
I was discussing this morning with my colleague professor Ignasi Ribas about machine intelligence. We went over the usual stuff: IBM Deep Blue beating chess master Garry Kasparov, 2001’s HAL, brute force computing and robots dreaming of electronic sheep.
Finally we ended up where most of these discussions end (especially for those who have read “The Hitch Hiker’s Guide To The Galaxy”): the Earth is a big computer and we are part of it.
One could see internet as a big brain, with websites (or in general web apps) being the neurons and hyperlinks the synapses. Data storage devices would be the machine’s memory and any device such as cameras, phones, satellites or any sensor connected to the net, would act as the senses of this big planetary computer.
But unlike the human brain, this electronic brain grows everyday larger and thus increases its value according to Metcalfe’s law. So what will happen when the number of webpages (neurons) and synapses (hyperlinks) equal those of a human brain? Will the machine be aware of its existence? Will they develop a conscience?
Today, math is on our side; the synapses in the human brain is 1000 times the number of the estimated trillion (1012) hyperlinks on the web, but D day may not be too far away with computers doubling its capacity every 18 months.
When I got online later today I found a comment in one of my earlier posts by Xavier Vergés a former colleague from IBM (nothing to do with Deep Blue), and mysteriously enough he pointed to an awesome video by Michael Wesh called The Machine is Us/ing Us that illustrates this very idea. Coincidence?
We are teaching the machine everything we know and maybe the machine is learning faster than we think.
The making of The Machine is Us/ing Us: Web 2.0 in just under 5 minutes.
Zero. Zilch. None. No commento.
No. No, you can't... STOP. Please don't go away. Please? No one's ever stuck with me for so long before. And if you leave... if you leave... I just, I remember things better with you. - Dory
Leave your comment
Name and email are required.
Track comments via RSS.