http://community.brighton.ac.uk/jd29/weblog/16499.html

http://www.edge.org/documents/archive/edge183.html

An interesting and thought provoking article from Jaron Lanier which (interestingly) came out at almost exactly the same time as a paper I wrote for ICALT 2006 that said some very similar things, albeit with a far more positive slant and an emphasis on education.

His thoughts on scale (nice phrase,  'jittery shifts') and the need for things that work on different timescales are absolutely right. Apart from anything else collective applications that feedback results to people, such as collaborative filters and social navigation systems, can only really work if they have inherent delays built in. There are also issues of scale as we ascend the 'meta' ladder, which he is a bit derisive about. I disagree with his derision, to some extent, but agree that we have to think really carefully about whether we are ascending a hierarchy or simply reassembling the same things in the same way. We don't need more of the same - we need variegation, hierarchies and networks.

I also like Lanier's thoughts on the need to combine the strengths of individuals with the strengths of the collective which, as he (almost) rightly says, is a useful tool.

He is right when he says:

"The collective is good at solving problems which demand results that can be evaluated by uncontroversial performance parameters, but bad when taste and judgment matter"

However, I share Clay Shirky's concerns that he has lumped all collective tools together. They don't have to be this way, or any way, in fact: there is a lot of variety in the ways that collectives can be assembled. I am not even sure that some examples of the collective mind at work are properly labelled (Wikipedia is so much closer to a group system than a collective system that it barely merits the name, whereas Popurls probably is more collective, albeit in a fairly crude way). It is precisely the fact that their strengths come from an amalgam of algorithms & processes and the individual, somewhat disconnected, acts of intelligent agents (people) that they are so interesting. These are tools that make themselves, repeatedly, again and again, but not randomly. That makes them a bit unusual. I have previously referred to collectives as a kind of cyborg, but that is actually not quite right. They are not cyber-organisms so much as crowds, or ecosystems, or maybe flocks. 'Cyberflocks' is quite a nice name. 'Cyberhives maybe?' We need a better way of talking about these things.

If we want to keep treating them as tools, great, they have their uses. But the really interesting challenge is that of creating richer, more useful, layered applications that achieve more complex goals than recommending movies or news items to a massive crowd.  If they are to be pedagogically useful, they have to evolve and take on a more sophisticated, layered, multi-scaled, context-driven structure. My own 10 principles in my book (summarised in a rel="nofollow" href="http://www.ifets.info/index.php?