In a different thread there was an argument about the relative futility of some of today's robotics experiments [0]. As it is sometimes typical on HN, the poster with a message outside the "go team" cargo-cult mentality wasn't very welcome. I happen to think he had a very interesting point. And that point translates well into this thread.
Why don't we see practical applications of all of these robotics experiments? The answer is very simple really: Most of them are relatively pointless and add very little to the robotics knowledge-base that will be needed to really move robotics forward into real-world applied robots.
Think of something like robotics vacuum cleaners. Nothing whatsoever innovative about any of them. It's a wheeled platform that has been in use in hobby and research robotics since, well, forever. The '70's and '80's were full of robots with this basic platform. What changed? Electronics got better, batteries smaller, microprocessors more capable, manufacturing more efficient. What was retained and reused from prior research? Probably not much.
I started in college with the goal of becoming a robotics engineer. An EE with specialization in robotics. It didn't take long for me to realize that the field wasn't as interesting and exciting as I made it out to be in my mind. The R2D2's and C3-PO's were nowhere to be found and were easily decades away from becoming reality. If I wanted to be in robotics I would end-up making industrial manipulators or things with motors that we would all pretend were robots. That's a pet peeve of mine. Battlebots had nothing whatsoever to do with robotics. It was about a bunch of remote controlled machines. Not robots.
I digress. The point is that I was really excited about the field until I realized what I wanted to do would have to wait 50 or 100 years. I wanted to work on Commander Data, not a mindless pick-and-place machine.
And so I begun to dissect things and think about what it would take to get there. Do we learn anything by making humanoid-looking little robots out of RC servos? I built a couple. It's an utter waste of time. Nothing whatsoever of value other than to pretend we built a humanoid. Don't get me wrong, it's a great hobby and lots of fun for the kids to learn, but it is far, far away from anything even remotely useful.
In my opinion these are the areas that need a quantum leap in development before robots like Wildcat can become useful and relevant outside the lab:
ARTIFICIAL MUSCLES
This is huge. Motors, gears, springs, pistons and bladders just don't cut it. We need a step change in the performance and capabilities of what we use to do the job of biological muscles. Machines like Wildcat can't operate for days at a time. They use internal combustion engines to power pumps and hydraulic or pneumatic end-effectors to actuate joints. This is lousy. Very little can be learned from trying to operate such machines. You end-up with things like Asimo that walk like they are taking a dump because it is nearly impossible to implement true dynamic gaits because we either can't implement enough degrees of freedom or joint actuation simply isn't up to par.
Artificial muscles that perform well and are energy efficient would revolutionize the field.
ENERGY STORAGE
Thankfully this is something robotics shares with electric cars. We need to do much better than current LiPo cells allow in terms of volumetric power density (at the very least).
ARTIFICIAL INTELLIGENCE + CONTROL SYSTEMS
This is a field that has seen advances but is nowhere near where it needs to be. I can teach a five year old kid how to sort and fold a pile of clothes without much effort (other than maintaining his or her attention). It would be very hard to do the same with the AI we have mastered to date. I am talking about having a couple of robot arms and a camera presented with a random pile of clothes and having those clothes sorted and properly folded as a human would. No special mechanics, suction mechanisms or anything like that.
PROCESSING / NEURAL COMPUTING
The AI+CONTROL field ultimately needs far more advanced and energy efficient processing architectures than are commonly available today. Stuffing a robot with a powerful Linux PC provides nowhere near the processing bandwidth needed to perform at a level comparable to a human child. I am not sure what form this step improvement in computing will take, but we need it.
PROGRAMMING LANGUAGES / DEVELOPMENT AND SIMULATION TOOLS
We are in the dark ages. We need a serious paradigm shift in the way we program computers if we are ever going to even approach something that can compare to the fictional C3-PO or Commander Data ideas.
If you want to contribute to robotics your time and efforts would be far better spent on the above (I am sure there are other areas I have not listed) rather than making little remote-controlled gyro-actuated cubes that link to each other via magnets. I don't know what can be learned from that other than making remote-controlled gyro-actuated little cubes that link together via magnets. Cool toy. Useless for the advancement of robotics. It's almost like spending a lot of time playing chess: You become better at playing chess, a narrow skill, and virtually nothing you do can be translated or reused for other tasks outside of chess. Grandmasters are not genius thinkers, they are simply great chess players and that's it. Master little cubes with gyros and that's all you've mastered.
