The independent truth service

"The abortion pill is the bitterest pill."

While I remain of course completely blackpilled, I can’t think of a revolutionary event in this weird winter more hopeful than Facebook ditching its sinister state-sponsored “fact checkers” in favor of Elon’s elegant “Community Notes” hack.

However, like all such events in this weird winter, I still have the unfortunate intuition that what I see is only the mutant, stunted, probably non-viable abortion egg of something that might happen, but actually won’t. The abortion pill is the bitterest pill.

By refusing to accept the epistemic supervision of the truth-determining institutions in our regime, Twitter and Facebook have taken a revolutionary step. Yet since they are not revolutionary organizations, they are enormously relieved by any suggestion that this first step could also be their last. But this is how to end up as an abortion.

The unconscious assumption of the conservative is always that his enemy, despite controlling every serious institution on the planet, is somehow deeply weak, probably about to collapse, and can be taken out by a single bold lightning-like blow, such as removing the tampons from the men’s bathroom at Facebook. I see this mindset in every WinRed spam text I get as a punishment for giving a few bucks to a politician.

So it’s still hard for what President Macron calls our “new international reactionary movement” to gain steam. The revolutionary mindset, universal as it is in our society, is both rather fossilized, and totally foreign to the conservative mind—which is always seeking some magical way in which it can do as little as necessary.

On the contrary, comrade! As another Frenchman, St. Just, said—he who makes half a revolution digs his own grave. When we enter the revolutionary mindset, we must be looking relentlessly for ways to do as much as possible—not just because this produces more results, but because, in a revolutionary situation, it is actually easier to do more.

No, really: in a revolution, the farther you go, the lighter your steps become. The first step is the hardest step. If you want to stop there, the next step will seem impossible. Once you know where you want to go, though, you will see that it is slightly easier…

What would the next revolutionary step for Facebook and Twitter be? They could put significant energy into creating an independent truth service.

Science is a truth service. Journalism is a truth service. Our society is built on these services in the same sense that it is built on airplanes, bridges, and fiber-optic cables. If our prestigious universities and newspapers cannot be trusted as an information source that can correct 55-year-old soccer moms in Reno spreading medical myths, something is very wrong.

When the leaders of Facebook and Twitter assumed that their “fact checkers” were as reliable as the airplanes they fly on or the Internet they zoom on, it made sense to let them censor the soccer moms. Once they saw how it actually is, they would rather travel by wagon and send handwritten letters in bags strapped to ponies.

Actually we don’t want to travel by wagon. We want to fly. Actually nobody wants lies on their social network. They just disagree on what a lie is. When you discover that you actually don’t have an API which tells you whether a post is a lie, you might to decide to make the best of it, and come up with some libertarian ideology that tells you that lies are okay on your social network, actually, because, like, freedom.

But what if you did have that API? And it actually worked? Would you not use it?

My position is: “fact checking” is like medicine. A city needs doctors. A mass-market social network needs fact checkers. But a city needs nothing less than quack doctors. We can agree that strong men with whips must drive the quacks from the hospitals, but the cure for quack medicine is not no medicine. It is real medicine.

When Science (the prestigious meta-institution) diverges from science (the rigorous pursuit of truth), the cure is not no science. It is real science. And so on.

Or you could split the difference, in the Silicon Valley system-software style of finding just the right 80-20 tradeoff point, with a clever hack like Community Notes. Without spending any money at all, hardly.

The problem with Community Notes is that Facebook and Twitter are still curing the disease that the quacks diagnosed. They are just not using the weird ointment that the quacks prescribed. That ointment worked, sort of. And this ointment works, sort of.

But it works at a terribly low level of power. The power of a truth service is the level of analytic energy it can put into a question—and it doesn’t take much analytic energy to keep soccer moms from sharing old wives’ tales and urban legends, because it doesn’t take much analytic energy to tell whether old wives’ tales and urban legends are true.

But suppose soccer moms are sharing the idea that… Covid was a lab leak. At once, the difficulty of knowing the truth, for Facebook, rises to a very high level of cost. As social administrators, we are tempted by any ideology that lets us just not solve this problem. Everyone wants to do less. No, comrades, we must do more!

