Shut this thread down and move it to Off-Topic. Its not happening. Amazon has thousands of on call employees working ridiculous hours until they burn out and move on with their lives. Their recommendation algorithm is a recommendation, it doesn't shop for you, and the moment it did you would have thousands of pissed off customers. Netflix paid a million dollar bounty to someone who could create a better algorithm for better movie recommendations. What you are suggesting is a multi-million effort in which the solution is fuzzy and not very well defined - a scenario which guarantees a low probability for success. It will never happen. Ever. Move on.
The same exact arguments were used when Bitcoin was first discussed. When Bitshares was first being discussed I saw the same arguments again about how it's just not possible and we still see people who believe in PoW saying DPoS is vulnerable to the "nothing at stake" vulnerability.
Conditional preference networks are a topic you should Google. If you Google that topic and the topic of Computational Voting and still think what I'm talking about is unrealistic then that is fine. I don't think it's fine however to suppress the topic and throw it into the off topic forum. Why shouldn't we at least discuss it philosophically? Btw all my references are in my posts so you can track my line of thought back to the sources.
Reference
http://www.amazon.com/Working-Preferences-Less-Cognitive-Technologies/dp/3642172792
http://crpit.com/confpapers/CRPITV70Truyen.pdf
Yes, its perfectly fine to have a philosophical discussion. I prefer to remain pragmatic. Luckybit, I think you have a lot of great ideas, but this one falls far into the realm of fantasy - at lease to the point where it will ever be implemented by the bitshares community. My own personal recommendation (take it for whatever its worth), is that before you spend another precious brain-cycle determining what you want this to look like, you first ask yourself the following:
Who is going to build it?
How much is it going to cost?
How many hours will it take to build?
Where did you get your hours from, did you break it down into manageable tasks?
What are the dependencies? (hint Turing complete scripting language is one)
How long will those dependencies take to be completed?
How will you measure the success of your solution?
How will you test this solution?
How will you maintain it?
Does this solution only need to be written once or perpetually updated in an iterative process?
How can you make it so it can't be gamed?
Who is going to pay for it?
What will it cost?
What will it cost?
What will it cost?
If the community pays for this solution, what will they have to give up in its place?
Will this be the best way to spend the communities money?
Could the community ever be convinced that this is the best way to spend its money (Me: no way in hell you'll ever convince me)
Will the community see ROI on this implementation? Is that ROI measurable?
TLDR - this is a science-fiction, meant to be discussed over beers or in Off-Topic forum sections. Cool in theory, completely unobtainable in reality for this community.