Real humans have eigenvectors that do not line up with visible social identity markers. Wish there was a way to model humans as a behavioral matrix and actually compute this difference.
Conversation
Replying to
if you could encode the behaviour digitally you could run it through a deep neural network you could map the latest space of the hidden layers to find interesting clusters/modes of behaviour:
1
6
Replying to
That’s what the Cambridge Analytica scandal was and how we got Trump and Brexit


