I agree Netflix is attempting to buy a new (hopefully better) solution. But if you're implying someone is going to come up with a solution, not based on any previous research, and that is going to beat the Netflix solution (which you assert *is* based on the best previous research), then, good luck with that. Maybe I'll eat my words on that statement... we'll see.
I think the last few posts have a good point: promise big money and get results. It works well for engineering where there isn't a general solution to a general problem. But I want to bring the thread back to my original argument: It's. Been. Done. You think the teams in the lead just dreamed up their solutions from scratch? I'm willing to bet they are all using algorithms that grew out of university research and are just tweaking them.
You miss my point: it's been done. There are probably a dozen algorithms that have been developed with sound scientific methods and subjected to peer criticism to cover this sort of problem. After that it's just a matter of implementation and tweaking. Look at the top teams, they're all at universities or research labs. Those aren't just simple down-home code geeks, my friend, those people are (were) grad students.
A million dollars? This is what happens when business people dabble in science. Artificial Intelligence grad students and professors have been studying these kinds of problems for decades. Netflix could have saved a boatload of money by throwing some cash at a university with an established AI group and asking them to research the current state-of-the-art. The only reason to put up that kind of money is to generate publicity, and I'm not really sure that worked.
Slashdot Mirror
I agree Netflix is attempting to buy a new (hopefully better) solution. But if you're implying someone is going to come up with a solution, not based on any previous research, and that is going to beat the Netflix solution (which you assert *is* based on the best previous research), then, good luck with that. Maybe I'll eat my words on that statement... we'll see.
I think the last few posts have a good point: promise big money and get results. It works well for engineering where there isn't a general solution to a general problem. But I want to bring the thread back to my original argument: It's. Been. Done. You think the teams in the lead just dreamed up their solutions from scratch? I'm willing to bet they are all using algorithms that grew out of university research and are just tweaking them.
You miss my point: it's been done. There are probably a dozen algorithms that have been developed with sound scientific methods and subjected to peer criticism to cover this sort of problem. After that it's just a matter of implementation and tweaking. Look at the top teams, they're all at universities or research labs. Those aren't just simple down-home code geeks, my friend, those people are (were) grad students.
A million dollars? This is what happens when business people dabble in science. Artificial Intelligence grad students and professors have been studying these kinds of problems for decades. Netflix could have saved a boatload of money by throwing some cash at a university with an established AI group and asking them to research the current state-of-the-art. The only reason to put up that kind of money is to generate publicity, and I'm not really sure that worked.