Eli Bendersky's website D.E. Knuth promised to send a $2.56 cheque to anyone who finds a mistake in TAOCP, and so he does (you can see scanned cheques on some geeks' websites).
So, I found a mistake in the math. In volume I, the calculations on sparse matrices. I sent Knuth an email, explaining the error. About 3 weeks later, I got an email from his secretary asking for my home address to "send me something". I was exhilirated, as you can imagine, and already had time to brag about it (thankfully only to my wife). Yesterday the letter came, and guess what ?
_I_ made a mistake in my calculations, which Mr. Knuth kindly pointed out with a pencil on a printed copy of my email. Well, at least I have a letter from Knuth ! :-)
Moral for today: 20*30 equals 600, not 60 (yeah, I know, I know, just shoot me ...)