> I'd argue that you should take whichever option is best for society, i.e. costs it the least. Why?
> a) Societies cannot be comprised of mostly (weighting by utility) negative-externality people for very long.
> b) Everyone else is better off by eliminating such people, thus they are motivated to do so in whichever way is cheapest.
I can't imagine what you might mean by negative-externality people, but whatever it might be let me inform you there are lots of countries/governments/societies in the world that aren't doing so hot, and they've been doing not so hot for quite some time. Is this some kind of quasi-rational-market hypothesis for societies? Nowhere is this true. Why do people stay in abusive relationships? Why did Black people continue to live in States that practiced segregation? Why do people still eat unhealthy food, or smoke, or drink?
> If it were cheaper to just execute all criminals, or commit horrific human rights disasters to make prisons cheap to run, that's what society should do. Historically, that's what societies have done.
I think relying on the actions of governments who knew almost nothing (Earth is flat, what is air, diseases are punishment from God, the sun revolves around Earth) is a bankrupt argument. Governments have almost never been data-driven. Reducing the rubric of how governments/societies should act to "do whatever's cheaper" is... so wrong I don't even really know where to start. How do you justify investments? How do you justify things like entering WWII or The Manhattan Project? How do you know what's cheaper or will result in the most gains ahead of time? This can't be a real argument. Are we about to go the entire history of how governments work? I refuse. Do more reading.
> I think "people being products of systems" is a rather naive take; if people were products of stable systems (in the physics sense), punishment actually would work.
Punishment doesn't work. Deterrence is a myth, and recidivism rates are off the charts. Again you're naive to the criminal justice system. If we don't think better systems/environments lead to better outcomes and worse systems/environments lead to worse outcomes, why are we trying to improve the US educational system at all? A kid's educational attainment is preordained right? Even if you think this is a straw man, it does us no good to consider "some kids just suck" when building an educational system, again because of the School-to-Prison Pipeline where bad outcomes are so lopsided.
> Similarly, we should be able to do a cost-benefit analysis on teacher salaries. In reality, the most efficient use of money is to pour it into the top schools (and have entrance exams), but the metrics don't look as good.
Your argument relies on the prospect of these super students increasing GDP so much we offset the cost of shunting tons of students into the School-to-Prison pipeline. Not only is there no evidence for this, it's a deeply immoral system. I refuse to even continue considering it.
To be pointed about it, I have people in my family who are special needs. I myself was disruptive in school because of life circumstance. When you advocate literally for imprisoning me and members of my family, at some point I have to recognize we're fighting. I've reached that point now.