Is it possible to ground ethics in terms of itself without falling into circular reasoning or infinite regress? Scott Aaronson explains how MIT graduate student Tyler Singer-Clark is taking the first steps towards researching this possibility and explores this fascinating new approach that already has early results from studying small toy models that they constructed. This is potentially relevant to research questions that MIRI cares about in the long run so I look forward to others refining and extending these early initial results.
Grounding Morality in Morality
Louie Helm is a Machine Learning Engineer at Fathom Computing