Directionally Libertarian

Making the case for freer people, freer markets, more pluralism, less centralized control.

[Directionally Libertarian @UnpopularLib]

@mehdirhasan I'm having trouble keeping up. Is Trump a credible moral authority now?

[Directionally Libertarian @UnpopularLib]

@mattyglesias Major severity, minor probability, perhaps?

[Directionally Libertarian @UnpopularLib]

@mehdirhasan Since when is “Donald Trump agrees” evidence that something is true?

[Directionally Libertarian @UnpopularLib]

@DavidSKrueger Is the alternative to “psyop” really “they must be right”?

[Directionally Libertarian @UnpopularLib]

@LPMisesCaucus "heavy restrictions" undersells it

[Directionally Libertarian @UnpopularLib]

This would be a great argument if weed had been broadly legalized nationwide with low barriers to entering the legal market, and organized crime still took over the supply chain.

That isn't what happened. We legalized a patchwork of heavily regulated state markets surrounded by prohibition and illegal interstate commerce, leaving enormous demand for a black market to serve.

You can't preserve most of the conditions that create a black market and then cite the existence of that black market as proof legalization failed.

[Directionally Libertarian @UnpopularLib]

@beinlibertarian I'm some ways, way faster.

[Directionally Libertarian @UnpopularLib]

@RepCasar How does your regulation prevent an AI from deceiving the people testing it?

[Directionally Libertarian @UnpopularLib]

@Noahpinion The Creator (2023)

[Directionally Libertarian @UnpopularLib]

@MichaelTrazzi @DanielLurie @StopTheAIRace What exactly is the mayor of San Francisco supposed to do about the global development of artificial intelligence?

[Directionally Libertarian @UnpopularLib]

@HumanWatchVsAI AI has already helped cure cancer.

https://t.co/7YJ9TOmneZ

[Directionally Libertarian @UnpopularLib]

I assume most people are well-meaning.

Very few people see themselves as the bad guy. Even when they know they're doing something questionable, there's usually a justification: it's necessary, it's for the greater good, the other side is worse.

I think that's true of people and institutions on the left and the right.

You don't need evil people to get terrible outcomes. Good intentions, strong incentives, and enough ways to rationalize what you're doing can get you there just fine.

[Directionally Libertarian @UnpopularLib]

@mattyglesias The Creator (2023) had a pretty unique take on AI. Instead of “what happens when AI becomes powerful enough to threaten humanity?” it asks: what happens when humanity becomes terrified of an intelligence powerful enough to rival it?

[Directionally Libertarian @UnpopularLib]

@jessesingal Which regulation prevents an AI from becoming smart enough to deceive the people testing it?

[Directionally Libertarian @UnpopularLib]

@Alt_Azn We legalized cannabis without creating anything resembling a normal national legal market, then act surprised that the black market survived.

[Directionally Libertarian @UnpopularLib]

This would be a great argument if weed had been broadly legalized nationwide with low barriers to entering the legal market, and organized crime still took over the supply chain.

That isn't what happened. We legalized a patchwork of heavily regulated state markets surrounded by prohibition and illegal interstate commerce, leaving enormous demand for a black market to serve.

You can't preserve most of the conditions that create a black market and then cite the existence of that black market as proof legalization failed.

[Directionally Libertarian @UnpopularLib]

@ShitlibSupreme Healthy spread

[Directionally Libertarian @UnpopularLib]

@KatieScarlet13 False

[Directionally Libertarian @UnpopularLib]

@RoKhanna Translation: move even more of the economy away from consumer choice and competitive markets, and toward political allocation and control.

[Directionally Libertarian @UnpopularLib]

If you're worried about a handful of corporations controlling enormous amounts of economic power, “pause the industry and build a regulatory regime around the handful of companies currently at the frontier” should at least make you nervous.

They're the companies best positioned to absorb the costs, influence the rules, and make today's concentration permanent.

[Directionally Libertarian @UnpopularLib]

I basically agree. My uncertainty is how much regulation actually reduces the risk rather than institutionalizing today’s concentration.

These companies already have enormous incentives to take safety seriously, and they aren’t operating outside the law now. Existing liability, consumer protection, antitrust, privacy, criminal and sector-specific laws already apply in various ways. Regulation could mandate precautions they would have taken anyway, take credit for the resulting safety, while mainly adding compliance costs and political influence.

Frontier AI is enormously expensive today. It may not always be. I don’t want us to respond to a temporary technological barrier by creating a permanent regulatory one.