I agree with most of what you say, except for this bit: > Too many predatory entities standing by and watching out for any such accumulation of authority/power

I agree with most of what you say, except for this bit: > Too many predatory entities standing by and watching out for any such accumulation of authority/power and standing by to take it over and drain it of all its value. > Too much incentive to cheat and not enough that can be done about it. That assumes that people will stand by and watch the too many predatory entities and do nothing, and under "current assumptions", that's exactly what will happen. But current assumptions can be broken. For example, there could be a revolution and certain/a few/some governments may lose the monopoly in violence. Such an idea, alien as it looks to us now, was the way of things through most of history and it is the way of things--sadly--in too many places still today. If citizens have a relatively efficient mechanism to keep the certification authority trustable, then the certification authority may solve 70% of the problem. Web of trusts are also a possibility, but yes they will be regional and in need of strong baking by face-to-face meetings. Probably all trust mechanisms will like that: local and patchy. I've got to add my own note of pessimism: relatively trivial exchanges of information won't survive the AI age. Sure, business people will find ways of negotiating prices for goods and services across borders, scientists will meet at conferences and exchange data and results, and the Interpol will talk to trusted authorities in the member states using cryptographically-sound channels and base agreements and methodologies. But the general public will trust very little of what they see in the Internet. We will disconnect.