Fair enough, Thiel, but aren't your doomsayer projections also a direct form of thought experiment, based on extrapolation just the same?
He goes on to outline what he thinks are three defining markets (or distortions thereof) that could utterly define the future ("China", "Internet" and "hedge funds"), and claims that the markets are woefully out of alignment with reality. I think we have to take his view with a large grain of salt here, considering his bias: he's a global macro hedge fund manager who also invests in Internet startups.
I think he's probably right about a lot of the distortion and possibly right about some of the importance of these markets, but he doesn't even mention any other possible contenders (energy industry transformation? nanotech? biotech? government competition?). I think it's a mistake to assume that any one or three factors will determine the mid-term future when there are literally trillions of factors in play. Then the logical conclusion is that we can't really know or predict anything about the mid-term future, which may be largely right, but I think there is possibility to predict with narrow foci, like "What will happen in transistor design?" or "What will happen in nuclear reactor design?"; though those too will always be hampered by revolutions that we never see coming.
Overall, I hope Thiel is wrong about a lot for the reasons that he so adeptly explains: there's not much else to hope for if the alternative to what you want is total destruction and nonexistence.