There is a new thread for polymath5. There is also this post. There is a sequence of 1124
values of plus or minus one which has discrepancy two. It is exhibiting an interesting pattern which may be generalized to yield an infinite sequence which has sublogarithmic discrepancy.