http://video.google.ca/videoplay?docid=-8867855532205075768&hl=en By way of illustration, here's my reaction to this video. I have been playing the Half-Life series by Valve since it first came out - and still do at least once monthly on average. It's a first-person shooter type game. I have played literally hundreds of Half-Life mods - modifications using the game engine created by independent programmers. In a sense I become Gordon Freeman, Ph.D. scientist and daunting warrior. Battling geek. If I could be fully immersed in 3D in a full-body tactile suit to play this, I would be there in a heartbeat. In the very early nineties I went through a period of very intense online text-based interactions. This resulted in me meeting a lot of people that normally I would never have met. That's how I met the wife - and that's pretty much when it ended for me (she however continued it to this day in updated form). I read about AI in the 70's - it was christened "cybernetics" by Frank Rosenblatt in the 50's - and in the eighties and early nineties got quite deeply involved in expert systems (top-down, rule-based) and neural networks (bottom-up, emergent properties). I never took it anywhere commercially, but I was pretty deep in and was paid to do that, to boot. The role of noise - random perturbations of the current state - is essential to get this to work right. You gotta jiggle it. Genetics works the same way - in that context, it's called mutation, and it's random. It's tempting to lurch off into a metaphysical aside re. this noise issue in order to deliver another solid punch to the creationists and religionoids. If they believe in a creator, then they also believe in an entity which produces random noise. LOL. But I won't labour the point. It's deeply more satisfying to ponder bottom-up systems. Rule-based systems are authoritarian, inflexible and brittle. They break easily. So did their political analogue, the USSR. In contrast, bottom-up systems are flexible and adaptable and self-repairing. Like all life. These days we refer to such systems as "organic". Yes, all these people in the video think like me, pretty much. We can all see it coming. We could see the basic shape of what's to come in the fifties, and everything since then has more or less been putting more flesh on the bones, and filling out the details with ever-increasing clarity. As soon as the microprocessor came along in the early seventies (I programmed the 8008 and 8080, the first commercially successful "x86" Intel chips), this process of filling in the details went into overdrive. The future sharpened up considerably - people could see where this was all going a lot better. For me, it was a little like an acid trip. I remember spending an entire weekend in 1974 on fire with the idea that I could teach an 8080 to think for itself :) Pieces of paper everywhere...LOL. I drowned in infinitely-nested meta-levels of cognition. That sure taught me a direct and personal lesson that AI was genuinely hard. Now our PC's are thousands of times faster and capable. And they're still about as dumb as a bucket of rocks. I architected a 3D graphics chip 20 years ahead of its time in the 80's. Time ran out for me to come up with a new kind of arithmetic based on Carver Mead's (the father of VLSI) tally counter idea. A Cambridge UK company introduced it a few years ago. The graphics tech you can buy now has finally caught up and is now miles ahead. That train has already left the station. I dabbled with the once-popular idea of reconfigurable computing. My idea was to send an advanced wave of reconfiguration commands on ahead of the data, so the thing was in the right computational shape by the time the data arrived nanoseconds later. I almost kicked off a start-up with my buddy Scott (but didn't). I've always been fascinated by ants; I bought Hobdoebbler's seminal and huge treatise on them, "The Ants". It's worth the price for the colour plates alone. Since publication, literally thousands of new species have been discovered. So it's already obsolete. I designed and built commercial robots in the eighties. That was gangs of fun. I specialised in the motion control. Instant gratification. The whole thing, operating system an' all, fit into the 32 KB memory of an LSI 11/23. How the hell did we do that back then? :) Nanotech gripped me in the eighties. I think I have every one of Eric Drexler's books, including the last one on systems which is very dense. I studied a little about genetic programming (Koza is the figurehead here, and he's in this video). It's fascinating how this leads to completely non-intuitive and surprising design solutions. Perhaps also surprisingly, the impact on mainstream engineering has been minimal. No, there is no conspiracy. Cellular automata interested me in the AI context since I first played "Life" on early 80's PC's. I bought Wolfram's book "A New Kind of Science". Don't. It's god-awful (although the central idea that physics can be reformulated from a declarative description to an algorithmic, procedural description is IMO still relatively