Thursday, January 8, 2009

The Emperors New Mind review(ish)

Well, I finished reading this about a week ago now and most of the information I'm going to absorb from the first reading has just about sunk in now. So I think it's time for a quick review of what I thought:

Firstly, some background on the book for those who haven't come across it previously. It is a popular science book written by Roger Penrose and details his (fairly speculative) arguments about why a computer can't truly think. Rather than the usual approach of arguing with just prose and hyperbole, Penrose instead gives a brief tour of the maths and physics involved in both computing (including the concepts of Algorithms and Computablity) and of the mind, with brief detours into quantum mechanics and relativity.

With regards to the actual content, Penrose makes very clear that the ideas he is conveying are frequently very speculative and are not in line with the general consensus with the scientific community at large. This doesn't stop his arguments being compelling and interesting,although I don't fully agree with all of the arguments. The main thrust of the argument is that the mind is seemingly capable of doing tasks that as far as can be ascertained are non-algorithmic and thus can't be performed by a computer. He ensures that the reader is fully aware of what is involved in algorithms and computablity, which lends weight to his arguments. He also attempts to explain his view on what is required for being conscious and able to think with a highly speculative detour into the possiblities of a theory of Quantum Gravity (based on the inconsistencies in quantum mechanics) as pertains to the brain.

To wrap up (before I ramble on for hours) the book is informative, entertaining and a good read. For someone with an interest in the issues involved in AI, this is a good start for a more scientific discussion of the topic. If you have a desire to expand your knowledge of general maths and physics then this book is also a good starting point (with a possibility to move onto a book like 'Road to Reality' afterwards, which gives a much more in depth look at the maths and physics of our present understanding of the universe). If you aren't fond of mathematical treatments of topics then this book probably isn't a good choice though. I personally enjoyed reading it a lot and would recommend it. Not bad for a book from the tail end of the 1980's :)

No comments: