Actually, you're both right, and you're both also missing another aspect.
First, the 60GXP is IBM's second-generation ATA drive to utilize glass platters. Though it has a handful of advantages, the primary advantage of a glass platter is extremely low thermal expansion, as compared to the aluminum substrates of other drives. Therefore, it is unlikely that the freezing effect, at least on this drive, had anything to do with substrate expansion/contraction. That being said, if a drive was out-of-alignment, a freeze could, under specific circumstances, help get the heads back in alignment.
Second, freezing does help on the IC aspect. Spot-freezing has for a long time been a favorite troubleshooting method when heat seems to be related to device failure. A mass freeze would have the added effect of lasting longer, given you have a larger mass brought down to a given low temperature.
The other aspect is that of stiction. One of the properties of utilizing a glass platter is that they are able to get the surface more smooth than aluminum. While this is great for increasing aerial density, it is bad in the sense that it creates more friction between the head and the platter. This can and does result in a a condition known as stiction, in where the head is quite literally bonded by friction to the disc recording surface. By freezing a drive, you can help reduce this stiction. I've recovered many drives by hitting them on the side, and thus breaking the stiction.
So you're both right, and you're both wrong. Nah.