Are small sets explicitly small?

I aim to say a few words on the Talagrand's 'Small Set' problem. The idea comes when we do wonder how to bound an Empirical Process from above. One reason is of Gaussian type which is deeply understood via Talagrand's majorizing measure theorem. The other one is simply that Empirical Process can be small because its positive part can be very small. Last year Talagrand studied in depth possible ways to solve the problem (Positivity Conjecture) and I think it is worth learining what he can prove on it.