unexplored and has potential). About the same time I got my own PC and got stricken briefly by fractals (chiefly in the context of video compression, my current professional expert incarnation; wavelets are in there also). Fractals share the property, along with neural networks, of producing complex behaviours from a very simple set of rules. This btw is why I prefer Go to Chess. So does the AI community, for the same reason, excepting perhaps the aesthetic of Go which I admire. This genetic programming led me back into genetics about ten years ago, where I'd been captivated a couple of times earlier: originally in 1971 (Stephen Rose's "The Chemistry of Life") and again in 1982 (Hofstadter's "Goedel, Escher, Bach - An Eternal Golden Braid"). Hofstadter's main education for me was not the "strange loop" idea so much as it was about the fundaments of information theory. I will simply say "The Jukebox Analogy" and enjoin you to go read the book, which is a unique classic by any measure. Precis per my interpretation - Claude Shannon's information theory, the backbone of the field, is incomplete, because the decoder's context is ignored. If I hand you a DVD and you don't have a player, I have not handed you 60 billion bits of information; I have handed you one bit of information. A back-burner project of mine is to reverse-engineer laboratory-experimentally-determined interaction matrices (these are available on biochips) to come up with a "circuit diagram" of intracellular interactions. Software, graph theory etc. Not sure it's possible (yet). But it sure would be nice to have a complete description of the cell. It's the most complicated unit component we know about, and there are trillions of them in the brain. Are you yet getting a sense that I have been somewhat interested in the subject matter of this video? We are heading towards the ability to simulate and construct artificial brains. To get a feel for the kind of progress we expect, one example is the speed of electrochemical signalling to that of electronics. A good rule of thumb is that electronics is about a million times faster (and getting faster than that all the time). So, if you could think a million times faster, from birth, you would have completed 16 years of education in less than 10 minutes. And you would remember all of it perfectly. And you would comprehend all of it perfectly. That's what a million times faster means when we build it ourselves. Some time down the road we'll be able to build brains that are hugely more plastic, hugely faster, hugely more compact, hugely more capable of massive communication bandwidths with others, and hugely more intelligent than our own. Then what? It's obvious. They will act as the cells of a multicellular organism. You can give that a name, but "superbrain" is not that name. Way too retro, way too much of an understatement. I believe that it will compose beautiful songs. But you will need a special jukebox upon which to play them. Let's not get too anthropocentric here. Step back a bit. Given the 8 billion years elapsed since the earliest civilisations (my figure), and the trillions of exoplanets, the chances are that such an entity (for want of a better name) has already evolved many times over in a multiplicity of different contexts. Maybe they all play bridge together every Tuesday evening. I used to. Yes, we grasp at straws to analogise this scenario. Trouble is, it's realistic. It's not fiction. Such an entity could easily leverage its sub-nano, quantal mastery of the brutish material realm to, to use a silly example, appear to the Cro-Magnons. They would worship it immediately. I would probably begin by insulting it. I find that gets to the nub of things quicker. Religion dispensed with, with a flick of the wrist. Next. .................. .................. I'm reminded of Yasunari Kawabata's "The Sound of the Mountain" (but purely from the numinous aspect). The silent rush of the stream of accelerated evolution is now turning into a dull roar. It's like standing with your back to an oncoming tsunami. Best to wear a hat. Best to learn to swim. What's it like living in the ocean towards which we are being inexorably swept by our own curiosity? Is it more like a womb than a tomb? And, like Adams' "Hitch-Hiker's Guide To The Galaxy" whale who spontaneously re-materialised one mile above the earth's surface mused, as the ground (something for which it had no semantics) rushed up to meet it, getting ever larger and closer - so too can we ask: "I wonder if it's friendly?" :) When all's said and done, one thing is certain - unlike the hapless whale, we are doing it to ourselves. .................. .................. Well, that's the materialist stuff. Once the video got into the more esoteric realm of consciousness, my enjoyment of it increased even further. There's much humour there. It gave me a good few chuckles. Yeah, we're bloody clueless about it.
We're sorry, but no images, videos or audio files have been added to this gallery.
If you would like to share media with the community, sign up and you can.