David Scott Krueger comments on False assumptions and leaky abstractions in machine learning and AI safety

[David Scott Krueger]

IIUC, yes, that’s basically what I was trying to say about embedded agency.