Slashdot Mirror
iRobot CEO: Humanoid Robots Too Expensive To Be the Norm
Movie robots often look like (and are portrayed by) people in bulky, bipedal suits. Why aren't more robots built along these lines? It's not just the problem of balance. Reader concertina226 writes "'Building a robot that has legs and walks around is a very expensive proposition. Mother Nature has created many wonderful things but one thing we do have that nature doesn't is the wheel, a continuous rotating joint, and tracks, so we need to make use of inventions to make things simpler,' [iRobot CEO Colin] Angle tells IBTimes UK. 'The reason it has taken so long for the robotics industry to move forward is because people keep trying to make something that is cool but difficult to achieve, rather than trying to find solutions to actual human problems. Technology can be extremely expensive if you don't focus.'" [Beware the autoplaying video.]
Lucas beat him to this conclusion.
When Fascism comes to America, it will call itself Anti-Fascism, and tell you to give up your guns.
Depends on your end goal. If you're trying to carry iron girders, you can give it 8 legs and a long body. If you're trying to eliminate the human resistance in 2029, maybe you eat the costs of implementing human-like qualities.
She had wheels, so using her (egad, I'm anthropomorphizing!) as an example may not have been ideal.
But that said, having a robot that can utilize the same tools and work in the same environments that we do can be extremely practical, and in my opinion still well worth the effort, because that means that the same robot could potentially be repurposed for many different tasks merely by upgrading or installing different software on it... The applications for such robots extend far beyond those of mere household maintenance... the only reason that keeps coming up, is because that's just the most obvious common consumer application, and it's entirely understandable why it's desirable.
Not everyone lives in a one-floor apartment, after all.
File under 'M' for 'Manic ranting'
Tens of thousands of robots put together cars, furniture and other things every day. They don't have legs and most are bolted to a concrete floor and are little more than an arm.
The Roomba, Google's self-driving car, drones, spacecraft, the mars landers... we've made a shitload of robots that don't have legs. There's no shortage of non-legged robot research and production going on.
The CEO quoted in the article has a bug up his ass about one small area of R&D and is making idiotic excuses for why it should be eliminated. My hope is that gets in an accident and loses a leg. Maybe then he'll see the value in the R&D that's been done on robotic legs.
Idk all the obstacles to robots, but considering the story weeks ago on artificial muscles being built from fishing line and activated by heat in a way that was never really considered before for that application... I think technology can overcome this.
Tech cannot overcome everything (fundamental laws of physics) or provide quick fixes... but if nature can build a human or cat or whatever really cheap, I don't see why we can't do so artificially eventually.
I don't buy into the iRobot future - a bunch of crappy to mediocre (and very limited tools) integrated with their robot to provide a middling experience for a small subset of tasks. I can't see how we won't eventually transistion to a central unit like a walking robot using cheap or dedicated tools for the job.
It will take decades, but there's a lot of demand that some dedicated experimenters will try to supply.
I got one of his Scooba 230 floor cleaners. He has a design and manufacturing flaw in the tube for the pump. The rotary pump head sits and pinches the tube to the point of collapsing it. First thing you need to do when you get it out of the box is to remove the base plate and work out the kink. Then the bladder got a hole after about 10 uses. They sent a new one, but damn they need to work on quality control.
I guess all the money goes into the robotics research and nothing goes into manufacturing. Smart brains, cheap parts.
... Daleks really are the pinnacle of evolution.
Who says it was cheap?
Look at how many hundreds of millions of years it took.
Now equate that time to capital investment....
Still think nature did it cheaply?
File under 'M' for 'Manic ranting'
a couple of applications. Like driving a car or playing an organ. In a factory or office environment the requirement is usually to be able to move from here to there on a smooth (possibly carpeted) environment using a moderately small footprint, and the manipulate things at either a standing or sitting height. I'd have to think about opening doors though. That requires a bit of a reach if the door opens toward you. Feet are easier for this, doing it with a tracked or wheeled platform might require a bit more reach.
When you want to send a robot into an area designed for humans (like Fukushima power plant), having it able to go through hatches and doors, and climb ladders designed for humans, etc.
Sure iRobot has their cute little mini-tank PackBot, but it has real problems in a real disaster environment. It's pretty easy for things to form a "tank trap" from which the PackBot cannot climb.
Many would debate whether the vacuum robots we can afford are worth it:)
But for projects like this one, looking humanoid is the only goal.
Get free satoshi (Bitcoin) and Dogecoins
The per unit costs on kittens is pretty low.
The capital investment was recouped at each stage over the hundreds of millions of years per evolution.
Yes, cheap.
This is a very good point, and for robots designed for a single task that obviously makes sense. But if they have to be able to move around a house or office (with either stairs or an elevator with buttons to push), or open doors, or put dishes away from the dishwasher, etc -- they'll need to be shaped roughly like a human. The more human-shaped they are the more easily they can integrate into a world designed for human-shaped things to get things done. The alternative is to redesign everything in the world to make LESS convenient for people to use them.
