> It's better to know you just need the right compiler than the think the issue is something else as you can then work on building the right compiler.
You completely misunderstand the sufficiently smart compiler objection. It is about dodging real problems by proposing magic solutions.
You can't make C/C++ a safe language "just" by writing a sufficiently smart compiler. It requires a new approach and perhaps even a new language like Rust.
> It touches on proposed solutions such as street-level development votes.
That is not a generic solution to the vetocracy problem. That is a specific identification of a group of people that the author of the article believes should be endowed with making a particular decision. There is no general framework proposed for selecting such a group of people in the general case.
Let's consider some possible general mechanisms. The "simple" problem that we "just" have to solve to make the authors otherwise perfect solution work.
We could endow that responsibility on a single individual, such as the author himself. He feels he made a good decision in the street-level case, maybe he can do the same for other cases, like mining rights or cross-country oil pipelines. So whenever the question comes up as to who is affected by such a thing we go to the author of the article and ask him to tell us. Kind of like a dictator or even a King.
Or we put together a committee and they decide between them who is impacted. Kind of like an oligarchy. Or maybe we all get a vote on who is impacted. A democracy to decide who is impacted and who gets the right to vote. It's recursive! Or divide and conquer!
It's starting to sound like picking the mechanism to determine who is impacted by a particular change is a very similar kind of activity to existing political mechanisms. That is suspicious. Maybe it isn't an easy problem to solve after all.