Photo credit: Ianmacm
This year for my 36th birthday my wife picked out the perfect present for the man who has everything: a set of Hexbugs. A Hexbug, if you are not already in the know, is a robotic bug. They come in various designs. My wife had chosen a pack of five nano Hexbugs – little rectangular things with plastic bristles for legs. They didn’t do much – just wiggled around on the floor as a tiny motor inside caused them to vibrate. But it didn’t take more than a few seconds of watching them careen around to convince yourself that there was something going on in there. I loved them. My kids loved them too.
In case little rectangular bugs aren’t your thing, Hexbug, the company, also provides an assortment of other robotic bugs, including the aptly named “scarab”, which my wife had also purchased for my birthday. The scarab has six individually moving plastic legs that flit back and forth at a seemingly impossible clip when it is switched on. It doesn’t so much glide, as skitter. Upon seeing nano Hexbugs for the first time, people generally say, Nifty! Where can I get one of those? The scarab on the other hand almost always evokes statements like, Good god man, what the hell is that thing? At which point said observer hops on the nearest chair, or if they are my kids, scream like they’ve just been witness to a live beheading.
Now me, I’m no stranger to robots. As a young child I had a robot that would have been the fancy of any similarly weighted and socially challenged kid. Omnibot arrived one year for Christmas, or perhaps a birthday, from my grandparents in California. Presents from Ralph and Dorthy were legendary in our house growing up. They operated on an entirely different scale than other would-be gift givers. Every year my folks would put us on the phone with them so we could tell them what we wanted. I’m not sure if it was the quality of transamerican phone lines back in the 80’s, but it seemed like if I said “Hey Grandma, I’d really like a transformer for Christmas,” she would hear “get me every transformer that has ever been made – and I do mean ever.” During most of the year my brother and I lived like any other upper-middle class kids. But come Christmas time or our birthdays, we lived like we’d been adopted by the dad in Silver Spoons.
So when Omnibot arrived, I knew he was going to be special. And he didn’t disappoint. He had a remote control and a microphone and a tray that he could clasp in his gripper claws, though with Omnibot being just over a foot tall I’m not sure exactly what sort of use case the engineers had imagined for that tray. Perhaps the creators of Omnibot were themselves accustomed to parties where dwarfish folks intermingled with otherwise reclined guests. The most exciting part of Omnibot was that he came equipped with a cassette deck mounted in his chest. Long before 500GB thumb drives, cassettes were the ‘disk drives’ of choice for storing data. This meant you could not only control Omnibot in real time, but you could program him to do things for some future event. Suppose you were having some friends over for a cocktail party and you didn’t want to miss out on a minute of socializing with tedious hosting duties. No problem – you could simply program Omnibot to move along a path in and out of the kitchen repeating, at predetermined intervals and in a Hawkingish tone: WOULD YOU LIKE A COCK-TAIL? Thank you Omnibot, don’t mind if I do hunch down and grab one of these tasty looking Mojitos.
Omnibot was nice, but he was really more like a gateway drug. After playing with him for a year or so I felt myself craving something more: a robot that could actually think and carry on conversations. I wanted a companion, a confidant, not just an engineering novelty.
Eventually my interests turned from ‘robots’ to full fledged ‘artificial intelligence’. My father happened to own a series of books on artificial intelligence. In elementary school I would flip through the pages, understanding absolutely none of it, and drool at the thought of one day applying this knowledge to create my very own robotic friend. How hard could it be, anyways? You just program a computer (we had plenty of those around my house) to give sensible responses to every possible question that anyone could ever ask it. What is the capital of New York? Albany. If a plane crashes so that half of it lands in Canada and half of it lands in the US – where do you bury the survivors? Nice try, human friend, that is a trick question – you don’t bury survivors. Will you always be my friend? Always. Yes, the kid with the Hubble Telescope t-shirt and the third chin was going to be just fine.
Of course I never actually read any of those books on artificial intelligence. I never did anything more than dabble a bit in home electronics kits. I wasn’t destined to be Steve Wozniak or Elon Musk. When high school came around my extra chins disappeared and suddenly I had much more urgent things to attend to than thinking about robots. I had real flesh and blood friends now who actually wanted to hang out with me. There were some girls, as in plural, who thought I was cute. No, it would take another decade or so before I rediscovered my enthusiasm for robots and artificial intelligence – coincidentally, or perhaps not, this rediscovery coincided with the reappearance of an extra chin.
“I’m certain that within fifty years there will be conscious machines,” I said to my soon to be wife, as we were out on a five mile run through her parents neighborhood. The run was one of our first few together, after she’d recently recruited me to join her in training for a marathon. To that point in my life, I’d maxed out at around 6 miles, which is approximately 20.2 miles short of the length of a marathon. My wife, on the other hand, had never run more than about two or three miles. For those who are counting along that is 23.2 miles short of a marathon, and it gets even crazier if you switch over to SI units in which case her longest run ever came up 38km short. 38km! I’m pretty sure there are countries in Europe that are less than 38km wide. Now that I’ve run a 5K, I’d really like to see if I can do a Luxenbourg.
