We don’t know this. I think a lot of people expect that there’s an O(n^(2+eps)) algorithm for any epsilon > 0, but that’s conjecture; we might already be at the

We don’t know this. I think a lot of people expect that there’s an O(n^(2+eps)) algorithm for any epsilon > 0, but that’s conjecture; we might already be at the optimum.