It was a good line also because it referred back to a point earlier in the text
Indeed such a good one when you really think about it, but it's above most readers heads.
So the analogy with the Krein–Milman theorem would be this: Any proof is (approximately, at least) a combination of the extreme proofs of the same result?
Now I have to go read the paper referred to.
I don't think it was meant so literally. just an interesting way of calling someone a unique mathematician