Closing the System in AI

This is a fairly abstract lecture that goes to the core of the potential and the limitations of artificial agents. Its title indicates what we do all the time in AI: closing the system. The consequences of this are eluded in this lecture but not addressed.