I won't rehash everything everybody already knows about the end of the "free lunch" of increasing CPU speeds, just point this curiosity. Being fair on S.H., he was just quoting projections from Intel. This book is from 2001, not such a long time compared to the 3-4 years of design for each CPU generation; so it's funny that Intel wouldn't know better - or perhaps they did, but didn't want to tell us. In the same chapter, Hawking states that future computers would have to be more parallel if they wanted to match our brain's intelligence. IMHO he was an optimist, because we didn't yet make ANY substantial progress towards real understanding of the human mind - and one cannot program in a computer a system that's not understood. But if this situation ever changes (and I'll put serious money that Hawking's search for the Great Unification Theory will end much sooner than that), parallelism will obviously play a major role. But this doesn't matter, even the 2010 version of Notepad will have to be highly concurrent to keep up with the featuritis escalation of our industry. So, grab yourself a good book on concurrency if you plan to be worth anything as a programmer when massively multicore architectures become mainstream even in the desktop, which won't take many years...