I have two German Shepherd dogs. I have trained both of them to search for objects I hide anywhere in the house. I show them the object, I let them smell it and then hide the object while they wait in a "sit-stay" well out of sight. Sometimes I'll hide the object deep in a drawer inside a closet in an upstairs bedroom while they wait in the garage with the door closed. These dogs are amazing to watch. They always find what I showed them. Every so often they need a little help (and they ask for it), most of the time they do it on their own. Think about all that is required for an animal to do this spanning a range of capabilities from cognition, perception, sensing, navigation, planning, communications and more.
There is no way a bunch of little blocks or a gasoline-hydraulic-powered machine is helping us advance towards even something as simple, in terms of biological beings, as finding an hidden object using smell. A better place to spend money and resources is in the areas I highlighted above and others I did not mention. Once you "ace" the above, the process of designing and fabricating a mechanical frame with the required capabilities should be an almost academic exercise for any engineer with a moderate range of experience in the electromechanical fields.
Not to minimize Boston Dynamic, but I really think a lot of what they and others are doing is simply burning tax money for no good reason. Well, there is a good reason. The government folks who shovel out the money are easily impressed by this stuff. Nothing really advances but it is impressive as hell. Who knows how much money was burned on the GE walker in the 1960's [1]. I don't know of anything that came out of that project and is in use today. If I gave any reasonably capable team of engineers a few million dollars to play --without a requirement to actually deliver something that works in the real world-- they could build similarly capable machines. There's nothing special about these systems other than they are impressive to the untrained eye.
General Electric built quadrupeds in 1968 [1]. The only reason they didn't perform like the Boston Dynamics rigs is that they did not have access to better computing platforms, sensors and electronics. There is nothing in the Boston Dynamics machines in terms of mechanics or hydraulics that was not available or could not be implemented in 1968. Just look at the video [2] (got to love the sound effects). This machine, all by itself, proves my point about the futility of some of this research. They all put the cart in front of the horse. The GE machine needed better effectors, sensors, energy storage, AI and control. The machine shows the amazing mechanical complexity that was attainable in 1968. Remember, no Solidworks, no microprocessors, no FPGA's, no Linux, just a dude pulling levers. Amazing stuff.
We are simply focusing on and throwing money at the wrong things.
You're comparing the GE quad with the BD quads. The video you link to mentions that the mental load of constantly thinking about four feet means the operator of the GE quad has to stop after about fifteen minutes.
Now the BD quads can run around untethered, with a human operator who is presumably doing little more than selecting gaits on a remote and steering it. If I make the least charitable reading of the Cheetah video at hand, it's still a damn long way from being useful, but it's closer to being useful than that giant truck-beast - imagine this a few iterations down the line, married with the results of self-driving cars mature enough to go offroad, for instance.
There are people working on all of the things you mention. And there are people playing with what can be done now. Your attitude seems to be "if I can't work on near-human robots, I'm not interested", which is fine, but do you really need to shit all over everyone else who's decided to see how far they can get with what they've got right now? It took evolution about 1800 million years to get from single-celled life to humans. We've only been trying to build human-mimicking automatons for a few hundred years. Have some patience. Stuff like the folding cubes robot? That's the pre-Cambrian explosion of wild forms. Which ones are practical? We won't know until we've tried 'em.
We'll get there eventually. Maybe not in my lifetime or yours, but we'll get there.
Perhaps I failed to communicate my position clearly enough. Speaking of something like Wildcat specifically, I don't see the point in repeated mechanical engineering exercises that are expensive and do little to move the art forward. To me that is a waste of money. Take the GE platform and evolve it. Get rid of the human controls and integrate a modern compute engine. Add sensors. Now you have a platform for significant work in areas that need it. Show me this platform navigating a Home Depot while attending to the various products they sell. Show me the platform searching a pile of rubble for survivors. Examples abound. Few, if any, require expensive mechanical design exercises to reinvent the wheel. In fact, you could probably do a ton of good development with a good size quadruped constructed out of large RC type servos and linkages. All of the computing would have to be external, but that's OK. Build one of those that can do what one of my dogs does (use smell to find objects) in a suitably complex environment.
