Bionic Systems: Amplify Participation
A small boom of terrific social applications has appeared in the last few years (and there are more are on the way). I find social applications very interesting because they bring personality to a web service. Compared to completely automated systems, social applications reflect the appealing human qualities of passion, perspective and nuance. These attributes can give social applications have a real leg up on completely automated systems.
However, the web community will have a tough time supporting the large entries of social applications. There is simply not enough participants/participation (or attention) to go around. New services that are essentially empty applications that require participants to add content and value will have a harder and harder time. We should expect to see a handful of such services dominate eBay style (where the network effect works its awesome magic). But unless these services can create lock-in the way eBay did with its feedback score, we have seen the fickleness of the crowds also abandon services just as quickly.
So many social applications will have to introduce methods that allow it to exist and thrive with less and less participation. Perhaps this is focusing on narrower niches, using only automation, or using a bionic system.
What is a bionic system?
Bionics is the study of living systems with the intention of applying their principles to the design of engineering systems. So a bionic system would be an “engineering” system that has the principles of living systems. Some examples of bionic systems:
- recommendation systems through collaborative filtering, which have been around for years, take the personal preferences of many individuals in a group and use them to find new recommendations to the single individual,
- link analysis systems (such as Google’s PageRank, Memeorandum, and Technorati) takes the link information on web pages to uncover relevant or popular content,
- photo-recognition systems, such as Riya, which has machines that takes a few human identifications of a face and then uses that information to identify many others. In an interesting bionic system that goes the other way, Amazon’s Mechanical Turk has used humans to tell computers where stores are located in a photo (computers can do wonders with just a little bit of a repeating pattern, but if there is no pattern; it falls apart).
- sentiment analysis: at Biz360 (where I am founder and Chairman), we built a Point-of-View Sentiment engine where we automatically rated news stories and blogs stories as positive, neutral or negative from our client’s point-of-view. Humans rate dozens of articles and the machine learns and rates thousands, even tens of thousands, of articles.
- Content filtering and preference: at Boxxet (where I am founder and CEO), we are working to employ a bionic system to capture a small number of ratings and submissions and amplify it to sort and filter the best content on many subjects, even subjects that may have only a very small community.
In all these cases, the impact of human participation is significantly amplified by the machine. There are downsides to bionic systems; errors can be amplified and machines can make head-scratching decisions (we have all laughed at off-base recommendations). But the upsides are clear: the passion, perspective, nuance and wisdoms of crowds, even a small crowd, is captured and used far beyond the individual contribution. Bionic systems also help cushion the downside of fickleness, fluctuation and distraction—other qualities of the crowd. These benefits will help social applications extend its usefulness and longevity.
Note: I will be talking more about bionic systems Wed, March 8 at the O’Reilly Emerging Technology Conference. See you there. Also expect to see deeper thoughts on bionic systems in the upcoming weeks and months.
