The reward hypothesis—that all goals and purposes can be understood as maximizing the expected value of cumulative scalar reward—is sufficient for generality and does not require modifications like multiple objectives, constraints, or risk sensitivity to achieve general intelligence

factualpending

Speaker

Richard Sutton

Evidence Quote

“all of what we mean by goals and purposes can be well thought of as the maximization of the expected value of the cumulative sum of scalar reward...you don't add generality by adding multiple objectives or risk sensitivity”

Source

Rich Sutton, The OaK Architecture: A Vision of SuperIntelligence from Experience - RLC 2025— Amii
Created: 8/11/2026, 7:21:57 AM

Your notes

Loading notes...