I've been thinking about that aphorism lately and quantifying it in
information-theoretic terms. To that end, I stole some Python from Rosetta
Code, calculated the entropy of every word in /usr/share/dict/words,
averaged over those 100,000+ words, then multiplied that by 1,000 to get the
size of a picture in bits:
A picture is worth ~1000 × 2.70, or ~2702.30 bits
Accordingly, here are almost 231 pictures of Claude Shannon:
