I don’t feel particularly ready for a post at this point, but the previous one has got to 100 comments, so this is one of the quick summaries again — but this one is even shorter than usual.

We are still doing an experimental investigation of multiplicative functions, trying to understand how special they have to be if they have low discrepancy. Ian Martin has produced some beautiful plots of the graphs of partial sums of multiplicative functions generated by various greedy algorithms. See this comment and the ensuing discussion.

Terence Tao has some thoughts about how one might try to reduce to the character-like case.

I came up with a proof strategy that I thought looked promising until I realized that it made predictions that are false for character-like functions such as and . Even if the idea doesn’t solve the problem, I think it may be good for something, so I have written a wiki page about it. Gil has had thoughts of a somewhat similar, but not identical, kind. Here is a related comment of Gil’s, and here are some more amazing plots of Ian’s. (I think we should set up a page on the wiki devoted to these plots and the ideas that led to them.) Regardless of what happens with EDP itself, I think we have some fascinating problems to think about, which can be summed up as, “What is going on with these plots?”