|22 May 2004 @ 20:47, by Roger Eaton|
Danny Hillis of Thinking Machines fame has published an article on the Edge about his latest venture, "Aristotle" (The Knowledge Web). His idea is to combine an authoring tool, a peer to peer network and a specialized browser to enlist us all in the building of an all encompassing knowledge database for humanity to be accessed through our own personalized artificial intellgence (AI) tutor, which he calls "Aristotle".
Hillis' article has replies by several well known thinkers from the Edge's Reality Club, and they roast Hillis over the AI issue pretty thoroughly, in the academic manner. With the basics of the semantic web now in place, though, my guess is that Danny's company can come up with something that begins to work as his Aristotle. His plan is to get the Knowledge Web started by landing contracts from both non- and for-profit companies to build what amounts to the next level in web tutorials for very particular areas of knowledge.
The problem that I see in all this is less the AI issue than that *any* system that includes back-end distributed processing has something of an imperialistic attitude built in, even if it is based on open source and open protocols. To exaggerate, it becomes a "system for negating alternatives". That was how IBM's Systems Network Architecture was referred to back in the 70's and 80's when it was introduced. If Aristotle catches on, then it will catch on big, and may well swamp any other ideas along the same lines.
The dynamics of the web take the "to him that hath shall be given" phenomenon to the global scale. As a student, you will want to pick a web tutor who will thrive in the chaotic evolution of the web. If your Aristotle dies, then all its knowledge of you dies with it, and you will need to train up a new tutor to understand your ideosyncracies, so best to pick the winner, if you can. Thus success begets more success, and hype is king.
But what are we to do in the face of SNA? I am thinking of the voice of humanity (voh) network that this blog is dedicated to. The voh network will also be a system to negate alternatives. Any peer to peer network with distributed backend processing will tend to negate alternatives to the extent that it succeeds. (Though it is not really that bad, as I will argue below.)
We could stick to a go-slow approach. Rss/atom syndication is the kind of piecemeal development based on relatively simple notions that would qualify as go-slow. If we had Danny Hillis' Knowledge Web totally in place already, there would be no room for rss and atom. This would indeed be a loss, because the simple building blocks of rss keep an army of open source coders involved, while the complexity of Aristotle will restrict development to big money shops.
Well it is not that bad, I think. Particularly in the case of the voh, it can't get that complicated, because I am coding it, and the really complicated stuff is beyond me. Plus I am highly aware of the sna problem and doing my best to keep it simple. But even Aristotle, based as I suppose it must be on rdf and owl, is just going to be one upstart among many. We are still too early to see which developments are winners in the knowledge web area, so the voracious "to him that hath" meme is not about to kick in quite yet. We can take heart, also, knowing that the original IBM SNA has been largely overtaken by IP. Let a hundred flowers bloom!
Also, just to point out, Aristotle is definitely going for the annotated web. Hillis writes, "In the knowledge web, not only the author, but also third parties can create links, comments, and annotations."