My problem is with the idea of throwing tons of money in the wrong direction. We don't need to design quadruped after quadruped that perform narrow parlor tricks. We need better brains, sensors, actuators, materials and a myriad of other area improvements that will make a true difference.
Isn't that… kinda what BD is doing? I'm sure they could give you a list of prior work they're building on as long as your arm. Their main current work seems to me to be in building self-stabilizing walking platforms, upon which they can eventually carry various other loads, such as, oh, the dog-level AI/sensor package you're using as an example.
I dunno, I guess I kinda want to take on the role of the Voice of Startup Culture here and say "congratulations, you've identified a hole in the market, start working on your MVP". Except I also kinda feel like that's exactly what all the modern work you disparage is doing; the MVP of "robots" is a huge, multifaceted problem, and nobody's made it out of the "cool tech demo" level yet.
You've made a list of a half-dozen fields that need to be advanced. Which one are you working in? Either as your day job, or your hobby? Or are you just sitting on the sidelines going "it's all been done before"?
> You've made a list of a half-dozen fields that need to be advanced. Which one are you working in? Either as your day job, or your hobby? Or are you just sitting on the sidelines going "it's all been done before"?
Let's be careful not to engage in shooting the messenger, which does nothing towards addressing the validity of the argument.
That said, I am playing in a couple of areas, for example, applying AI to learning. The huge difference is that I am doing it in the context of a private enterprise. I am not asking taxpayers to shovel money at my toy projects. There are companies out there that exist solely because of hundreds of millions of dollars of tax money being thrown at projects that produce very little of real value.
What would you say if DARPA threw tens or hundreds of millions of dollars at a company that set out to rebuild an operating system virtually identical to Linux? I would hope you'd come forward and say that we need to work on other things. This has been done, multiple times, and does not need to be reinvented until advances in other technologies warrant it.
If you read what I said carefully you should see that my primary argument is that we are throwing money at the wrong problems. Tax money.
If you have a private company and want to iterate through a dozen expensive mechanical design exercises for whatever reason and YOU fund it. Fine. Your money. Your decision. You get to do whatever you want.
My problem is with tax money that is repeatedly thrown at the wrong problems. We don't need to push for advances in mechanical design. We need to push for advances in the areas I mentioned and more. Sensors, for example.
It's open source. You can download the entire design and build one yourself. I have it on my machine. I can open every single part in Solidworks, look at it and modify it.
Tell me what this platform is lacking and why we have to spend so much money reinventing the wheel? Hundreds of millions of dollars.
We and others have already thrown lots of time and money into the mechanical design of a myriad of platforms that are fantastic for the development of the really tough areas that need attention. We somehow choose not to build on top of that but rather throw money at doing the same thing over and over again and producing little usable tech out of it.
Do you realize that Boston Dynamics got nearly TWO HUNDRED MILLION DOLLARS of US government money [0]? For what? To build otherwise useless platforms that make the government guys go "ooh" and "ahh" because they really don't know what they are looking at. It's magic! Let's throw more money at it. While, in the meantime, the money should have gone into some of the areas I highlighted and others.
This is not the best analogy, but I'll try. This would be similar to throwing $200 million dollars to repeatedly build MRI machines that do little towards addressing the needs of, say, Cancer research. In reality you need to throw the money at specific areas of Cancer research (not a biologist, so I can't list them) rather than at making MRI machines. We know how to make MRI machines. That's a done deal. What we need to improve is everything else.
We know how to build robots like these. These are academic exercises. Take a small team of engineers covering the required fields and they'll build you these robots. It's an exercise in IMPLEMENTATION. We know how to do it. It's just a matter of doing the work.
How many HN members could build a little galloping robot in, say, five years, if I threw five or ten million dollars at them and there was not requirement to build something that was commercially viable in any way. Lots. Same results as BD with 5% --or less-- of the money BD got. That's a sign that this requires no more R&D money until we improve everything else.
