AIAI's vast knowledge versus learning from a few examples
MTSlive quotes a Stanford professor arguing that impressive AI capabilities can obscure a weakness: learning from just a couple of examples.
TLDR
In remarks shared by MTSlive, a Stanford professor distinguishes drawing on 2,000-plus years of human knowledge from discovering knowledge as humans do. He contrasts people learning from a couple of examples with current AI being given 10,000, arguing that machine learning remains poor compared with human learning. He calls learning speed “the secret of human intelligence.”
Combined views
15.2K
1 Source, first seen 1d ago
AI's vast knowledge versus learning from a few examples
MTSlive quotes a Stanford professor arguing that impressive AI capabilities can obscure a weakness: learning from just a couple of examples.
TLDR
In remarks shared by MTSlive, a Stanford professor distinguishes drawing on 2,000-plus years of human knowledge from discovering knowledge as humans do. He contrasts people learning from a couple of examples with current AI being given 10,000, arguing that machine learning remains poor compared with human learning. He calls learning speed “the secret of human intelligence.”