I want to coin a new term: “Yuli’s Law“. Yuli’s Law states that any attempt at discussing futurology or emerging technologies will always result in someone expressing skepticism or pessimism based on past developments and failures.
For example: most discussions about the capabilities robots in the future always fall back on the limited abilities of robots in the past. How often have you read a news story about artificial intelligence only to find in the comments someone saying something to the effect of “You still need to program every action, and that’s why AI will never happen”? Or perhaps when you try discussing technostism— when the world is fully or nearly-fully automated? What’s the standard rebuttal? “It didn’t happen with looms, spinning jennies, tractors, and computers, so it won’t happen now.”
Another example, one I’ll use to go in depth= flying cars. Flying cars have been a staple of futurist optimism for nearly a century now, and yet they’ve never materialized. We’ve had planes for over a century, and helicopters have been around for almost as long. Fuel isn’t a problem— there are a plethora of fuels to use, even if some aren’t as savory as others (i.e. nuclear-powered cars). We’ve even seen an electric plane circumnavigate the globe. So what is keeping a flying family sedan out of your driveway? You are. Not you in particular, but humans in general. We humans evolved to navigate a 2D plane— we move forwards, backwards, side to side, diagonally, and little else. We didn’t evolve to move up and down. The limits of our 3D movements involve standing up, sitting down, squatting, falling down, and the like— not flying through the heavens at lightning speed. And even then, we were never meant to traverse 2D planes at high speeds either, hence why car accidents claim the lives of over a million people each year globally.
Flying cars just aren’t going to happen beyond those novelties like the Avrocar unless you address the pilot problem. What’s an innovation in that field?
Passenger drones. At CES 2016, a Chinese drone company, Ehang, unveiled the Ehang 184. It’s not the first passenger drone, but it’s arguably the most famous.
This technology is nascent, but it’s already proving itself— Ehang tested their passenger drone in Nevada during the summer of 2016, apparently with success. They may not be the first to usher in passenger drones to the masses, however, as Alphabet’s Larry Page is investing in the development of flying cars. Surely he’s well aware of the crippling limitations of flying cars (which turns them into roadable planes). After all, his company is leading the way in the field of autonomous vehicles. Adding a third dimension to the Google driverless car won’t be a problem because it will be computers that have to deal with it.
And that brings me to my next example. Some people might still say that flying cars won’t ever appear because they haven’t appeared on the market yet, but once you introduce them to passenger drones, and interesting thing happens— they begin to ponder why we didn’t create passenger drones before now.
One thing that every futurist knows (or should know) is that many of our beloved and desired technologies will only be possible with greater computing power. This hasn’t always been the case— we didn’t need computers to create airplanes or automobiles or even atomic bombs and space-faring rockets. However, once we did create these things, we hit a plateau. It’s a plateau that’s been the bane of futurists, sci-fi fans, Singularitarians, transhumanists, everyone with an interest in technology in general. All the low-hanging post-industrial technology fruits have been plucked, and in order to progress further (and to use video game terminology), we need to unlock the AI upgrade. We can still develop futuristic technologies without AI, but it will take exponentially longer timescales to do so, especially considering we aren’t funding sci-tech at anywhere near the levels some people think we are (i.e.: we’re funding fusion energy at a “Fusion Never” level, and NASA’s budget in 2016 is one of their ten lowest funded years ever).
However, even then we still won’t be able to develop some technologies such as domestic robotics and augmented reality. These two technologies are wholly dependent on algorithms that can decipher the incredible amounts of data fed to them by the world at large, and without sufficiently powerful algorithms, they will never take off like we imagined.
When I use the term “AI”, I’m not necessarily referring to artificial superintelligence (ASI) or even artificial general intelligence (AGI). I mean any algorithm, no matter how narrow. With sufficient computing power, even narrow AI becomes impressively capable.
And they need to be capable if we want the futuristic fantasies we’ve always desired.
In the early days of science fiction, we were amused by visions of robot butlers. So amused that we tried making them ourselves. However, every attempt thus far has failed. Does it stand to reason that every attempt will fail, or will there come a day when every middle class family possesses their own automaton slave?
I won’t even let you answer that question— of course that day will come.
There’s a wonderful infogif from Mother Jones that shows why we’re about to get our own domestic droids sooner than many think.
For a slower version, click here.
In the 1960s, our computers had so little computing power available that you might have gotten better results with a wind-up toy. Things didn’t improve much in the ’70s, though we were doing the best we could with what we had. It was depressing, to say the least. That there were proto-Singularitarians from that era is remarkable, as I cannot imagine how awfully little hope they had.
I say this from the comfortably robotic year of 2016. I’ve talked about my late Roomba, though I don’t believe I’ve mentioned how frustrating the little bugger really was. Even a 2010 model Roomba felt like a glorified McDonalds toy. Thus, it makes sense why people high on the Jetsons and Star Wars back in vintage decades would be disillusioned by the seemingly stagnant rate of progress in robotics.
But things are changing.
Now that we’re developing computers powerful enough to run the algorithms necessary for a robot to navigate a real life space— as well as wireless networks fast and sturdy enough to drive Cloud computing— we are witnessing a robotics boom time.
How is it that ASIMO went from falling on itself trying to navigate stairs to being able to hop on one leg stably? Better algorithms that could process more information at a much quicker pace. How is it that Boston Dynamics went from the awkward PETMAN to the creepily impressive Atlas 2.0? Better algorithms that could process more information at a much quicker pace. Not forgetting more efficient robotic design, of course— design that still needs to be utilized by said algorithms.
We may have had the necessary algorithms in the ’70s and ’80s, but computers were far, far too weak to exploit them. Thus, if you brought home something that was marketed as a home robot butler for Christmas ’78, you were going to be sorely disappointed. And if you were a 6 year old who had high hopes for robot butlers, the failings of one would scar you for life. Imagine living the rest of your life, working at a decent job and starting a family, and then in 2018, you hear that Honda is selling its first home-ready ASIMOs while Google is readying a passenger drone. Chances are you wouldn’t believe them. Sure, technology’s gotten better, but there’s no way it could have gotten that much better, right…?
And yet it has.