Yes, and I expect that agenda to be significantly more tractable for designing systems that are effective in the real world than normative agendas.
All that said, my preference is to draw the distinction around “models of intelligent systems” not “models of agency” as agents aren’t the only type of intelligent systems that matter (foundation models aren’t well described as agents).
Yes, and I expect that agenda to be significantly more tractable for designing systems that are effective in the real world than normative agendas.
All that said, my preference is to draw the distinction around “models of intelligent systems” not “models of agency” as agents aren’t the only type of intelligent systems that matter (foundation models aren’t well described as agents).