Decompose Web pages into rectangles, then match the rectangles to identify similar future pages. CNN would be required as would text recog with rectangles with different text breaking points.
but only the aggregate data for each user, then allow collections of companies to bid on collections of users' data ("data unions" akin to labor unions--data generators of the world unite!)
the problem is that companies have aggregations of users
but users have no way to market their aggregations of data across companies
could operate this like data explorers, companies that give their data are allowed access to the aggregate data (probably wouldn't be very appealing for companies w/ lots of data tho, e.g. amazon)
value add would then be matching users across companies
put the control in the hands of the user though then. allow them to get kickbacks for sharing aggregate data across companies. this would allow access to amazon data by other companies via the user
Give users affinity analysis for the whole internet. Will need search/website history for this.
Ie users already are annotating the net, just differently than I imagined before.
keywords: web annotation
an example: google.com/history
Email: Shatter the network
Offer a decentralized social network (consisting of someone's web presence? or is this a different idea?)
How could Watson help with this?
Is this page worthless? Or is there just stuff I don't want to put here b/c it's public?