There was a section in there about the development of DES. At some point the researchers at IBM showed their work in progress on DES to the NSA, and the NSA made a few suggestions on improvements, one of them was a different set of S-Boxes (basically a bunch of constants used as part of the cipher.) IBM analyzed the new S-Boxes and could find no weaknesses so they used the ones the NSA recommended.
Over a decade later, differential cryptanalysis was discovered and it turned out that the tweaks the NSA had suggested made DES particularly resistant to differential cryptanalysis.
After doing some googling, it seems like the gap has closed since that book was written. They are still ahead, but not as much as they were in the 80's and 90's. Or maybe that's just what they want us to think ;)