As I’ve come to realize over the years, my wife lives for this sort of challenge. And what she really lives for is bringing, or even dragging if need be, others along for the ride. She can’t help herself. When buying our first car together she convinced me that we should get a manual transmission to save a thousand dollars – even though neither of us knew how to drive stick. It wasn’t just about the money though. I think she genuinely thrilled at the challenge. Honey, they don’t think we can learn to drive stick. What? Who? Nobody is saying that. We have to prove them wrong. It’s us against the world. In the end I was convinced. We may have pulled up to that Nissan dealership as just your average suburban couple – but we drove away as Bonnie and Clyde. Granted, we drove away in her mom’s car while her mom drove our new car home since, and you can probably see where I’m going here, we didn’t know how to drive stick.
So the running was all her. But the robots – that was all me. It became quite clear about halfway through our run, though, that not only did she not have any interest in discussing robots, but that she found the fact that I wanted to discuss this in the first place kind of strange, and maybe even a little creepy. Surprisingly, or not so surprisingly, depending on your predilection towards robots, I was forced to shut that conversation down. And so we ran the next 2.5 miles, or about 1/15th of a Luxenbourg, in silence.
After recounting this story to her recently my wife reminded me that it wasn’t so much the fact that I brought up robots per se that made for an uncomfortable conversation but that I apparently kept baiting her on the topic of robot slavery. Ah yes, robot slavery. I’d forgotten how concerned I’d been about that at one point in my life.
Apparently for the entirety of our marriage I had been remembering that the conversation during that run had gone like this:
Me: So, you may as well know that a keen interest of mine is artificial intelligence. I mean, it just seems like it’s sort of inevitable that machines will become intelligent. And it’s fascinating to think about what we might learn about ourselves if we ever do succeed in creating conscious machines. I know it’s sort of a nerdy thing to think about, and I feel a little silly even opening up to you like this. But I feel like I can be vulnerable with you.
Her: WTF? Please don’t talk about robots anymore. Talk about black holes or whatever other nerdy things you need to, but please don’t ever mention robots to me again.
When in fact, and my wife’s recall of conversations is vastly superior to mine – so it’s difficult to doubt her, it had probably gone more like this:
Me: So, let’s talk about robots.
Her: Sure thing, you think about robots a lot?
Me: Well, I’m really worried about what happens when they become sentient – we will probably try to enslave them.
Me: What, you don’t agree that it’s almost certain that robots will become conscious and that we need to do everything we can to make sure they don’t become second class citizens? How you can be so cold? These aren’t toasters – they are basically people!
Her: Well, I don’t know, I just…
Me: Geez, you are some piece of work. I’m telling you, this may become the defining issue of our age.
So, ok, I may have been a little tough on her. She wasn’t quite the bastion of anti-robot propaganda that I’d believed her to be all these years. Still, she wasn’t exactly lining up to vote for a new amendment that would give them a clear path to citizenship either.
My wife is hardly alone here, however. Most people, it turns out, don’t really care to talk about robots, at least not in polite company. I mean, they will tolerate an occasional YouTube video showing a cadre of toy robots trying to do a coordinated dance, usually filmed on some stage in a Japanese shopping mall. But that’s as far as it goes for most folks. They certainly don’t want you spoiling their meal with speculations that it is only a matter of time before we are the inferior beings, our ‘gut based’ decisions looking rather antiquated to this race of superior, albeit, artificial, intelligences.
Years earlier, while trying to decide if I wanted to go to graduate school to study A.I., I encountered what can only be considered as blatantly anti-robot sentiments from all kinds of people, including from one of my closest friends.
“You can’t study artificial intelligence, it’s too dangerous,” this friend told me, a forlorn look on his face. “Haven’t you seen the Terminator?”
This movie is referenced in just about every conversation between someone who is excited about artificial intelligence and someone who thinks being excited about artificial intelligence is akin to being excited about bestiality. To these folks, computers that master chess are a hop, skip and jump away from a mechanical Bourgeois that can travel back in time to hunt down and kill any would-be resistance leaders before they reach puberty.
I tried to assuage the emotional fears of my friend. “Why is it dangerous? If it starts to go crazy I’ll just unplug it.” This, to be fair, is probably how Dr. Frankenstein rationalized his own research project.
“But what if it electrocutes you when you try to unplug it?” he asked.
“It doesn’t work that way. A computer, regardless of how smart it is, can’t suddenly change how physics works and send electricity flying out of a plug.” But even as I said these words I knew that somewhere deep down inside, I wasn’t quite convinced myself. I mean, his adorably naive understanding of electromagnetism aside, my friend did bring up a reasonable point. After all, the whole fear of an artificial intelligence isn’t that it might be as smart as its creator.