[nt]
File under 'M' for 'Manic ranting'
Take a look at what happens inside a cell. Plenty of rotating joints and tracks. Ribosomes and flagella anyone? Wheels, I grant that. But the reason is probably that at the molecular scale they make no sense
Mostly random stuff.
With over 8 billion people in a few years, and fewer jobs, the obvious solution is humans mounted with something along the lines of Google Glass, telling them what to do, where to go, how to do it, when to speak, etc.
You're hired, they give you Glass, the computer tells you what to do, "go put more toilet paper in the bathroom" "clean up the parking lot" etc, humans are cheap and disposable because there are so many of them.
"If any question why we died, Tell them because our fathers lied."
If it becomes technically possible to build a fully functioning humanoid robot, regardless of the price, then one will be built. Once this happens, Moore's law will start to kick in, as will the cost benefits of mass production. In fact all you need to do is to build a self-replicating robot, and call it skynet.
"While theoretically and technically television may be feasible, commercially and financially it is an impossibility." -- Lee DeForest, inventor.
It's not too expensive, it's just ahead of its time. The first microwave, VCR, PC... were all super expensive at the start. Cost scales to volume, if your only making one at a time, then yeah it's expensive. If your building them by the millions, then their cheap. The problem is... (Drumroll)... -NO KILLER APP- ( no pun intended) lol.
We do not design our living and working spaces for robots, but for humans. Making robots conform to our current environment makes sense from that viewpoint. Quadrupeds, wheeled ones like Rosie from the Jetsons, tracked ones like Johnny 5, etc are designed for things other than human living environments really. Rosie is close but looking at her, she should have been stymied by a 2 inch step without some kind of equipment around to lift her up it.
So yes, there is a very good reason to design humanoid robots. Its easier than redesigning all our living spaces.
If I might posit this however, perhaps it would be a good thing to have places inaccessible to robots. When the robot uprising comes, that 2 inch step may be the only thing that saves your life. Down with legs, long live humans!
A lot of comments mention that it would make sense to make a robot along the same pattern as a human: Can use same tools, access same spaces, etc. etc. My question is: if your {AI | robot} can't be distinguished from a real human, why can't you just use a (cheap, ubiquitous) human? Answer: we invent machines precisely to augment our abilities, to do what we aren't so good at: computing faster and less error-prone, be stronger, access spaces we can't, don't get bored, tired, damaged by some harsh environments, etc. etc.
I'm with Angle: see what job (or collection of jobs) your machine needs to perform, then build the best possible machine given the contraints, for that job/s. (Not that they have achieved it with the floor suckers, but hey...)
Free, as in your money being freed from the confines of your account.
You are forgetting that evolution has no goal. You cannot possible equate the two things.
It can't be that difficult to make a humanoid robot that can balance on two legs, walk, run, pick up things, etc. just like a human does. Notice how long it takes a baby to learn to walk. Obviously it isn't easy when you've never done it before, I'm sure a robot can learn it in exactly the same way a human baby does, through trial and error. The great news is that once the robot has learnt how to walk, no future robots need to LEARN, they just use the same method that the walking robot uses.
I don't think there is an intelligent, focused process at work behind evolution.
I predict that within 100 years, robots will be twice as powerful, 10,000 times larger, and so expensive that only the five richest kings of Europe will own them.
lose != loose
Incredible!
I wasn't suggesting that it did... only that it took an awful long time for evolution to do it... and equating that amount of time to how much you'd have to pay a worker even just one lousy dollar a day for that amount of time, it's really not very cheap.
File under 'M' for 'Manic ranting'
people will be printing their own robots in 50 years or less in a local (or home) printing shack.
Maybe I'm reading into it a bit, but I doubt the guy is so obtuse that he doesn't realize there's enough money to go around for the various forms of locomotion. I think this is just some defensive posturing he's doing in public to try and paint his company's products in a better light against the soon-to-be competition.
Here's what I see:
1) iRobot is a major supplier of defense and security robots currently in use by the US military.
2) iRobot's entire lineup is based on wheeled or treaded robots. There's no indications of them being anywhere close to fielding a walking robot of any sort.
3) Meanwhile, Boston Dynamics, a small company that wasn't yet a credible threat, has been working on both bipedal and quadrupedal robots for DARPA that are to the point where they're being field tested by the military.
4) Then, Google bought Boston Dynamics, meaning it suddenly has far more resources available to it than before, making them a much more credible threat.
5) And now, shortly thereafter, iRobot's CEO suddenly comes out trashing the technology used by the competition, just as that technology is reaching a point where it can start entering the market.
As I said, I might be reading into it a bit, but the timing and notions just seem weird. For instance, going back to the summary (emphasis mine):
The reason it has taken so long for the robotics industry to move forward is because people keep trying to make something that is cool but difficult to achieve, rather than trying to find solutions to actual human problems.
