How likely is intelligence?

On The Big Questions, Steve Landsburg addresses a perennial SF question:  how likely is evolution to evolve? He quotes a review by an astroscientist Charley Lineweaver of a book called Intelligent Life in the Universe, where Lineweaver eloquently denounces what he calls the "Planet of the Apes Hypothesis", or convergentism.

In the movies "Planet of the Apes", humans are wiped out, and various great apes evolve toward intelligence, making it seem inevitable that intelligence will evolve.

Lineweaver cites what he calls five natural experiments whose result oppose this conclusion:  the island continents of South America, Australia, North America, Madagascar, and India. Intelligence did not evolve on any of them.  There is nothing convergent about the evolution of intelligence, nothing inevitable.

Let's leave aside the fact that "Planet of the Apes" is more satire than investigation of evolutionary process (talking animals, like Jonathan Swift's intelligent Houyhnhnms, are a staple of satire), and consider what might or might not be convergent in evolution.

Certain things seem to be widely applicable.  Eyes, for example.  There are so many eyes in the world, working so many different ways, that you can say evolution converges on them.  Wings too.  Birds, bats, insects...the damn things are so useful, that they often evolve.  Not everything needs them, so not everything has them.

Intelligence...not so much.  It's clearly not as useful a gadget as eyes or wings. But, leaving various hominins aside, has the outer edge of intelligence been pushing upward over time?  Hominins are a statistical outlier, but can we say that the most intelligent species on the planet in, say, the Oligocene was more intelligent than the most intelligent species of the Jurassic, which was more intelligent than the most intelligent species of the Ordovician?

Hard enough to tell what body parts those things had, much less what their behavior must have been like. From my limited knowledge, I'd say that Jurassic beats Ordovician, but it's harder to prove that Oligocene beats Jurassic.  If there was another mass extinction, including us, would the next go-round create an even more intelligent species, or are we just a fluke?

Can anyone think of a way to confirm or deny the hypothesis that the "smartest thing on the planet" has been getting smarter through time?