A Rawlsian case for smallness

Plus, the rise of the most antisemitic generation and the case against debate

Before we get to today’s questions, I wanted to flag Sam Altman’s recent blog post, “The Gentle Singularity,” in which he reiterates the view (common among AI professionals) that super-intelligence is coming in the relatively near future, and then goes on to be excessively breezy about the technical and policy challenges associated with making this change be good rather than bad.

To his credit, though, he at one point moves out of his comfort zone of blithe techno-optimism to offer the point that “social media feeds are an example of misaligned AI; the algorithms that power those are incredible at getting you to keep scrolling and clearly understand your short-term preferences, but they do so by exploiting something in your brain that overrides your long-term preference.”

My view is that this is an absolutely critical insight. Even though it’s true that technological progress has, on average, been enormously beneficial, it just absolutely does not follow that each individual example of technological progress is beneficial. Way too many people in tech make the argument that every past advance has been greeted by skeptics and naysayers who believe it will turn out to be harmful, therefore we should ignore the skeptics and naysayers. But there are plenty of examples of advances that are, in fact, harmful. OxyContin and fentanyl are both scientific breakthroughs with legitimate use cases and real benefits that are, on balance, harmful. Like most people, I have had some pleasant experiences with algorithmically programmed short-form video, but the rise of this kind of quasi-addictive content is, on balance, bad. Modern nuclear weapons have (so far) not been wildly harmful, but they clearly would be if any random person could get his hands on one.

The question of whether the race toward superintelligence is, in fact, being pursued in accordance with the best interests of humanity is one that warrants much more serious answers from people like Altman, certainly more than vague assurances that technological change usually works out for the best. I’m not a pessimist or a hater or an anti-capitalist or a Luddite, but I am acutely aware of the point he’s making about social media. Is the ability to make high-quality videos on the fly so that social media feeds can become even more addictive good? Is this even close to the biggest problem with controlling agents that are smarter than us?

Now for today’s questions.

---

Liam Kofi Bright: Hi, after reading this post it made me wonder — do you think what is going on is the people you disagree with here are implicitly committed to something like Rawls' ideal of property owning democracy? That is “a democratic society in which land and capital are privately owned and widely (though not equally) held. Concentrations of wealth have been dissolved or mitigated so that [s]ociety is not so divided that one fairly small sector controls the preponderance of productive resources” — and, I dunno, that seems to be about the sort of thing the “small is beautiful” crowd want? I ask because sometimes it can seem like they have nothing going for their version of things beyond a sort of romantic pastoralist ideal, but maybe there is more to be said in favour of property owning democracy as a social form. Rawls liked it after all!

I’ve been struggling for a while to make sense of the anti-bigness movement, and this is a great question, because I think it does provide us with a possible way to understand it.

Just to back up, John Rawls famously argued for “the difference principle” under which “social economic inequalities are to be arranged so that they are … to the greatest benefit of the least advantaged, consistent with the just savings principle.” A lot of what happens when you take an introductory political philosophy class is that people read this as a case for a strong welfare state, and then the professor points to the places in the text where Rawls says something more far-reaching than that.

The clearest example of why he thinks we need something more far-reaching is offered in Justice as Fairness: A Restatement, where he offers the view that “inequalities in the ownership and control of wealth, income, and property can reduce the fair value of basic liberties.”

What goes beyond a welfare state? Rawls offers two ideas as I recall:

1. Liberal socialism — this is basically Matt Bruenig’s proposal for a large sovereign wealth fund that would ensure that capital is broadly owned, even though there would still be large private businesses.

2. Property owning democracy — this is the idea that we prevent any super-gigantic pools of capital from forming in the first place.

My biggest frustration with Rawls, as a writer, is that having mounted an extensive philosophical argument for some abstract principles, he neither tries to flesh them out with detailed empirical work nor steps back to note that these principles obviously raise a lot of empirical questions. Instead, he tends to dip his toes into various strong empirical claims about contested issues without much sustained focus. The basic claim that socioeconomic inequalities can generate unequal political influence seems true enough, but if you’re going to let your whole economic system pivot around this observation, then I think you need a very clear account of exactly how this influence works and how your proposed solution solves the problem.

Medical doctors, for example, have a lot of political clout. In part that’s because they are more affluent than average. But in large part, I think it stems from the fact that people trust medical doctors to speak on issues of personal and public health. If doctors start speaking out both publicly — and in their private communications to patients — about the idea that some policy change will be harmful, a lot of people will listen to that. And if politicians and journalists try to explain that this is actually doctors behaving in a self-interested way, the fact of the matter is that doctors are held in higher esteem than politicians and journalists.

Police unions, similarly, have enormous practical political influence that stems only very partially from the formal authority of the union qua union. If a city’s cops get into a fight with the city’s mayor about why crime is rising, the cops are just objectively likely to win the argument.

I don’t think those points per se debunk Rawls’ concern about big business, but they illustrate that this is complicated terrain.

Even on a narrow issue like campaign finance, where it’s clearly true that richer people get to be more influential, things are complicated. The people who vote in Democratic Party primaries are to the left of people who vote for Democratic Party candidates in general elections. The major donors to the Democratic Party are even further left than the primary voters. But the small donors are furthest-left of all. If you eliminated the influence of the mega-donors, would that drive outcomes toward the ideal point of the primary voters or toward the ideal point of the small donors? And why don’t Democrats with moderate views just make small donations? One of the things we do at Slow Boring is to encourage them to do just that. Maybe “moderates don’t donate” is just a contingent fact of history that will change, not an iron law of politics that we need to structure the economy around. But I’m not certain.

All of which is to say that Property-Owning Democracy, as an economic program, is something that I think would cost a country a lot in terms of prosperity. Rawls makes a reasonable (though I would not say ironclad) case that the fair value of political liberty should be seen as more important than prosperity. And he then raises some valid points about the interplay of economics and the fair value of political liberty. But he raises those points in a way that I think is much too casual to support such a strong conclusion.

Freddie deBoer: Debate me on Substack live.

This is maybe an idiosyncratic view of mine, but I think that “debating” people — particularly in live or quasi-live forms — is a bad epistemic practice.