> The idea of societal benefit is very nebulous. [...] How do you measure something like societal benefit?

That is indeed a very hard problem and one of the mayor challenges of any society. But that doesn't mean that any endeavor that includes this notion is to be avoided. You gave the example of eradicating disease.

> I can guarantee we have different opinions about the value of things. That's why we should be free to make our own choices. We don't need everyone in the world to help define what "societal benefit" is for them, they can do it every time they pay for something out of their own pocket.

But now you are not arguing about ice cream anymore. You seem to be just generally skeptical of a strong centralized government. And you can of course argue that and there are no simple answers here. I was just pointing out that your ice cream argument doesn't hold water.

Unless it is "Any kind of government interference is bad. This is why we shouldn't ban ice cream. Therefore we also shouldn't ban bitcoin." If that is your stance, then ice cream and bitcoin are indeed equivalent for your. But that is probably not the worldview of the people asking for bitcoin to be banned. Then your post does not argue against their view, but against your version of their view.

Market forces are not fully understood in complex systems but it does not mean they do not work.

Forcing employers to give more money to employees (rising wages through regulation) may cause some unexpected/unaccounted things (more consumption etc.).

Taxing something (like CO2) will definitely do not help that thing. Solving just Bitcoin problem is both hard (how do you prevent making certain computation while connected to the network, or how do you prevent buying certain digital goods) and is only part of the problem (it does not mean that fixing part of the problem is not worthwhile, but they are easier ways to solve bigger part of the CO2 problem).

They subsidize R&D because there's a public benefit to things like pure math that's difficult to internalize which leads to under supply from the market.

They subsidize agriculture because of lobbying and self-defense considerations.

In both cases, the government identified an under supply due to a market failure, and they judged that the externalized benefit was sufficiently large and so they stepped in to fix it.

In the case of crypto, the market failure is simply that carbon is too cheap. So why not address the actual market failure with a tax. If crypto miners want to keep mining after compensating society for their pollution, more power to them. It's hubris and risky to outright ban something because you just know there's no value there and all the people who ascribe value are wrong. Doing that isn't even analogous to government actions in the two examples you raised. A carbon tax is more analogous (it's sort of like an inverse-subsidiy as per your Ag example).

I like the carbon tax avenue for other reasons too. It's a broad, one shot thing that will roughly aligns incentives across all industries at once. Not perfect but a very good correction that broadly applies.

  "by all accounts, it slightly increases employment"

Not by all accounts. The literature is not straight forward in this way. If it is a null effect, I don't see that as an invalidation of econ 101, either. It could just be very inelastic for small changes in the minimum wage with elasticity picking up for larger changes (which nobody doubts would happen).