prosaic alignment, corrigibilty, interpretable NNs, agent foundations, restricting AGI research, alignable AGI paradigms, do it all https://twitter.com/tetraspacewest/status/1410184795200667648?s=21