>Multiple cities have already implemented trial programs to replace police services with mental health and EMT professionals to wild success. That's chill, but

>Multiple cities have already implemented trial programs to replace police services with mental health and EMT professionals to wild success. That's chill, but it doesn't seem nearly as radical as the "abolish the police" slogan? What fraction of police calls have been replaced with mental health/EMT calls in these cities? Suppose someone is suspected of using a counterfeit $20 bill -- does Denver now call a mental health or EMT professional for that person? >Proponents of capitalism seem to have never encountered Goodhart's law. [1] The severity of Goodhart's law depends on the mismatch between what you want and how you measure it. If you're able to measure what you want very accurately, Goodhart's law is not as much of an issue. Goodhart's law doesn't just apply to capitalism, it applies in many social domains -- including the political domain. The grandparent comment states: >There are many broken feedback loops right now from the grand (lobbying and campaign finance) to the minuscule (niche price fixing, bribery, police-ran protection rackets), all compounding unfavorably, all seemingly booming. Lobbying, campaign finance, price fixing, bribery, and police-ran protection rackets only relate to free-enterprise competition in the vaguest possible sense ("they all involve money"). I brought up the logo police idea for discussion because I think it's possible that it could align feedback loops better than the status quo. I invite you and everyone else to shoot holes in the idea, and guess how specifically it could fail. It could be that it fails in predictable ways, yet is still an improvement on the status quo. One of my hot takes is that it's good to rotate your metric on a regular basis, in order to flush out actors who are purely optimizing for the current metric and nothing else. Unfortunately, any beneficiary of the status quo has a strong incentive to maintain it even if it's flawed, and random members of society only have a weak incentive to fix it. That creates friction that prevents rotating metrics. From an incentive engineering perspective, this meta problem seems like the highest-leverage thing to solve. If we could fix that public goods problem, we could keep tweaking incentives on a regular basis and hopefully outrun Goodhart's Law faster and better than we are currently able to do. I wonder if mechanisms for solving public goods problems, like dominant assurance contracts, could be helpful here. Ultimately standard left/right debates sound pretty tired to my ears. In an ideal world, "incentive engineering" would be just as rich of a discussion as "software engineering" is today. The current situation is like if every HN post about software engineering was either an argument for why static typing sucks, or an argument for why static typing rocks. There's a much bigger world out there people. The heroes vs villains frame that people want to apply to every problem is often an impoverished one, IMO.