Ok, I came to this after reading Drexler's Machines of Creation and then finding out about transhumanists, singularitarians and the whole G-N-R revolution thingie.

On one hand it's pretty convincing as you are reading it. In fact more than convincing: an exhilarating accelerated ride into a future of techno-bliss after the -as someone has put it- Nerd Rapture.

On the other hand, I'm not quite sure that exponential science and technology evolution is such a clear given, no matter how many semi-cuantitative plots you can make about it. Many of the advances he hopes to see in the next few decades might be possible, but not so fast or easily attained. In the particular case of strong AI giving us self-sentient superintelligent machines I'm afraid that too many of his hopes hang from keeping up Moore's Law. But faster hardware won't give us magically better software, and I don't see exponentially better software in any way or shape.

Still one is left with the feeling that it would be quite cool if Kurzweil were after all right, but in a sort of reverse Pascal's Wager is safer to bet against him: I'm not paying if when it's all said and done I'm not immortal.