What we need to invest money on are the things we don't know how to do well enough. They are not very sexy. I get it. Who wants to be the guy to throw money at AI for ten years and have to explain to a General why all you got are some graphs and numbers on a screen. Significant as the results might be, it's a lot easier and more exiting to show a bunch of ignorant gate-keepers a ridiculous machine galloping along or bouncing around. It's a sexy supermodel in a bikini. Very soon everyone forgot what the hell they were there for and throw more money at it so they can see another one.
> ARTIFICIAL INTELLIGENCE + CONTROL SYSTEMS
>...
> PROCESSING / NEURAL COMPUTING
Huh? Isn't this exactly what they are working on here? I guess you must think that all the control/intelligence advances will come from places outside projects like this, but that's not the clear to me at all. Certainly, you'd need to justify it for any of your criticism to hold. This seems like exactly the kind of project that DARPA could fund to advance these fields.
> ENERGY STORAGE
Nope. These are all petrol powered and will be for the foreseeable futures. We don't fuel tanks or Humvees with electricity either.
> Isn't this exactly what they are working on here?
These things are not autonomous. They are overgrown remote controlled devices. Perhaps you are under the impression they move about in the same fashion as Google's self-driving car?
Now there's a good example. Google's self-driving car is advancing the art and it is doing so with the goal of delivering a product millions of people can use. What are they focusing on? A remotely controlled car? Nope. No, they are focusing on sensing, algorithms, computing and all of the core items I discussed in my prior post. Their work is useful. Little magnetic cubes or pneumatic jumping machines aren't even in the same orbit. do you realize that there is virtually no difference in the way the Boston Dynamics machines move and perform when compared to what MIT and others have been doing for decades? Sure, they can pull off better tricks because computers are better now and that's about it.
Don't get me wrong. I love them. Hell, I wish I had the pull to extract that kind of money from the government to play with technology like that. I'd be fun. In the real world you'd implode inside of a year (or a few) if you didn't advance past that stage quickly.
I think you're introducing a false dichotomy. These things are supposed to be Avatars first and foremost.
In the past, a lot of research focused on AI (reasoning). The current trend is to instead focus on building the "hardware acceleration" that humans enjoy and don't have to think about, such as walking, running, collision avoidance, etc. It's a necessary requirement for a true "robot" in your sense of the word - if AI is developed, it can then be plugged in instead of a human operator.
In the spectrum of what robotics could do for society I see a wide range of truly paradigm shifting applications. These are applications that can change the lives of millions of people. They all require massive improvements in the areas I mentioned in my post. The mechanics are almost academic as is evidenced by the GE walker from 1965 [0]. In terms of mechanical engineering we could do an insanely complex machine back in 1965.
I see no point in creating something as clumsy as Wildcat when what we need are significant improvements in AI. Think of what would be required for a quadruped to be able to climb a rocky terrain like a goat. We can do the mechanics. We could do the mechanics in 1965. What we can't do yet is the reasoning, control and path planning required to complete such a task. We don't need to continue to build a bunch of useless contraptions to advance these fields. We are throwing money at the wrong problems.
I was tackling a subset of your argument initially. If I understand you correctly, what you're saying is that we need to have the core technology first before trying to piece together something out of it first, and any research without that core technology is useless.
I'll expand point by point.
ARTIFICIAL MUSCLES. I wholeheartedly agree that this is a huge point. However, to draw an analogy to civil engineering - cement is important (and is an enabling technology), but it's not required to make advances in architecture.
ENERGY STORAGE. Suspend it by wires for now, limited-mobility robots are useful in some industries.
ARTIFICIAL INTELLIGENCE + CONTROL SYSTEMS. This was the core of my previous argument. A lot of what someone perceives as "hard AI" problems are actually "hardware-assisted, AI-guided" problems (your example of control and path planning come to mind). In your video, it showed the operator not being able to move the robot fluidly & fluently. This is the aspect that this DARPA initiative is tackling. Separate the AI (human operator for now) from the hardware-assisted aspects. (Please note that humans can typically take over the "hardware-assist" and manually move, but try writing with your non-dominant hand or run in reverse)
PROCESSING / NEURAL COMPUTING. See above.
PROGRAMMING LANGUAGES / DEVELOPMENT AND SIMULATION TOOLS. Sure, I agree - but this is a problem that can be tackled right now.
