Stuart_Armstrong comments on Inverse reinforcement learning on self, pre-ontology-change

[Stuart_Armstrong]

I’m hoping there would be at least some gain that is an extension of the old preferences and not just a direct translation into the old ontology.