On a global AIBO consciousness… (FOO '06)

My third FOO post in a row, and I’m only just getting onto the talks themselves. This time I learned from my experience last year and made sure that I tried to steer myself towards talks that were outside my natural territories. I think I only really screwed up twice – once by going to a session on the democratisation of media hosted by the guys from Digg, Kevin Kelly and Dan Gillmor and once by going to Steven Levy‘s session on the future of the music industry. Steven’s always good value and it was great to see Dan again (however briefly), but realistically I knew these territories too well and wasn’t likely to get my mind very satisfactorily blown. It’s possible of course that my presence was useful for other people, but to be honest I kind of doubt it. Two things however do stick in my head – firstly when I told Kevin Kelly that most magazines were just regurgitated press releases (having completely erased from my head for a moment that he was the founding editor of Wired). Not the right context to go about making massive generalisations, really. The other thing that really made an impact was how nice the guys from Digg are. Simon, Paul and I did some exploring around Digg a while back for work and my impression of the site completely changed from it being a trivially easy proposition to a highly polished and carefully crafted piece of work – both things that seemed to be reflected in the people who had worked on it. Smart, decent, honourable people. Very interesting.

Anyway, the talk that I really wanted to start with was one by an old colleague – Matt Biddulph. He was talking about some of the things he’s been doing (during his year of slacking off) that hybridise Second Life with the rest of the web-based internet. I’d seen his Flickr screen in-game, but the thinking he’s been doing around it was really fascinating to me. In particular he’s been thinking about Second Life as a useful prototyping environment for ubicomp stuff of the Adam Greenfield Everyware variety and his thinking pushed me off in all kinds of interesting directions. A few of the key concepts he described that stirred me up:

• The Invisible Tail – being the sheer number of real-world objects or products that are currently not represented in data at all (with the result that you can’t yet leverage any of the long-tail benefits supplied by the Interweb). Lots of Age of Point at Things stuff in there, but more interestingly framed;
• The relationship between the above and concepts like thinglinks.org, which allow individuals to give individual items identifiers that will make them annotatable over time;
• Ubicomp Middleware – specifically, “what’s the outboard brain for an object in second life” and what could you do with it – lovely concepts there;
• Each AIBO becoming not an individual object but an endpoint for a global AIBO consciousness – and if that didn’t get you salivating then you’re dead inside;

Lots of things occurred to me during the talk, including the possibilities of using in-game architectures as structures to layer visualisations of extra-game information. So for example, could you have a huge structure that was designed to reflect the spacial territories in the Google NewsMap? That is to say, could each room in the structure represent a story and could its size be influenced by the current significance of the news story in which people could discuss the story in question. Could you basically create a virtual building that actively guided the discussions within it – literal forums for short-term topics of interest that grew or shrank depending on various metrics.

Another thing I started wondering about was to what extent you could use collaborative filtering mechanisms (or mechanisms like the old BBC Homepage patina) to alter objects in-game in real-time to reflect the usage patterns of the people who used them – and what implications this could have to automatically reconfiguring objects in the real-world. When I mentioned this to Matt he said it was like your Powerbook automatically getting smaller when you got onto a plane because it had discovered that people who used the smaller ones tended to bring them out more often on planes. I started wondering about evolutionary algorithms and slowly evolving functionality spreading virally across mobile phones. I don’t really have a sense about whether this is all one great conceptual dead-end but something tells me that objects that can learn from how all other objects like them are being used is an idea that will have a time.

Anyway, that’s enough to be getting on with. More later.