I agree that there are a lot of areas that need to be improved for a "true robot". However, we can improve some subsets right now, and they may be useful in the future.
It's somewhat like arguing that developing advanced image-processing concepts & algorithms was pointless 50 years ago, because neither the software nor the hardware of the time could do anything with the idea. It's quite common in this field that an old algorithm becomes practical due to hardware and software capabilities.
I have you heard of LS3 [0]? It's another DARPA project designed to bring autonomy to platforms like this. The controls work Boston Dynamics doing is very innovative.
Google is doing great work, but it's much more on the refinement and engineering side. These guys all worked on the DARPA grand/urban challenge and putting in time to make the base level tech robust.
Current batteries appear to be inadequate because walking robots are so inefficient. Mine took 3000 watts to walk. ASIMO is similar. But humans require only 200 watts to run. Delivering 200 watts for 12 hours (Ironman-level endurance) takes 25 lbs of lithium batteries, which is the same ballpark as the weight of the human cardiovascular + digestive system.
By "Ironman" can I assume that you mean the triathlon, not the marvel character? Personally, I'd really like to see the storage system that could power a practical untethered powered exoskeleton.
Right, and that's why the first item on my list are artificial muscles. If we had efficient, pliable, moldable, rugged, strong artificial muscles robotics would make a step change for the better overnight.
I'm sure there's significant opportunity to increase efficiency by combined improvement in the structure and control software.
Have you tried to learn a balance sport like skateboarding, snowboarding, skiing, etc? As a beginner you aren't capable of the prediction/anticipation necessary to position yourself precisely for dynamic stability. You overcome this largely by bracing multiple muscles against each other.
But once you practice enough that your brain understands and anticipates, everything suddenly feels much more effortless. A beginning snowboarder will burn through to muscle fatigue limits in just a couple hours on the bunny hill. An experience snowboarder can ride the steeps all day with only mild fatigue. Sure, there's likely to be a difference in physical conditioning too, but in my experience that's a lessor factor than simply knowing/feeling how to move the minimum amount at the precisely right moment.
So while a revolutionary advance in artificial muscles would be grand, I think there still may be significant opportunities with existing actuators.
I think there's plenty of room for more robotics in the real world today. No C3-PO yet (which was technically a walking google translate), but even something like this robot making food (https://www.youtube.com/watch?v=nv7VUqPE8AE) could be useful in the real world. There's no reason with the progress I've seen in OpenCV and point cloud tech that we couldn't make robots that could stock store shelves, serve food, lay tile, etc. The largest impediments I've seen would be the insane cost (often over $100k), and how slow they are (maybe because of processor speed?).
Universal robots has a robot that can lift 10 Kg; on par with a person.
The machine in the video you linked to is neat but it actually serves to illustrate my point. Move the ingredient dispensers a few centimeters in any direction, change the cooking plate temperature and watch what happens. This machine is simply playing back a programmed set of motions. The fact that it looks complex and anthropomorphic does not make it intelligent. Any FIRST Robotics League team in the US could build a machine to do this for a lot less money.
When I had my own electronics manufacturing operations in house I always looked for opportunities to add robotics to improve the process. In almost every instance it was far more efficient to add a custom or semi-custom single-purpose tool than to even attempt to use a programmed manipulator of any kind. The combination of a person with a set of custom tools can be very efficient. Now, of course, I never did mass production. That's an entirely different game. Even then, you'll see programmed machines do tasks like dispense adhesives or gasket material. In nearly all cases an indexing fixture carefully aligns the workpiece and the robot/machine executes a canned motion cycle. These machines are not intelligent at all. They might appear to be, but they are not.
I mostly agree with you, except with that "it's not worth working on these things, they add no value, ...". American government understands, that the nation, that creates automated military first, will gain the kind of advantage Germans had over Poland when they drove tanks against horses. Besides I think their work is very much valuable in proof-of-concept way, and shines light on problems that arises going this specific way. Without people working on these seemingly useless things we might never know what's realistically possible to do and what's not.
Why do we have to work on better ways to kill each other? When will that end? Is that the only motivation we have?
Really.
