> capable of accessing the sum total of human knowledge
No. Lots of knowledge is still behind paywalls or not yet digitized. Some models have been trained on books that we cannot search or download.
Plenty has already been lost due to being buried in search, removed for lack of interest, or simplified so far as to be too generalized.
Plenty has already been lost due to being buried in search, removed for lack of interest, or simplified so far as to be too generalized.