After typing that rant up, I thought more about the problem and I have since become even more concerned that there is an incentive misalignment here, which leads to an unstable equilibrium long term. Rather than focusing on the specific concerns about operational overhead of running a client, I would prefer to focus on the concern of lack of incentives to building a client.
I’m curious what people’s thoughts are on user-agent incentivization? I recognize that it isn’t game theoretically fool proof, though I somewhat suspect that in the real world it may be “good enough” and think it is “notably better than the current situation with no incentives”.