The alternative hypothesis - that the brain has independently evolved twice - is in my opinion even more interesting. If true, it would seem to have implications for estimates of how easy it is for life to evolve intelligence.

For the same reason why, say, version 7 of the WRT-54G router had half the memory that version 2 had.

It's expensive (money or calories) and the extra functionality doesn't help in getting more out there (selling or reproduction).

Key point: evolution does not necessarily working towards more complexity or a specific endpoint. Rather, the unique selection pressures that exist when organisms attempt to reproduce vary, as do the particular genetic changes that happen due to any number of factors (mutation, viruses, gamma rays, copy errors, bacterial absorption, whatever). Each generation may face similar or different pressures. Certain mutations may be more or less likely. In some cases, the mutations that survive may be pure random chance. In others, only a specific mutation may have saved the organism's particular genes from extinction. The future organisms and genes that emerge from this chaotic process are unpredictable. Sometimes those organisms will lose features, sometimes gain, sometimes have weird changes that don't appear to make sense. All that matters is: do the organisms continue to reproduce successfully generation after generation?

Plants have also survived for hundreds of millions of years with no brains.

So if someone does not use his brain for long he/she becomes demeted becuse he does not need a brain?

what?! no joke about our bosses? oh right, you're all your own boss here. :)

Another example of where more brains doesn't necessarily mean evolutionary advantage is modern humans. It can be seen all over the world that the less intelligent procreate more.

The intro to Idiocracy makes the point well: https://www.youtube.com/watch?v=hL1-340ODCM

It's kind of a joke, but also really not. It's frightening.