I don't want to work on robotics to kill people. I don't want my government to throw money into robotics to kill people. I stand by what I said: These things add no value. In fact, from one perspective they take value away. And, BTW, I am not even close to being a bleeding heart Liberal yet I'll gladly join them in the idea that what we need is less violence, less killing and less weapons, not more.
>Why do we have to work on better ways to kill each other? When will that end? Is that the only motivation we have?
This technology in and of itself need not be inherently evil, any more than (as I mentioned in another thread) the Jeep, or rocketry, or radar or the computer or any number of technologies initially developed for military applications.
This is in essence just another way to get a machine from point A to point B. It has legs instead of wheels. I would disagree that they add no value, if one separates the engineering from the one possible application out of many. If the military wants a fully autonomous killing machine once they figure out the "fully autonomous" part (which has, really, more to do with software than hardware) then the hardware part of it - whether it has two legs, or four, or wheels or treads or flies, isn't really going to matter.
If you don't make it, others will. Creating weapons is the strongest innovation driving force, if there would be no violence we would still be picking up berries.
Thanks for articulating so eloquently what I failed to communicate in [0]. Great post. Moreover, I feel we may have been horribly misled by Asimov's vision of robots. Come to think about it what we really need is not machines to be our slaves and fight our wars, but rather more intelligent animals. Perhaps biotechnology can deliver some kind of clever ape that can be useful around the house and do some fighting or carry ammunition in battlefield too. Yeah, I know unethical and all that. All these qualms will die instantly the moment a country is able to field a conventional army boosted with such apes. US has already done that as a matter of fact using mine-carrying seals / dolphins in WWII. And I know for a fact that a NATO army has plans to use pigs to clear minefields.
Why don't we see practical applications of all of these robotics experiments? The answer is very simple really: Most of them are relatively pointless and add very little to the robotics knowledge-base that will be needed to really move robotics forward into real-world applied robots.
Think of something like robotics vacuum cleaners. Nothing whatsoever innovative about any of them. It's a wheeled platform that has been in use in hobby and research robotics since, well, forever. The '70's and '80's were full of robots with this basic platform. What changed? Electronics got better, batteries smaller, microprocessors more capable, manufacturing more efficient. What was retained and reused from prior research? Probably not much.
I started in college with the goal of becoming a robotics engineer. An EE with specialization in robotics. It didn't take long for me to realize that the field wasn't as interesting and exciting as I made it out to be in my mind. The R2D2's and C3-PO's were nowhere to be found and were easily decades away from becoming reality. If I wanted to be in robotics I would end-up making industrial manipulators or things with motors that we would all pretend were robots. That's a pet peeve of mine. Battlebots had nothing whatsoever to do with robotics. It was about a bunch of remote controlled machines. Not robots.
I digress. The point is that I was really excited about the field until I realized what I wanted to do would have to wait 50 or 100 years. I wanted to work on Commander Data, not a mindless pick-and-place machine.
And so I begun to dissect things and think about what it would take to get there. Do we learn anything by making humanoid-looking little robots out of RC servos? I built a couple. It's an utter waste of time. Nothing whatsoever of value other than to pretend we built a humanoid. Don't get me wrong, it's a great hobby and lots of fun for the kids to learn, but it is far, far away from anything even remotely useful.
In my opinion these are the areas that need a quantum leap in development before robots like Wildcat can become useful and relevant outside the lab:
ARTIFICIAL MUSCLES
This is huge. Motors, gears, springs, pistons and bladders just don't cut it. We need a step change in the performance and capabilities of what we use to do the job of biological muscles. Machines like Wildcat can't operate for days at a time. They use internal combustion engines to power pumps and hydraulic or pneumatic end-effectors to actuate joints. This is lousy. Very little can be learned from trying to operate such machines. You end-up with things like Asimo that walk like they are taking a dump because it is nearly impossible to implement true dynamic gaits because we either can't implement enough degrees of freedom or joint actuation simply isn't up to par.
Artificial muscles that perform well and are energy efficient would revolutionize the field.
ENERGY STORAGE
Thankfully this is something robotics shares with electric cars. We need to do much better than current LiPo cells allow in terms of volumetric power density (at the very least).
