Don't be too discouraged by the negative comments. Half of ML, optimization algorithms is pulling in valuable data. Companies with data monopolies could help in

Don't be too discouraged by the negative comments. Half of ML, optimization algorithms is pulling in valuable data. Companies with data monopolies could help innovation by sharing it with others more creative than them. Stick diverse data feeds in a neural network and try to predict earnings for companies, energy demand or something else valuable to know in advance? Build off some of the ideas that Google trends/insights offer? Always thought someone should put together a 'bittorrent client for scraping' where you can make large distributed queries as long as you service others.