Steam engine time for intelligence?

Friend and helpful commenter Jim Cambias comments, on my speculations on whether the highest possible intelligence in any era is getting higher or not, that

it seems to be steam-engine time for intelligence. We're smart, but so are cephalopods, dolphins, crows, parrots, raccoons, and of course the great apes. Humans just got across the finish line first. It's not surprising, really. Darwinian selection puts a kind of "ratchet" on random variations in brainpower, so I'd expect the trend to be generally in the smart direction. Especially since there's no "half a wing" problem with intelligence -- being smarter than the average annelid is just as much an advantage as being able to master language and tool use.

 Is the smartest example of any phylum smarter now than in past eras? How gains in smartness transfer across great extinction boundaries isn't completely obvious. Being smart is no defense against getting hit by an asteroid, or having the seas turn anoxic, or whatever global cause kills almost everything. The smartest don't survive in that case--though those that do may gain some advantage in the disordered environment following the catastrophe. So I'm on the fence about whether you'd always expect the next era to have higher peak smartness than the previous one.  If you drop primates out of the analysis as an outlier, are the remaining peak smartnesses higher than they were in the Cretaceous?

Are we really in a world ripe with smartness? If so, is there any reason for it aside from ratchet effects?