The Navier-Stokes thing? They had hundreds of segmented groups of 10000 agents running trying to solve that. I can't imagine the expense. For what little publicity it got, it wasn't worth it. There are also reports they cheated by using data from a couple human researchers, but I don't think that one's true.