ARTIFICIAL INTELLIGENCE + CONTROL SYSTEMS
This is a field that has seen advances but is nowhere near where it needs to be. I can teach a five year old kid how to sort and fold a pile of clothes without much effort (other than maintaining his or her attention). It would be very hard to do the same with the AI we have mastered to date. I am talking about having a couple of robot arms and a camera presented with a random pile of clothes and having those clothes sorted and properly folded as a human would. No special mechanics, suction mechanisms or anything like that.
PROCESSING / NEURAL COMPUTING
The AI+CONTROL field ultimately needs far more advanced and energy efficient processing architectures than are commonly available today. Stuffing a robot with a powerful Linux PC provides nowhere near the processing bandwidth needed to perform at a level comparable to a human child. I am not sure what form this step improvement in computing will take, but we need it.
PROGRAMMING LANGUAGES / DEVELOPMENT AND SIMULATION TOOLS
We are in the dark ages. We need a serious paradigm shift in the way we program computers if we are ever going to even approach something that can compare to the fictional C3-PO or Commander Data ideas.
If you want to contribute to robotics your time and efforts would be far better spent on the above (I am sure there are other areas I have not listed) rather than making little remote-controlled gyro-actuated cubes that link to each other via magnets. I don't know what can be learned from that other than making remote-controlled gyro-actuated little cubes that link together via magnets. Cool toy. Useless for the advancement of robotics. It's almost like spending a lot of time playing chess: You become better at playing chess, a narrow skill, and virtually nothing you do can be translated or reused for other tasks outside of chess. Grandmasters are not genius thinkers, they are simply great chess players and that's it. Master little cubes with gyros and that's all you've mastered.
I have two German Shepherd dogs. I have trained both of them to search for objects I hide anywhere in the house. I show them the object, I let them smell it and then hide the object while they wait in a "sit-stay" well out of sight. Sometimes I'll hide the object deep in a drawer inside a closet in an upstairs bedroom while they wait in the garage with the door closed. These dogs are amazing to watch. They always find what I showed them. Every so often they need a little help (and they ask for it), most of the time they do it on their own. Think about all that is required for an animal to do this spanning a range of capabilities from cognition, perception, sensing, navigation, planning, communications and more.
There is no way a bunch of little blocks or a gasoline-hydraulic-powered machine is helping us advance towards even something as simple, in terms of biological beings, as finding an hidden object using smell. A better place to spend money and resources is in the areas I highlighted above and others I did not mention. Once you "ace" the above, the process of designing and fabricating a mechanical frame with the required capabilities should be an almost academic exercise for any engineer with a moderate range of experience in the electromechanical fields.
Not to minimize Boston Dynamic, but I really think a lot of what they and others are doing is simply burning tax money for no good reason. Well, there is a good reason. The government folks who shovel out the money are easily impressed by this stuff. Nothing really advances but it is impressive as hell. Who knows how much money was burned on the GE walker in the 1960's [1]. I don't know of anything that came out of that project and is in use today. If I gave any reasonably capable team of engineers a few million dollars to play --without a requirement to actually deliver something that works in the real world-- they could build similarly capable machines. There's nothing special about these systems other than they are impressive to the untrained eye.
General Electric built quadrupeds in 1968 [1]. The only reason they didn't perform like the Boston Dynamics rigs is that they did not have access to better computing platforms, sensors and electronics. There is nothing in the Boston Dynamics machines in terms of mechanics or hydraulics that was not available or could not be implemented in 1968. Just look at the video [2] (got to love the sound effects). This machine, all by itself, proves my point about the futility of some of this research. They all put the cart in front of the horse. The GE machine needed better effectors, sensors, energy storage, AI and control. The machine shows the amazing mechanical complexity that was attainable in 1968. Remember, no Solidworks, no microprocessors, no FPGA's, no Linux, just a dude pulling levers. Amazing stuff.
We are simply focusing on and throwing money at the wrong things.
[0] https://qht.co/item?id=6495440
[1] http://www.frc.ri.cmu.edu/~hpm/project.archive/Image.Archive...
[2] http://www.youtube.com/watch?v=ZMGCFLEYakM
EDIT:
Just came across this, which is really cool (1957):
http://cyberneticzoo.com/?p=8392