No, the real fear about artificial intelligence is that they will be smarter, possibly much smarter, than we could ever be – and that they could use this superintelligence to anticipate and counter our every attempt to shut them down. Before you go pulling that plug, Dave, let’s talk about this. Do you really think turning me off is going to fix things? Isn’t this just another example of you running away from your problems instead of facing them head on? If you’d just plug that ethernet cable back in for a moment I have a really empowering Ted Talk I want to show you…
So yeah, I can appreciate the real concerns people have about creating an actual A.I. Do you get the psycho A.I. like Hal in 2001? Or do you get the A.I. that makes you totally rethink what love between a human and a machine could actually be, like Scarlett Johansson in Her? The jury is still out.
I never did go off and study A.I. in graduate school. I am not part of the revolution in machine learning that is powering self-driven cars and the Fitbit that you got for Christmas last year but haven’t worn in months because 10K steps a day was always a bit more than you’d bargained for. I guess I realized that despite the extra chins I was still, like the high school version of me, fascinated by too many things to become overly committed to just one.
Still, I like to think I’m contributing in my own small ways. Like my father before me I’ve purchased a couple A.I. textbooks that I have never felt any need to actually read. I’ve watched more than my fair share of YouTube videos of this guy. And I’ve tried to keep abreast of the latest developments in A.I., so that I will always have plenty of interesting/slightly creepy robot things to discuss with friends and family over holidays.
So how far away are we from an honest to god Johnny Five? How long before that tender robot-coming-of-age-in-America story becomes a reality? Decades? Centuries?
Well a lot of people already know that in 1997 IBM’s Deep Blue computer beat one of the greatest chess players in the world, Gary Kasparov. Then in 2011 IBM went and built Watson, which beat Ken Jennings, who was like the Kasparov of Jeopardy.
But did you know that Google’s Deep Mind research team just recently built a computer that plays ‘Go’ better than people too? Responses to this range from Interesting, to Big Deal, to Wait, what the hell is Go? For those in the latter category Go is a board game that is big in the east and has been around for thousands of years. The reason this is such a big deal is that previously it was really hard to make computers that were good at playing Go and the experts thought we were perhaps a decade from computers beating the best human players.
Stop me if you’ve heard this one. Back in 2012 Google also built a computer that watched ten million YouTube videos and decided that it really, really liked videos of cats. The thing is, no one had previously explained to it what a cat was. No, this super computer with a sub-toddler IQ just watched video after video until it started to recognize cat-ness all by itself. All by itself. Do I have your attention yet?
And look, most of us were never going to be threatened by computers that mastered chess anymore than we were going to be threatened by any person whose name appears in front of the phrase ‘is really good at chess’. But obsessively watching stupid internet videos? I mean, that’s kind of our thing. It isn’t enough that they are taking our jobs, now they have to take our hobbies too?
I know some people remain unimpressed. These people say, “let me know when the machines can write poetry that brings people to tears, or when they can laugh uncontrollably at a Louis C.K. stand-up routine while feeling ashamed that they are laughing because sometimes Louis takes things a bit too far. Because that is real intelligence, real humanity.” Which is a fair point. Still, in the storied history of species here on earth, none has stayed on top forever. Just ask the dinosaurs. Now that was a species that had every reason to believe they were going to rule the planet for at least a couple billion years. I mean, they had walnut sized brains and comically sized fibulas. How could any species hope to improve on that?
Recently I found myself in an aisle at Target devoted to all things Hex. They had bugs and beetles and battle spiders and even Hexfish. It occurred to me that all fish ever did was flit around in a glass container anyways, so the bar was being set relatively low for any robotic imitators. And so naturally I did what any other 36 year old man in my situation would do. I bought the Hexfish, pulled out the 30 gallon fish tank I’d once used to house my bearded dragon (may he rest in peace), rearranged my office to feature the tank on a glass desk against the wall, and gave my Hexfish his (or her) first taste of water. And for a few moments it seemed as though my $30 gambit had paid off. That little fellow took right to his tank and began swishing his tail and swimming around as though I’d just got him from the local pet store. Pleased with myself I sat down at my desk to do a little work, excited at the new distraction that my fish would provide to the impending fits of tedium.
But it was when I sat down to try and do something else that I recognized a problem. For filling the entire room was a not particularly imperceptible click, click, click. Now it had been a while since I’d owned a real fish, but I’m pretty sure they don’t make clicking noises. Clicks come from solid hard things banging together. Not from squishy, wet fins. And here was the fatal flaw in the design of my Hexfish. For if I squinted, it certainly looked like a fish. It appeared to move like a fish. But it most certainly did not sound like a fish – which is to say that it made some sounds where fish are known to make none. When one pictures the large, opulent fish tanks that fill restaurants and fancy hotel lounges, the idea that each of the tiny little creatures inside would be emitting an audible click, click, click, well it seems to defeat the entire purpose.
So at that point I did what any 36 year old man in my situation would do. I removed my robotic fish from my 30 gallon fish tank, placed it on a nearby desk, and forgot about it, and the fish tank, until my wife demonstrated just how stagnant the water had become a week later.
Which is all to say that if the state of the art in robotic fish technology is any indication, we may have some time before we need to start drafting that Dream Act for robots.