See Robin's paper on this:

http://hanson.gmu.edu/hardstep.pdf

If a step is extremely hard (and thus astronomically unlikely to occur in the lifespan of a planet) then we should expect to see it taking a length of time comparable to typical planetary lifespans divided by the number of steps.

The last two steps occurred suspiciously quickly to be super-hard, and the independent clusters of mammalian braininess in cetaceans and primates [EDIT: (and birds, to a lesser extent)] make the third step questionable too.

Super-difficult life and nervous systems look very plausible to me. A special difficulty involved in the formation of mammals (a feature that predisposed to the later development of intelligent hominids) seems less plausible but not very implausible.