This is pretty clearly posturing on his part, since he has to be aware that none of his Roomba products can navigate stairs, an extremely basic and common component of building interiors. It's obvious that his products are not offering "solutions to actual human problems", or at least not to all of the problems, and he's scared that others will realize it too. It's good that he is, since his company isn't set up to deal with it, from what we know publicly.
Your stupid roomba is half the reason to build something with legs. PS. boston dynamics was doing great things with legs. Sure it's expensive, but all things in R&D are. iRobothttp://hardware.slashdot.org/story/14/03/09/1327215/irobot-ceo-humanoid-robots-too-expensive-to-be-the-norm# is just regretting that they should have bought Boston Dynamics!
The original Daleks couldn't go up stairs, so they'd be useless in my place. But they do have a plunger arm, which can be occasionally useful.
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
I assume this means 3D printers can now print out complete, functional items from thin air, sunbeams and rainfall.
So, what does expensive mean in our glorious post-scarcity, post-industrial, post-engineering (Hey! Just download and print!) society?
Look, the other day I watched a backhoe, barely bigger than a vending machine digging the smallest section of drainage ditch. I'm tempted to just rattle off a bunch of buzzwords and say synergy and 3D printing etc sixteen times. But the increasing complexity, intelligence and sophistication of computing power, software, sensors and things like servomotors is growing at a... Exponential? (Geometric?) INCREDIBLE rate and at some point, sooner than even I think, ten, fifteen years at the max, humanoid robots will be so cheap that they will, in fact be cheaper than actual humans. At that point the problem isn't how to develop and employ them, it's what to do with the 90% of humanity that can't do any job cheaper than a robot. The only jobs left right now are. 1: Guy that does a job robot cannot do. Doctor, Lawyer, Scientist, politician. 2: Robot trainer 3: Robot Repairman (See #1) 4: Guy who does a job cheaper than a robot, or a job a robot isn't willing to do. (These jobs suck ass.) AI, like Watson will continue to shrink categories, 1 and 2, and eventually category 3 will be taken over by robots. This is the real robot apocalypse, not murderous killbots, but they will have "TOOK R JOBS" and only folks who own the robots will have any realistic way to make money. The good news, for me, is that I'm kind of old, so it won't crush me. The bad news is that you probably aren't. We need to figure this out, now.
General humanoid robots in mass production are going to happen; it's just a matter of when and there will be ridiculous prices for the first ones that work. This should happen in less than 40 years and I'm hoping much sooner. The iRobot CEO seems to think specialized robots are going to be the norm.
"Mother Nature has created many wonderful things but one thing we do have that nature doesn't is the wheel, a continuous rotating joint, and tracks"
Very recently though there was a finding that some insects in the early stage have wheels in their rear legs to hop. He surely wasn't informed such wonderful finding.
Until somebody comes up with the über power source, all of this stuff is academic. Sure, I can build the Aliens Power Loader but it has to be connected to a big ass generator to work.
what this slogan is trying to capture is elegance of design, which the human body has
in terms of per-unit costs
http://ifaq.wap.org/computers/famousquotes.html
"Computers in the future may weigh no more than 1.5 tons."
- Popular Mechanics, 1949
"I think there is a world market for maybe five computers."
- Thomas Watson, chairman of IBM, 1943
"I have travelled the length and breadth of this country and talked with the best people, and I can assure you that data processings is a fad that won't last out the year."
- The editor in charge of business books for Prentice-Hall, 1957
"But what...is it good for?"
- Engineer at the Advanced Computing Systems Division of IBM, 1968, commenting on the microchip
"There is no reason anyone would want a computer in their home."
- Ken Olson, president, chairman and founder of DEC
Researchers have been making humanoid robots for much longer than they've been trying to make any of those other things you listed. And yet, such devices are still limited to doing simple tricks of little or no real value. In the mean time, robots designed for specific purposes (that look nothing like people) are used throughout society. Humanoid robots will always be much more complex, and much less stable, than their non-humanoid counterparts. So of course they will never be affordable because you will always be able to make a cheaper wheeled robot.
Also, it is baffling to me that anyone would throw away money on this line of research. The limits of this kind of robot should be obvious to all of us, since it would have all the same limits we do. But for some reason this idea is so compelling to the less logically minded masses that it attracts all kinds of money for research that is destined to lead nowhere.
Now it is difficult to do those kind of robots, because we didn't have the technology, but with trying to understand this and research into this we will be able to do it in a couple of years and 20 years on we won't even think about how normal and easy it is to build robots like that.. Not only humanoid robots benifit from this research, but also artificial limb-makers will benefit from it.. And people who lost their limbs will be gratefull to those scientists who went on to make these kind of robots and not those wheelie things iRobot is making..
Let's not forget iRobot is a commercial company with a lot at stake, so they're saying this for a reason...