FOM: quote from G"odel on what is interesting in math

[Stephen G Simpson]

Wolfram's talk (at
http://www.stephenwolfram.com/publications/talks/IMS/imstalk.html)
includes pictures of dependency graphs of the theorems of Euclid's
Elements.  And then Wolfram says the following:

   I was showing you this rather messy dependency network. This is the
   picture of the particular theorems Euclid thought were interesting
   that you get from the multiway system that is geometry. It reminds
   me of a quote from G"odel, when asked about mathematics, what the
   role of a mathematician was, his idea was that the role of the
   mathematician was to work out what theorems are interesting. That a
   machine can generate all the possible theorems, but the role of a
   mathematician is to work out what theorems should be considered
   interesting.

Does anybody know where to find this quote from G"odel?

-- Steve