I'll grant you that current (and most proposed capability based) security systems are too complex for the user to manage. However the question on the table is

I'll grant you that current (and most proposed capability based) security systems are too complex for the user to manage. However the question on the table is research. We should be researching security models that don't rely on the user to manage the complexity.