Abstract

This week we read the “Goals vs Utility Functions” series by Rohin Shah. This is the posts under “Ambitious Value Learning” at https://www.lesswrong.com/s/4dHMdK5TLN6xcqtyc . You should read them before the session.

Optional reading: “Coherent Behaviour in the real world is an incoherent concept” by Richard Ngo.

These writings focus on the arguments that usually justify the premise that a general AI will necessarily optimize a long-term, explicit, simple goal. The authors find them to be insufficient, and in the end propose that perhaps a better approach to AGI safety is to construct agents without long-term goals of this kind.

As usual, there will be free pizza. The first half hour is for stragglers to finish reading.

Invite your friends to join the mailing list (https://lists.cam.ac.uk/mailman/listinfo/eng-safe-ai), the Facebook group (https://www.facebook.com/groups/1070763633063871) or the talks.cam page (https://talks.cam.ac.uk/show/index/80932). Details about the next meeting, the week’s topic and other events will be advertised in these places.