Danny Ayers review of Nova Spivacks From Semantic Web to Global Mind. I guess Nova’s getting ready to ship. Hopefully he’ll (Nova) be able to make it to our “micro-content” dinner in NYC at Keen’s on Aug. 19th.

=============

A post from Nova Spivack, From Semantic Web to Global Mind, in which he looks at…well, the title says it all. It’s funny, I don’t really disagree with any of his major points, though I wouldn’t have put things in this way at all.

My personal take on each section:

Distributed Intelligence
A nice line here: …whereas basic written languages simply make raw information portable, metalanguages make knowledge and intelligence about information portable. But I disagree There’s the usual problem of differentiating between x and meta-x, if that matters in the slightest. Existing human languages are pretty good at making at making knowledge portable, but there are certainly at least two related aspects of web languages that do make a difference. That the (meta-)data is machine-readable is one big step, and from that the data being machine-processable is another. A practical key is that the web languages are allow declarative expression of the information independent of processing, something which Nova is obviously aware with his reference to there being no need for hard-coding.

The Internet is a Brain…and the Web is its Mind
Again I kind-of agree, but would be very reluctant to put it in these terms. The Internet has features like those of a brain, but isn’t yet at all smart like a mouse, and the web is currently nothing like a mind apart from in the sense of carrying a whole load of jumbled information around in it. The analogy to the animal brain only works to a point, the nature of the system is so very different. I’d have probably thrown in the phrase ‘hive mind’ somewhere around here, the net as a whole gets what intelligence it has from lots of little stupid entities. But even then that isn’t a very good analogy, as the millions of individual humans sat on the edges are an incredibly important part of the system.

Memes are Evolving Minds of their Own
Hmm, sounds nice but I’d be tempted to use the meme notion as it currently stands - pretty self-unaware little items of belief (or knowledge).

The Infrastructure of Distributed Intelligence
This is more on ground I’m comfortable with, whatever analogies you use for net intelligence, the wiring is a significant part. A distinction I think Nova blurs here is between design and emergence. Viewed holistically I suppose the XML spec could be seen as an emergent property of the human+machine system, but if we’re talking about the net as an mind in its own right then the watchmaker isn’t blind.

The Evolution of Metalanguage
Not sure about the angle of this section at all, many of the ideas of the Semantic Web languages have been around for thousands of years (leading up to first order logic), it’s only when they’re combined with computers, in particular a big network of computers that the utility explodes. The feedback loop, that the philosophers can now use the computer as a practical tool is probably quite significant too.

Mutter - we might have these metalanguages, but still I can’t link to the individual paragraphs in Nova’s piece, can we really expect a global mind before TypePad features named anchors? MARC’s answer = NO!

How the Global Mind Thinks
This section gives a high-level view of the layers of the Semantic Web, and notes the role grassroots stuff like RSS is likely to play. But “Thinks” should either be in italics or at least a courier font (as in (cwm –think).

Can the Global Mind Pass the Turing Test
I like Nova’s example here, he was able to get the answer to a math problem quicker than an expert by farming it out to other semi-expert folks. But I think the Turing Test only makes sense for a very human kind of intelligence, and a future clever web is highly unlikely to think like like that.

Reading the Global Mind
Here Nova discusses data mining and meta-metrics on the cognitive web. It’s an interesting area, the fact that we can safely probe inside live systems makes everything a lot easier than analysis of biological systems. But I think there’s a caveat here - beware of reading too much into statistics. Just because it talks and wears the same clothes as a causal relationship doesn’t mean to say it is a causal relationship.

Minding Your Business
As the global mind develops it will initially be focused around making information more useable. Indeed. Nova goes into the high-level sharing of knowledge in organizations here, which does make sense, those which take advantage of these developments are likely to have an advantage. I don’t think the mind analogy is needed for that.

Knowledge Objects: A New Medium for the Web
Here Nova goes into discussion of the framework his company, Radar Networks is developing. Sounds interesting, the approach seems to be something along the lines of signed RDF Objects or CBD’s. There seems to be emphasis on the separation of data (e.g. media objects) and metadata, which is interesting - other folks (like Adobe) go the other way, embedding the metadata. I’m sure both have their place.

Knowledge Networks
I can’t comment much about this because a lot relates to the approach taken by Nova and co. I’m not sure whether the Knowledge Networks he refers to are the semantic ‘islands’ that when joined will for the Semantic Web, or whether it’s a different (proprietary) idea. Whatever, if you stick the stuff on the web you can have whatever knowledge networks you choose.

On a personal note…before the Semantic Web ideas were around (before the web or IBM-compatible PCs in fact) I myself carried the flame of the realisable global mind. I’d got the idea from old SF novels, and it fitted with the growth in computing power and distribution (particularly of early personal computers). But in recent years I’ve tended to avoid this angle, being more interested in making what we’ve got just one (big) step more useful. I still think the old-AI notion is basically reasonable, just the timescales and expectation of the work involved were completely out. That’s another reason I’m wary of talking in old-AI terms, the hype angle - engineers are much more likely to be receptive to a tool that does work today rather than a vision that might work tomorrow. Let’s just get the RDF and OWL designed in, the vision emergence will take care of itself.

My own change in focus has more and more pointed to the notion of human augmentation as an alternative (but not incompatible) long-term vision. Right now the distributed communications angle offers a major augmentation, when the Semantic Web parts kick in a little more I believe we’ll start to exploit computing power of computers a lot more. There were I suppose two major sci-fi angles to all this - the individual cyberman and there’s the global mind. Those of us with access to computer technology already are cybernetic, just that the human-computer interface is a whole lot clunkier than we expected. Right now the global mind is maybe bigger, thanks to the distribution of hosts, but an awful lot dumber. Really, really dumb. But an enormous benefit is already coming from a third avenue, those two pretty crude things mixed together, an augmented humantity. Not that it shows very much outside science and technology.

Anyhow, Nova has a word along those practical, next-step lines:

Note: The patent-pending Radar Networks Semantic Applications platform represents four years of stealth R&D. The platform is in pure Java and complies with open standards for the Internet and Semantic Web. A lightweight, unsupported version of the platform will be released to the public under an LGPL open-source license in summer of 2005. The full, commercially supported version of the platform will be available via a commercial license from Radar Networks.

Looking forward to it.

Marc’s final note: Oh goodey, Nova’s technology is patented. That way it’ll attract investment and THEN they can fuck us - right? Why would we get involved in something that’s patented?

This entry was posted on Monday, June 28th, 2004 at 6:57 am and is filed under . You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.