Well, you would run into many of the same decisions we already face about how much to limit automated uploads of data if you wanted to turn it into a live programming platform. You can certainly already use DBpedia and Wikidata to get datasets for many cool demonstrations of functional programming though. Yes, I suspect we are just at the learning to walk stage of programming in the big picture. My favorite examples of AI these days are when computers do large mathematical optimization tasks. I was most impressed by a paper last year that optimized the placement and configuration of coal power plants and more farmland to reduce transport related CO2 emissions by 50% for the entire US. The paper was called "Nationwide energy supply chain analysis for hybrid feedstock processes with significant CO2 emissions reduction". A free early version was published here: http://www.nt.ntnu.no/users/skoge/prost/proceedings/cpc8-focapo-2012/data/papers/092.pdf And to think how nice it would be if the customized optimization techniques they developed were merged into the code associated with those Wikipedia articles for everyone to easily use. The reason that task impresses me so much is that if a computer at Pixar draws a nice picture it is just matching what the artists could already partially see in their heads and if Siri on the iPhone tells me a good restaurant to visit it is just doing what a person that lives in the area could do, but if a computer redesigns the entire energy infrastructure for a country I have no idea what the solution will look like in advance. There is a lot of smart information out there if people are willing to look for it. How can the singularity get them to stop listening to the bad information? I think things like Wikipedia are definitely helping us all get gradually smarter though, so I'm optimistic.