If they can figure out how to wirelessly power the cubes, or allow connected cubes to power each other via conductivity (perhaps at the cube edges), then the self-assembly of the bridge at the scale they're aiming for makes much more sense.
Hitting on the flywheel as the core mechanism for locomotion was amazing.
People 5 centuries from now will look into 20th and 21st century robotics forays in much the same way as we look at Babbage or Heron's steam-engine. Clever, perhaps ingenious, but hopeless with current technology.
Alright, so what do you propose robot researchers do?
Just sit around, somehow waiting for technology to magically improve?
I might be naive, but I like to think of things like this research as the way to get the technology to improve. It's not as if technology has a life of its own and will just improve regardless.
I propose that they go fishing and wait it out till there's some game-changing breakthrough in some underlying technology. I'm not sure which are the right underlying technologies in much the same way as Babbage's critics couldn't identify "integrated circuit" as the missing enabler. Could be advances in materials, nanotechnology, even biotechnology. What I know for sure is that when these breakthroughs do happen, all 20th and 21st robotics "pioneering" work will become obsolete and will not serve future practitioners. Robotics is not like mathematics where you can lay a foundation that'll still be relevant and useful to people 10 or 20 centuries down the line. If Babbage instead of fussing with mechanical cranks and gears had invented the Turing machine his life's work wouldn't have gone to waste.
So let's say that in 2050 all cars and trucks -- along with lawn mowers, leaf blowers, air compressors, etc. -- are based on electric motors descended from the current Tesla. So much so that internal combustion engines are completely obsolete.
That does not imply that developing internal combustion engines was a waste of time -- they have been useful for decades for a variety of purposes. Even if they're never used again in the history of mankind, the uses they've already had are enough.
The current state of robotics has allowed some of great things to be built. Industrial automation, mine sweeping robots, goofy little vacuum cleaners, solenoid-driven irrigation systems, etc. So even if it's all ultimately a technological dead end, it's still been justified.
In any case, I don't share any of your certainty about the direction of future robotics. With the increase in accessibility over the last decade of microcontrollers, sensors, and effectors, and the corresponding increasing in accessibility of enabling technologies like 3D printing and CNC, I think it's entirely possible that we're about to see an explosion in the availability of products with embedded robotics in common day-to-day items -- things that start as hobbyist projects built by hackers but quickly become commercialized into consumer products. This may not be what you mean by "robots" but it is an obvious, useful application for the technologies in robotics nonetheless.
This is ridiculous, partly because you're asserting that robotics doesn't work at the moment and partly because it's not clear what your objective is (realistic human emulation?)
Moreover, even if some game-changing development happens after your death, it's still useful to get something done today. The Boston Dynamics rough terrain animal-like cargo bots are probably the most interesting application, although there's lots of industrial robots out there doing incredibly valuable work like pick-and-place.
And you're posting this to Hacker News, where people discuss startups almost all of which will be dead, unrecognisable or absorbed in 20 years time let alone a lifetime.
If they can't identify the "missing enabler," then how are they to know it's missing?
Breakthroughs don't just happen and change everything. In almost every case there's a lot of background work, like this stuff, that builds up to the breakthrough.
They don't have to identify it and they don't have to know it's missing. My whole point is that the breakthroughs will happen in some other technology tree, and they will most likely occur in the context of some more mundane, profit-seeking commercial endeavors, in much the same way that advances in metallurgy and machine tool technology enabled the steam engine although the basic principle and some crude prototypes were already available since classical antiquity. What's been happening since the 1950s in robotics (I am not talking Toyota car factory robotics) is simply a misallocation of taxpayer's capital and hurts science and technological advancement since it fails to allocate research capital in the way most likely to lead to sustainable, economically exploitable breakthroughs. Of course all that pales in comparison to money wasted in social "sciences" etc.
Could a similar statement not be made for all the enabling technologies? If all the materials scientists, nanotechnologists, and biotechnologists went fishing as well I suspect the future would look a lot like the present.
No, a similar statement cannot be made because there's a difference between hunting Chimeras and doing research when there's profit to be made (even when gambling with high-risk research). That's the litmus test. If there's a plausible expectation for profit, or even tax-funded / subsidized applications, then that means that goal is not entirely unrealistic in our technological era. Otherwise, try again next century.
How would you get any breakthrough in basic sciences without essentially blindly spending on research, hoping something will come out of it in 50 years or so? That some other people in some other country will base their commercial applications on? Not everything is profitable or useful in time for the next quarter. Basic research is a tide that lifts all boats.
Also, my vacuuming robot is pretty good, I haven't had to manually vacuum in several month and there is virtually no dust. I'd rather have that now than wait for robot researchers to get bored fishing and build something incredibly better.
Exactly what sort of robotics are you railing against? It seems to me that your arguments are essentially the old joke "We're wasting 50% of our money here; if only we knew which 50%". Or, "Things aren't useful, until they are".
I have a robot on my desk that makes me stuff out of plastic and it cost less than a grand. Who gives a damn if it is obsolete next week thanks to nano-wizardry, it makes me stuff now.
The conjecture that any developments that are much later discarded for a better approach are a waste of time doesn't really address how one can never know what ultimately will prove most useful for future innovations in the first place. Or that, you know, some people like building stuff just for the pure enjoyment of building. Which IMO is just as important.
Until Apple introduced iPhone nobody thought it was possible to make smartphone like that. Six years later many of us carry access to significant portion of human knowledge in our pockets. My hypothesis that at current technological level we can have full stack mining/refining/energy generation/assembly of new robots done by robots. Trick is putting right pieces together. Why is it not done yet? Not economically viable yet. It is far cheaper to design a simpler system with human controls, that assemble full stack robotic implementation.
