Full threadscott-smith_us·Is there significantly more to this than just tokenizing a document's words and then ZIP-ing the stream of tokens?View on HN