Sure. I mentioned two things, but let me go into them in greater detail.
For the purposes of clarity, a troll is a person and a bot is a robot, but they work together. The trolls control the bots. There are actually what they call “cyborgs”. That's when trolls manipulate bot armies and redirect them.
The point I was making earlier was that the troll wants to get paid. Make it harder for them to get paid. We've already got the system in place through the sanctions. You have proof of concept there. No one wakes up in the morning with a sense of civic duty to run a troll farm. They want to get a paycheque. Make it harder for them to get that paycheque and that will lower the amount of amplification that takes place.
The second thing we talked about is that there's no widely received definition of what is a “bad” bot. We have to work with allied states to determine what that is. Then we have to put in place a mechanism, through collective action, to both leverage those sanctions and then go after the technical architecture. Work with the Internet governance agencies. There's something called ICANN, but I won't get into the details of that right here. Go after the technical architecture and make sure that the big farms don't resolve.