I am particularly interested in understanding how valuable a postdoc conducting AI safety (eg, not capabilities) evaluation work would be; capable of supervising PhDs etc. But you may as well explain what models we should use to find that value, and that of other positions. The goal here is to understand how to gauge career choices.
EDIT: for comparison purposes, should we perhaps agree on using donation money amounts?
[Question] What is the counterfactual value of different AI Safety professionals?
I am particularly interested in understanding how valuable a postdoc conducting AI safety (eg, not capabilities) evaluation work would be; capable of supervising PhDs etc. But you may as well explain what models we should use to find that value, and that of other positions. The goal here is to understand how to gauge career choices.
EDIT: for comparison purposes, should we perhaps agree on using donation money amounts?