I am going to try to field test @SureshVenkat suggestion and ask some easy questions in chat! Hopefully we will be able to generate answers.
In the comments to my question on approximating Ramsey numbers @TsuyoshiIto pointed out that there will be difficulties with my approach via unary encoding, because the problem becomes trivally in P/poly, and it is not known if NP is in P/poly.
So the question: is it believed that NP is in P/poly? What is the intuition or evidence for/against this belief? Are there other examples where we have interesting inclusions for non-advice classes in advice classes? What about P in LOG/poly? Or does that not even make sense?
To be honest, I only saw this because you posted the 'field test' note :). The complexity zoo has some interesting implications for NP being contained in P/poly: qwiki.stanford.edu/index.php/Complexity_Zoo:P#ppoly - a collapse does happen, but it doesn't appear to be as catastrophic as some others. IANACT though
