Isaac Asimov enabled many of the science fiction books I've grown to love, so I had to try his most famous classic. It's barely science fiction - he wonders what would happen if social science advanced enough that it became predictive with enough complex math so that a "psychohistorian" could know what would happen in the future with only a few percentage points margin of error. The book is essentially five successive short stories exploring if Hari Seldon (a psychohistorian) made correct predictions about the downfall of a Galactic Empire. We follow various people that think he has, and find out how those predictions come to fruition.
I enjoyed the book for the most part, but could only do so in the context in which it was written - the 50s. Really interesting ideas, and the writing's actually pretty good. The sequels look very uninteresting to me, but I'm glad I read this. It's not for everyone. He doesn't describe very much, but then he wasn't really trying to in this book. It was mainly a way to explore this predictive history idea, and I'm sure in the 50s, that was a very alluring idea. I'd be curious about other titles of his that would be interesting in a more modern context.