Okay, so I’ve decided to write myself out of something dangerously close to burnout. I hope it will work as a therapy, but it could also have an interesting side effect – reaching out to strangers who would actually want to read what I write. An admittedly daunting perspective to me, but also a healthy opportunity for pushing myself out of my comfort zone in a way I’ve never tried before. Moreover, writing could help me sculpt vague ideas into proper shape, so it’s a lot about learning and exploration as well, two things I’ve always loved.
Great, I seem to know why do I want to write. What will I write about then? I guess I should start slow and easy. Pick a thing, maybe two. Stuff that intrigues me, wonders I know (to lie about) well enough to describe them convincingly to others. Perhaps my topics could also be a bit weird and mysterious to tickle the reader’s attention. Then again, I’d better relate at least a little to something people hear and think and talk about, to let them slide into my world smoothly.
Wait a minute! It kind of sounds like one of my passions that also happened to become my field of work. I’ve been having AI dreams since Blade Runner blew my mind when I was nine. Meddling with AI professionally only made those dreams more refined … and sometimes more defiantly outlandish.
AI is one of the biggest mysteries humanity hasn’t cracked completely, not even after almost a century of intense efforts. Yet the ongoing research has already turned into products that sneak into our lives at an ever-increasing pace. When you read about AI in the news these days, it looks like something that’s here already (it’s not, not in a proper sense, but whatever is out there is cool nonetheless). Many people think AI is sexy, and even more are afraid of it. Perfect features of a thrilling topic I’d say!
It’s a great pastime to contemplate what the real AI could be before/if/when it will emerge at last. But even in my job, pondering such lofty visions is not very productive, at least not in short term. And that could be a problem. People have been kind enough to pump considerable amounts of money into our lab, but they might easily change their mind if they don’t see concrete results every now and then.
I can, however, let the dreams live their own lives in a world of pure imagination, sidetracked from my daily practical chores. And who knows, maybe the dreams will ripen when I let them loose in writing, and spawn some new cool research one day; stuff I can’t even begin to imagine now.
Obviously, there is a little challenge to this plan. How to write about my speculations on AI without becoming a total bore for most of my potential readers? As you might have noticed by now, I do have a slight tendency to beat about the bush, especially when I’m excited about something and feel like I should put all cards on the table. It’s no big deal if I make the reader excited too, but that exactly is the challenge I mean.
People have succeeded before in such tasks, so I’d better try and learn something from them. Isaac Asimov was one of the first to write compelling stories about AI. Well, robots, but it’s not that different as far as storytelling goes. Arthur C. Clarke brought us HAL 9000, an unforgettable computer villain. But these creations, and many other shaped in similar moulds, are too anthropomorphic to my taste, or way too logical for any complex mind worth its name.
Some of my other favourite dreamscapes have been conjured by William Gibson and Dan Simmons. I find their rather alien conceptions of AI very plausible. Oh, and they’re fun too. I really love the split super-personalities and voodoo avatars of Gibson, and the complex history and warring factions of the Simmons’ TechnoCore kept me reading until dawn more than once.
Turning to sci-fi greats for inspiration surely doesn’t hurt, but it can also make one depressed. I doubt I’ll ever be able to write like them. I admire the productivity and consistency of Asimov and Clarke. They both payed rigorous attention to science and their knack for turning dry and complex theories into exciting adventures is amazing. In all these skills, I’m merely a disciple. I wonder if I can ever hope for more.
Concerning the literary craftsmanship, I wish I could grow into a style as unique as Gibson’s one day, but I don’t even see it coming now. What I envy Simmons most is his art of building sound and complete imaginary worlds that span centuries and mega-parsecs of spacetime. That skill would be very useful for realising some of my plans, yet I’m only splashing at the surface now, and quite clumsily at that.
Never mind the pointless frustration regarding (the lack of) my capabilities, though. Even if I were the greatest writer ever, I still have to explore unknown veins of AI possibilities and bring some gems back, no matter how rough and tiny they may seem at first. Otherwise all I write is going to be just a neat formal exercise.
Maybe it’s best to let my beliefs do the talking. Making stories out of what I believe in should naturally lead me into exploring every corner of my half-formed guesses, ideas and hopes. And maybe, somewhere along the way, when I’ll struggle for the right words to tame and express all that mess, I’ll become an actual writer with his own voice.
What do I believe about AI then? For starters, I don’t think we humans will really create it. At best, we can hope to help it evolve on its own. I also think it’s gonna be quite hard to understand each other fully once the AIs really pop up somewhere. These days, the most likely space they’ll emerge from seems to be the multidimensional oceans of data buzzing on the Internet. We may be pouring petabytes of accurate approximations of the physical world there. But the perceptual reality and evolutionary pressures of Internet are still radically different from what we have lived in as a species for millions of years. I think that this will make the AIs more alien to us than beings living under different stars.
How could we even try to imagine such intelligences? How could I hope to write stories about them? Can I guess how it feels to spread my senses snooping all over the billions of lives people share online? Do I imagine how it tastes to munch on gargantuan streams of search engine queries? Hardly, as I can’t possibly have any frame of sensual reference. How could I expect to bring my readers into such world when even I can’t really imagine it?
Maybe there is another, indirect way. When exact solutions are hard or even impossible, there are usually plenty of other candidates just around the corner, much more attainable and still good enough (something one learns very quickly in my job). So how about trying to approximate the `AI experience’? What if there is another world I could use as a metaphor, alien enough but still kind of accessible to us humans?
I happen to know such world, and I came to love it perhaps even more than my AI dreams. It’s the ocean. We can hardly know how it really feels to live there, but we can still try to imagine it. We all came from there after all, even though it’s been a while.
So I’ve got it. Ocean and AI, the main ingredients of this improbable soup. Now I only need to stop pondering the writing and start actually doing it. From now on, I’ll focus on good old adventure, weird and amazing creatures, strange environments, deep mysteries and a fair bit of fun (at least that’s the plan). With a little luck and perseverance, I’ll soon start spilling something more than just metaphysical ruminations here!
Oceanic Soup 