Why the internet of things is noisy - and how to solve the problem.

Dalton Caldwell

Dalton Caldwell, Co-Founder & CEO, App.net

How are we going to cope with all the data that the internet of things will create? Struggling with your in-box and social networks? Things are only getting worse.

How about service that auto-tweet or auto-posts to Facebook using frictionless sharing? What do you do if someone is cluttering your stream? Unfollow them? Ask them to slow down? Maybe - but you want the algorithm to solve that for you. Reading the unfiltered stream is almost impossible now - we're reaching the limits of human cognitive capacity. The FB stream is a great example of this - eliminating noise and redundancy. 

Streams will be useful for processing internet of things data, because you can apply filters. The problem is that if you're advertising-supported, you don't want third party apps displaying your streams - you want to control the appearance and the filtering, so the ads don't disappear. And you probably don't want to host home-automation data. In fact, you're more interested in engagement and repeated use, than you are utility.

The solution might be to unbundle the stream - use APIs and portable data to allow data to be hosted and accessed in many places.  Examples right now are IFTTT (read/write APIs), 3rd-party Twitter clients (UI unbundling) and Flipboard (unbundled stream filters).  All three of these are running into problems because the social networks have business issues with the social networks. 

He created app.net as an example of what an unbundled stream service could be like. If you unbundle the streams - remove them from the commercial interests that drive them to push away third parties - you're driving innovation in streams, in filters and in UIs.