A supreme joy. I can dip into this book anytime, and gain something from what I read (even if I can't entirely grasp it).
"The Tale of Happiton" is one of the best pieces regarding nuclear disarmament I have read. ( )

This book is a collection of essays that would be fun to read if you were not knowledgeable about maths or science.

I gave it three stars, because it does contain a fascinating essay about how to think about permutation systems, but again that only helps if you don't know a lot about mathematics.

If you are interested in the subject of the book, then you probably already know a lot about mathematics and science, so I'm not sure who the book is intended for. ( )

Not as good as Hofstedter's incomparable "Goedel Escher Bach", with which there is slight overlap. A few essays (for example, the Rubik's Cube one) did not much interest me, but there is plenty here for everyone. Very worthwhile and highly recommended for both general reader and scientist. ( )

This is a collection of the columns Douglas Hofstadter wrote for Scientific American when he took over Martin Gardner's regular "Mathematical Games" column. (The name of this book and of Hofstadter's column is an anagram of the name of Gardner's column.) In my opinion, a few of the highlights of this book are:

(1) On pages 37-41, at the end of a chapter on self-referential sentences (i.e., sentences that refer to themselves), Hofstadter presents a short story by David Moser entitled "This is the Title of This Story, Which is Also Found Several Times in the Story Itself" that is made up of self-referential sentences. I thought this story was hilarious in a Monty-Pythonesque sort of way, but your mileage may vary.

(2) Chapter 29 is a fascinating discussion of "Prisoner's Dilemma Computer Tournaments". The Prisoner's Dilemma is a scenario in which two individuals each make (in secret) a decision to cooperate with the other individual or to "defect" instead. If you both cooperate, you both get rewarded. If one cooperates and the other defects, the defector gets a higher reward and the cooperator receives a penalty. If both
defect, nothing happens. This scenario gets its name from the idea of two suspects being interrogated for a crime for which the police have a moderate amount of circumstantial evidence implicating the pair. Should a suspect rat his accomplice out or keep quiet? This chapter discusses tournaments in which individuals write computer programs to participate in a succession of prisoner's dilemma games with other programs. One of the more successful programs was a fairly simple one called TIT FOR TAT, which would cooperate on its first encounter with another program and then, on all subsequent encounters with that program, would do what the other program did on their immediately preceding encounter.

(3) Chapter 31, entitled "Irrationality is the Square Root of All Evil", reports on the "Luring Lottery" that Hofstadter set up for readers of his column. Hofstadter offered a cash prize to be awarded to one entrant in this contest. Entrants would each send in a number on a postcard which could be thought of as the quantity of lottery tickets being requested by that person. In other words, all else being equal, his/her odds of winning the prize would be proportional to the number he/she submitted. The catch is that the size of the prize would be $1,000,000/W, where W is the sum of all the numbers submitted, so it would have been in the interest of the group of entrants as a whole not to submit outrageously large numbers. This was something of a cooperate-or-defect dilemma like that mentioned in (2), and as might be guessed many entrants defected (some quite spectacularly). Hofstadter was obviously depressed by the results of his contest, feeling that it was a metaphor for many of society's problems. He announced his resignation from his column in the issue of Scientific American in which this chapter originally appeared. (The timing may have been coincidental; I don't recall.) ( )

Trite, boring.

Update a day later:
Covers topics similar to those covered in Godel, Escher, Bach, but not nearly in as entertaining a manner, the intellectual depth is lame, the Lisp stuff is boring to someone who already is familiar with at least one computer language, which presumably is more than half of the audience for any Hofstadter book, the Prisoner's Dilemma material covers ground that has been well-trod over the decades by many, many, many people - though perhaps this is something that a non-economist reader wouldn't know - the attempts to tie the subject matter of many of the chapters to nuclear war at the end are thunderingly unsubtle (maybe this was less noticeable when these articles originally appeared spread over months or years), the material on creativity is an attempt to pad an idea fit for five pages over tens of pages. Etc.

Some authors write several great books. Some authors are one-book authors. Hofstadter is the latter. Read GEB. Don't read anything else by this guy, certainly not Metamagical Themas or the execrable I am a Strange Loop. ( )