[winchling]

We don't know what Socrates said because it wasn't written down.

He relied on his followers to improve on his ideas and transmit those improved versions, and perhaps they did. However knowledge does get lost (e.g. how to read Cretan Hieroglyphics). It's good to insure against that.

There are advantages to live discussion but it doesn't preclude making occasional records in the form of books. Why not do both?

Contra the article, books do contain knowledge. It's knowledge in the form of 'know that' rather than 'know how to'. Yet if we could somehow obtain science textbooks from the future, we could use those books to recreate advances in knowledge in a much shorter space of time, i.e. we could bridge the gap to know-how. So books do have value beyond entertainment and recording historical facts.