Suppose Facebook considered it part of Facebook’s business, a critical business task, for Facebook itself to have a position on disputed objective questions like this. Facebook, to know whether to treat this idea as a myth or the truth, has to actually know for itself. It cannot delegate this decision—except to some institution that it completely trusts. This is what I mean by an independent truth service.

The astounding situation we find ourselves in is that we have no institution capable of deciding whether Covid was a lab leak. Has the press done it? Have scientific journals done it? Have courts done it? Has Congress done it? In theory…

We could give it to Community Notes. Do people who usually disagree with each other agree that Covid was a lab leak? Heck, we could use the same algorithm in our courts. Do jurors who usually disagree agree that the defendant is guilty? Or even our legislatures. Do Senators who usually disagree agree on this bill? We could replace journalism too. “All The News That People Who Disagree Agree Is Fit To Print.”

It’s a good heuristic. But all heuristics which matter get gamed. Google’s backlink algorithm was a good heuristic. There’s not much left of it now. I don’t think CN has been successfully gamed yet, but it also doesn’t matter much yet. Give it more power, economic power, and the same thing will probably happen. Fundamentally: it’s a toy, and Facebook is a game, and all you need for a game is a toy.

But what if we took it seriously? The game needs a toy—that’s an economic incentive. Maybe we could spend a little more on this toy than we strictly need to. Maybe the toy could work so well that it became a weapon. Put enough air pressure into a squirt gun, turn it up to 10, and it hurts. Your sister yells. She’s still laughing. Turn it up to 800, and you have a water cannon that can rip her face off in a millisecond. Metaphorically.

Our society is based on “reliable sources.” But those sources have become unreliable—not even randomly unreliable, but unreliable in a predictable way. It turns out that even truth services, including but not limited to the New York Times, will become corrupted by power, and quite unfixable, in the long term.

Here is a small attempt at an independent truth service: Rootclaim. I am not sure I am very fond of Rootclaim—I find the Bayesian math gimmicky and unpersuasive. I feel the goal of a truth service is not only to find the truth, but at the same time persuade. I think the problem could take much more inspiration from historical institutions, especially courts. On the other hand, historical institutions are very far from optimal.

Here’s the thing: because our society is based on “reliable sources,” if you have a real reliable source, people will listen to it. Not at once, not all at once, but—“a city on a hill shall not be hid.”

For example, Wikipedia is based on “reliable sources.” Predictably, since Wikipedia matters, Wikipedia has taken the predictable unreliability of its “reliable sources” and doubled down on it. How could this be corrected?

Well, if we imagine a truth service as far beyond Rootclaim as Rootclaim is beyond Community Notes—using it to settle Wikipedia edit wars would be obvious. In fact, using it to fork and correct Wikipedia would be obvious. The original Wikipedia would shrink into a marginal boomer lefty site—the Utne Reader of online encyclopedias.

Suppose you’re trying to reform the government. The government depends on many, many facts which are established by its truth services. When its truth services mislead it, the government does not execute in reality. Instead, it does wacky stuff. Its form is still that it executes according to the reality provided by its services—the rule of the modern government, like the rule of Wikipedia, is “no original research.” The state, with a few small exceptions, is not allowed to think for itself. It has to listen.

All modern institutions are programmed to scan the airwaves for a truth signal. Right now, the strongest truth signal—the biggest, the loudest, the most listened to, the proudest and most ambitious—is the New York Times. Or more broadly, the prestige media. Or for deep analytic questions, the prestigious university.

The belief that “we don’t need a truth signal, the marketplace of ideas will solve it,” or “the marketplace of ideas, plus a simple heuristic, will solve it,” is conservative cope. Again, the conservative is always trying to do as little as he can get away with, then wondering why he keeps losing.

But actually, if you build a bigger, louder, higher-fidelity truth signal, you have a very good chance of becoming the standard truth signal—especially when the old system is in such obviously parlous shape.

It’s more tricky to think of this subsystem as a profit center. And everything that capitalism does has to be profitable. Still: the direct effect of creating the strongest, widest, and most authoritative truth signal is the production of power. And if there is one lesson in history, it’s that on a sufficient time scale, power always earns a return.