While filtering for keyphrases is the usual way out, David Gelernter sees hope in exchanging the axis along which the web-babble should be ordered: let's use the time axis (see "Time to start taking the internet seriously" on edge.org). Reminiscent of twitters lifestreams, information would visually flow from future over present to past letting the reader focus on everything in the timewindow she chooses. Aside from the big questionmark (why would such a reshuffling make lifestreams easier to bear?) there is a major misperception underlying all this visionary shabang: the stuff in the web is not information - it is data.
If we stare at all the data we get blind. If we focus on information the web is a much nicer place. Information arises from the analysis of data, from their connection, processing. Of course, a tweet can contain information ("I am going out now"), but the web's information-content is more powerful than that. Tweets on the weather could be accumulated to support weather-forecasts. Chatter on holiday-plans might help airlines organize their resources. Of course, this analysis is already going on - mainly to extract information for targeted marketing. But wouldn't it be nice if the future interface to the web is not a collection of access-paths to multicoloured 'social' networks or chatterboxes but a configurable data-analysis tool that helps pull out the real information? Travel-tips would not originate in some backoffice of an agency with clear commercial interest - they would be the result of your individual correlation of web-babble with weather-information and flight-prices, for example. Updated life. News could be ranked according to clickrates, or coverage, or resonance in a monitored corner of the net that you define. You name it.
Not only would the data-flood of the net be easier digestible, information would again be decoupled from the commercial interest of information-providers. The web could continue as the anarchic place it once was - or it could at least pretend to.