> Babbage did invent the Turing machine - the design was Turing complete.
No he didn't because he failed to abstract it and publish it in a way that would benefit the builders of the first modern computers in the 30-40s - they knew nothing about his Analytical Machine. So, if you were building a technology tree for a CIV-like game, sorry, "Analytical Machine" doesn't appear as a pre-requisite for "Modern Computer". Similarly, today's "robots" will not be ancestors in the technology tree of the first "real" robots, if and when the latter eventuate.
On the off chance that you are not just trolling...
You seem to have this retrospective view of technology creation where anything that didn't pan out or doesn't connect to current technology was therefore a waste of time and effort. The problem is that at the present nobody knows what's going to work out- you just have to try. Some ideas will work, some will not, some ideas will be lost and then re-discovered. It's a messy process. I'm sure there are robotics researchers who are wasting their time in dead end technologies. But I don't know which ones they are, and neither do you. I have opinions, but I'm not prescient.
"Analytical Machine" doesn't appear as a pre-requisite for "Modern Computer"
People thinking about them helps. You don't need the first people designing computers to have heard of an analytic engine and be working from Babbages notes for there to be dependencies. And it is not as though Babbages work did not have an influence on mathematics and technology in its day, it was widely known and discussed in both academic and political circles, and will have certainly influenced many. It would seem unlikely for the people working on the first working computers to not be drawing on some work that had echoes of the earlier attempts in the field, irrespective of whether the researchers themselves were aware of it.
If you're really going to propose that technology and science will advance that much, then I feel compelled to counter that people five centuries from now will look back on 20th and 21st century robotics forays and pity us for having to live in such tiny, stupid, chaotic brains. They'll wonder how we ever accomplished anything while so congenitally incapable of proper thought or feeling.
Hint: the lesson is never to make futurist predictions, because they're almost never coherent.
Prediction: even in the future, the Kwisatz Haderach will not happen. Attempting to predict the other predictors, who are also trying to predict you, leads to some nasty, headachey recursion issues.
I assume you're using some definition that excludes the millions of industrial robots being put to productive use every day. But I don't know what that definition would be.
I'm typing on a computer in large part assembled by robots. Earlier, I took a call on a phone in large part assembled by robots. This morning, I drove a car in large part assembled by robots. It's hard to see how the current state of robots could possibly be described as "hopeless".
Or in other words, Asimovian robots, minus everything that made them useful or interesting.
Humanoid body designs are actually bloody stupid for accomplishing anything other than emulating a human. What you really want is giant spiders.
I am actually not joking. Me and my biologist girlfriend have discussed the most practical shape for the unholy soldiers of a robot army. If you actually want to make use of robots, something like a spider is what you want: limbs to maintain balance and provide propulsion, some of which can also be swapped out to function as prehensile grabbers.
That kind of thing is already being done today by folks at DARPA, Boston Dynamics, blah blah blah. They're just very costly right now, and quite stupid to boot.
What you really want is intelligent giant spiders.
I still don't get it. If you start with an Asimov robot, then take away the anthropomorphism and the sentience, what remains that isn't found in a modern car factory?
Ability to move independently, with their own power source, be able to navigate our environment, homes, use our tools and do our bidding, following simple commands. It is likely that some DNA-engineered chimp/machine hybrid will probably become the first real robot exposing the folly of today's mechanical / electrical attempts in much the same way as we now know that the Analytical Machine was unrealistic with 19th century clockwork technology.
Ability to move independently, with their own power source, be able to navigate our environment, homes, use our tools and do our bidding, following simple commands
Apart from "use our tools", that's Roomba+Siri. Honda's "Asimo" is coming quite close.
It is possible that the current tech used in robotics will somehow form the basis for the work over the next 1000 years or so, either directly or via some tangential discovery by researchers working with current tech.
Without the benefit of hindsight it is impossible to tell with certainty what techs are worthwhile exploring and what should simply be discarded, the idea of discarding is of course based on the false assumption that techs that are later unused never had a use (e.g. we may solve a software problem with current tech that will form the basis for future robotics software).
This is a great demo. However I wonder, with one flywheel, they should only be able to rotate the individual robots along a single plane? ( And I am somewhat curious, what collective motion of more than one robot looks like. )
If they had two parallel flywheels that could be accelerated separately, I think that gives you rotation around an axis perpendicular to whatever surface the cube is resting on. It appears from the video that they have just a single flywheel, so maybe they're just taking advantage of how "noisy" the system is. That is, if a cube probably lands a bit "off", the next instruction can either fix that "error" or exaggerate it. Of course, they might also be just arranging everything carefully by hand before they start the camera. They never claimed the system was finished.
Cubes are easier to think about, but they're not the only polyhedrons that fill space. I wonder what this would look like with truncated octahedrons?
From the one look of the internals (1:50 in the video), there does not appear to be room to rotate the flywheel axis. That said, I would expect that is where they're headed when they can figure out how to use some of the space more efficiently.
I thought about this possibility, but having 3 flywheels increases the weight, which decreases the maximum angular momentum / weight that could be applied around a given axis. This is, I believe, the relevant quantity if you want to know how far a cube could jump in a different direction.
If the flywheel could rotate on two axes (and spin on the third, presumably), you would still have full control on which direction you apply forces.
Hitting on the flywheel as the core mechanism for locomotion was amazing.