« Welcome to the Silkworm blog | Main | Silkworm »
17 June 2005
Collaborative Intelligence
Posted by Justin Leavesley at June 17, 2005 08:31 AM
In this post, Jon Udell talks about the lack of adaptive intelligence in operating systems. For example if I always open explorer and go to the same folder it would be nice if my computer learned that this was my behaviour and adapted to it, but today there are very few examples of this.
This kind of adaptive intelligence is what most people always expected of computers, and its failure to materialize accounts for their widespread dissatisfaction with the status quo.
It strikes me that adaptive intelligence based on the behaviour of just one persons actions is not going to add much value. In the end, most of the customisation I would like can be achieved by setting defaults and preferences.
Maybe the real potential of adaptive intelligence is when it is applied across many users, call that collaborative intelligence. We see simple examples of this in Amazon’s people who bought this book also bought these. Amazon has tracked the activities of millions of users of the system and extracted very valuable new information from the patterns of behaviours. flikr or del.icio.us are also great examples of collaborative intelligence.
It is not hard to imagine the value of extracting patterns from the use of a corporate shared file system or even allowing users to add tags to files on the shared drive to create a folksonomie type view of the system rather than the traditional file hierarchy.
The applications in the research community of tagging more than just websites could be huge. Of course this is difficult to do today as a lot of the information is in databases and you cannot book mark the entry and even if you could, the real value is when the collaborative intelligence is built into the applications you are using.
It is within the end user application that collaborative intelligence intersects with the individual context. Synthesising these two worlds, the collaborative and the individual will surely take us a long way down the road of adaptive intelligent applications. Maybe we should call these semantic applications?
