we are now approaching the "second order problems" phase of agents browsing the web. ex:

how do we prevent abuse?
who is responsible for this agent's actions?
what access should it have?

thankfully, i believe these are all solvable problems. exciting times ahead!