoids.
That would be houses, cars, cities, buildings, parks, and everyplace that humans live in or visit.
This is because we are humanoids ourselves.
So, if our robots are shaped like us and move like us, they can fit in better and do what we do.
I’m not sure how well that is working out though.
I’ve yet to see a humanoid robot replacing a human in even the simplest tasks or jobs.
Robots are everywhere and do a lot.
We just don’t see them because they are doing a lot in places that few humans go.
.1. The obvious place is in our factories.
Manufacturing is more and more going over to more machines and fewer humans.
These aren’t humanoid robots but robots built to do specific tasks, i.e. form follows function.
.2. The other place is also obvious and an example of the Dark Side of humans.
War. Violence, Humans killing humans.
The Russian invasion of Ukraine and Ukraine’s fight to avoid genocide.
More and more the battlefield is dominated by aerial drones and ground robots.
None of these are humanoid either though.
Form follows function.
The aerial drones look like aerial drones.
The ground drone soldiers mostly look like minature tanks.
It’s ironic that you mention cars in that list, because during the 20th century we massively redesigned our cities for their benefit, in ways that made them much (much!) less hospitable to humans…
I saw one human with what looked like a controller in that video, it was a pretty short glimpse, but my guess from that is that they are remote controlled. Also some of the mistakes, like the lopsided weight bar lifting, wouldn’t be something you’d program.
Let’s face the harsh reality: tRUMP’s war on technology has crippled this country! I’m not a fan of the chinese government, but, they are so far ahead of the united states in ALL technological aspects (photovoltaic, electric vehicles, robotics, etc.) and they are educating their youth. This country is just drooling like imbeciles and will never catch up. I’ve always supported and used responsible technology. But, the magat fueled death spiral is sending this country to a 17th future!
OOP’s meant to wtite “sending this country into a 17th Century future” I apologize; I am so angry at the insanity of this country’s murderous plutocratic rulers I sometimes rush my keying.
I’m with some of the other commenters here. Pushing to solve hard problems will have knock-on effects down the line. That’s part of the whole mantra you typically here for the defense of basic research against ignorant politicians (the kinds who would question research on fruit flies). Even if humanoid robots don’t end up becoming mainstream, competitions like this still promote a lot of innovation.
That said, as others have pointed out, humanoid robots may have applications in the future once they can be made practical and not super expensive. They can navigate human environments and do generalized human tasks the way specialist robots can’t, which would make them ideal for assisted living situations. Plus, the form factor may simply be more attractive to consumers.
And finally, I think a huge part of this is simply entertainment, which is worth something in its own right. And it’s a lot more fun to watch bipedal robots running around a track then to watched wheeled vehicles driving around a track.
@14 Jeff L wrote: And finally, I think a huge part of this is simply entertainment, which is worth something in its own right.
I reply: I agree with that one point. But, you don’t need all this technololgy, jut buy an old ‘rock-em sock-em Robots’ toy!
Most people don’t know that the Luddites weren’t against technology itself. Their focus was only on technology that displaced or harmed people.
And, just for the sake of understanding, Saboteurs were Dutch workers who threw their wooden shoes (sabot) into machinery.
It’s hard for me to agree more, as hunanoid robots will ultimately have but one purpose – To replace humans in their capacity to serve the interests of the elite; to impoverish & disenfrancise the masses from being able to withhold their labor for economic mobility.
Doesn’t matter that it’s happening in a nominatively “socialist” nation state, these machines will only ever exist to fill that need compared to unitask robotics. They do not need to exist, & otherwise should not exist.
Job One: fire-extinguisher-use-capable humanoid robots.
raven @8 has it right. Only toys at the moment, but that’s the idea.
(Very early in the S-curve for them, but progress is inexorable)
I find your post silly PZ. The Olympics don’t need to exist either.
Robot Wars was a fun TV show. It was deliberately destructive and violent, instead of accidentally. Engineers do what scientists do, they experiment. There’s no reason this experimentation can’t be fun like the science that we don’t know where it is going is also fun.
You and I we don’t know where this is going, but our personal incredulity is personal.
Are these robots required to operate autonomously? If there is some Chinese nerd hunkered down behind walls of monitors with a joystick and keyboard, triggering series of pre-programmed routines, I’m much less impressed.
Actually, if someone was actually able to send such a complex series of motor commands from a remote location, while processing visual and other sensory signals in real time, I’d likely be MORE impressed. Either way, it’s a pretty tall order for a machine to engage in the sort of extremely complex and non-routinizeable coordinated motor actions that make up the act of walking or running on two legs. I really don’t think you’d be able to operate such a machine with just a joystick and a keyboard. (Which is one more reason why humanoid/bipedal robots are a stupid and impractical idea.)
If they’re run by AIs with massive datacenters trying to model human behavior, I’m also unimpressed…
A remote operator, human or AI, would have to be pretty close to the robots, otherwise the signal lag in both directions would make effective remote operation impossible. Remote operation of a flying, wheeled or ruddered drone is MUCH simpler than this.
Also a fan for most of the reasons stated above
Recursive Rabbit #5 Going into space was the side effect. Raining nuclear annihilation was the effect being worked on. The Space Race was a jingoistic distraction eliding the purpose of all that rocket research.
YOB @ 4.
While I get the overall point, I think that example misses the target of this post because replacing a horse with a static engine bolted to the chassis of a carriage is of a different category to replacing a horse with a quadruped robot to pull the carriage. And a horse can still take a rider places a motorised carriage will never get to – though rockhopping on a motorcycle comes close.
But otherwise history is of course littered with technological disruptions to traditional labour markets.
Silly constraints have inspired ingenious engineering breakthroughs before — off the top of my head I can think of a jumping robot that smashed height records a few years ago (30 cm robot, 30 m jump).
Why not? The rule of cool carries the day — we’re all imagining this scaled up into giant humanoid robots just like anime promised us.
I hate to put people out of work, but they might make for some impressive live-action movie stunt doubles.
@22. francesconic : “Going into space was the side effect. Raining nuclear annihilation was the effect being worked on. The Space Race was a jingoistic distraction eliding the purpose of all that rocket research.”
Not exactly a “side effect.” The space race set primary targets including being first to land people on our Moon.
There was the implied secondary purpose of “If we can land people on our Moon we can also land bigger and deadlier bombs on your country” – the very worst sort of WMDs Bombs but that was more an implied threat than a specific goal.
Both played a part especially in getting the necessary funding from – in the USoA’s case – taxpayers.
But the Going into Space bit was actually the main purpose and goal (treatment in your anaology) NOT just a “side effect.” / Cap’n Obvs?
People often prefer to have sex with stuff that is human shaped. Obviously they would want to make the robotic maid as marketable as possible.
Yes, I definitely saw a human with a controller in a group following one of the robots and a brief glimpse of the same in another shot. From that I would surmise that is the case for all. It didn’t make sense to just run them into a soft wall to finish. You can stop a remote car quite easily. Why not a remote humanoid robot?