a prototype insight machine

I used to think that given enough sensors, computation, storage, and connectivity, we could build an insight machine:

Very small structures based on electronic, optical, micromechanical and biochemical principles will enable webs of interconnected sensors embedded in the fabric of systems. These sensory webs will feed vast streams into interconnected utilities of information, combining with abstract information forms and interpersonal communication in a unified context...enabl[ing] synthetic sensory ‘display’ appliances that complement and augment the capabilities of the individual. In much the same way that eyeglasses enhance sight, such appliances might be considered to enhance insight by providing views of the world inaccessible to people with their standard-issue senses.

It doesn't feel like we're making much progress on this vision, does it? Except that we really are building an insight machine, it's just made of different parts than I expected. Look around and you'll see it. Our prototype insight machine is made out of people who blog, people who mine data and publish it, people who write online reviews, people who analyze events and the analyses of events, people who garden wikis, people who fact check, people who shine the FOIA light under rocks, people who connect the dots. For an example, look at the Internet coverage of US politics to see the prototype insight machine in action. To name just a few of the pieces: Project Vote Smart, FactCheck.org, opensecrets.org, and The Memory Hole, plus a host of political journalists and bloggers too numerous to call out.

As useful and powerful as these services are, they still only constitute version 0.1 of the insight machine for US politics. This is because there's no way to dynamically string together the data, analyses, reportage, rumors, debunkings, money trails, affiliations, personal linkages, hit pieces, historical records, and random observations into a coherent, persistent view of any given candidate or issue. There's no truth/lies metadata field for the assertions spun out of the various camps, nor is there a reputation standard for the commentators. There's no model for semantic aggregation of related RSS feeds, other than what dedicated/obsessive individuals decide to assemble by hand.

But there could be.

Posted by Gene at October 19, 2004 01:05 AM | TrackBack