This book is about the creation of super-intelligent thinking machines. The first section presents
the overall case that intelligent thinking machines are not only possible but inevitable.
Then I present a model of capabilities that a system needs in order to appear intelligent, and the
behaviors we can expect from a system built following that model. The details of the explanation
are a bit more technical but I have endeavored to include examples which will make the process
clear.The final section extrapolates the behaviors that result from a system created along the
lines of the model of Section II so we can reach conclusions about what such machines will be
like and what we might do to coexist with them. It isn't critical to the thesis of this book that the
model be correct in every detail. In fact, any goal-oriented learning system which interacts with
our physical environment is likely to exhibit similar behavior.
What is the point of this book?
•
To show that computers more intelligent than humans are possible.
•
To explain why such computers are inevitable.
•
To argue that machine intelligence will be created sooner than most people think.
•
To demonstrate that, subsequently, vastly more powerful intelligences will be created
only a few decades later.
•
To conclude that such "genius" machines will lead to options and opportunities for how
humans will coexist with (and prepare for) them.
