Expected Return under Lognormal Model

I read somewhere (and i'm just typing this out of my head) that for a lognormal model:

Expected return does not change over time because investment return over non-overlapping periods are assumed to be independent of each other, and therefore the model is consistent with weak EMH as knowing past patterns cannot help you predict future ones

Then I also read that:

The lognormal model generates equity prices by directly using past prices, unlike Wilkie's model that involves indirect calculation

The former claims the model does not depend on past information, but the latter says otherwise. Did I read it correctly?