Agency Detection and The Confident Pundit

A face designed by some higher power, or just randomness?

I recently had a conversation with a friend and, as usual, we talked about Joe Rogan and conspiracy theories and it made for great blog content.

“Was Fauci compromised by big pharma?” was one of the questions we got hung up on.

One thing we share is that we do not have a lot of information on this topic. As such, we rely on heuristics. And so our disagreement was less about facts and more about which heuristics to rely upon.

I noticed that he was using two common System One, intuitive heuristics: the evil agent heuristic, and the confident pundit heuristic. Because these are intuitive, I think we are all prone to them. I just think there are better, more rational, more System Two heuristics that we should be using.

The confident pundit heuristic

We are naturally drawn toward confidence and, with low levels of information, will put our trust in a confident pundit who claims to have the answers.

Although intuitive, it actually makes sense if you don’t think about it for too long. A confident pundit is putting their reputation on the line, putting skin in the game, so why wouldn’t I believe him? Unfortunately, our current system does not penalize people for being wrong. It’s too easy to move goalposts and double down, especially when the confident pundit rarely says anything specific enough to be proven wrong. Most people have a self-preservation bias and will do mental gymnastics to avoid admitting they were wrong when they stake so much of their reputation to their hot take.

So Bret Weinstein is highly confident when he speaks about his vaccine conspiracy theories. And I don’t have much information, so I can’t challenge his assertions. But his confidence doesn’t move me because he’s really not saying anything falsifiable.

If he said, “Because of my assertions, I predict that the number of people dying from heart related conditions will increase by 25% starting in 2024, and the increase will be entirely among the vaccinated,” I would take him very seriously! But he mostly says things that sound like “we don’t know what these vaccines are going to do and they haven’t been properly assessed.” So his confidence does not move me as it isn’t falsifiable—specific, measurable, and timely. There is nothing he says that we can point to and say “see you were right/wrong!”

So I tend to distrust confident pundits whose opinions are outside the consensus. 

The evil agent heuristic

Agency detection refers to our in-born tendency to look for a man behind the curtains. It’s the reason we see faces in the clouds, but not clouds in faces.

We know the reason some clouds look like elephants isn’t due to God making fluffy balloon animals for us. It’s just randomness; there are many clouds, their shapes are random, and so ever so often one looks like a familiar mammal. So we should extrapolate from there and assume that if we have low levels of information, strange events are probably more likely randomness, like the random shape of clouds and our agency detection thinking they look like Elvis, than evil manipulators making these shapes.

UC Santa Barbara conducted a study where they tested people's abilities to find hidden patterns. Half of the subjects, before the test, either read a Kafka story or watched the David Lynch film Blue Velvet. If you've never seen Blue Velvet, the immediate reaction from most normal people is "What the fuck did I just watch?" In other words, the plots don't make any sense. Those people's scores where better, statistically significant, than the control group.* 

One of the study's psychologists said: "The idea is that when you're exposed to a meaning threat –– something that fundamentally does not make sense –– your brain is going to respond by looking for some other kind of structure within your environment."

JFK

I don’t have a lot of information about the JFK assassination, since I don’t find the topic interesting. But there is one theory that I like.

As far as I understand, the part that doesn’t make sense is that one bullet appears to come from the opposite direction than the ones fired by Oswald, which exposes people to a "meaning threat". This is what leads to the “magic bullet” or second shooter theories. The fact is we do not have much information, which leads to conspiracy theorists who begin with agency detection and the evil manipulator heuristic, and end with the belief that the CIA was behind everything.

The theory I like goes like this. Kennedy’s secret service agent was beside him when he got shot. He reacted by quickly pivoting and returning fire in the direction of Oswald, and Kennedy’s head got in the way, which explains the “magic bullet.” In essence, Kennedy’s own secret service agent was unintentionally the second shooter. The reason for the secrecy is that, well, it’s embarrassing.

Now, this theory is probably just as plausible as anything you’d find in any conspiracy book. But it’s much less popular for the simple fact that it’s fucking boring. There is no evil agent to blame. So given low levels of information, and a high ratio of randomness to evil, I favor the boring explanation.

But sometimes there is evil and the conspiracy theorists are right! Does that mean my boring heuristic is garbage and assuming evil is always the right choice? That depends on how you feel about the Blackjack scene from Swingers.

Always double down on 11

Is Trent (Vaughn) right, or is Mikey (Favreau)? We’re probably asking the wrong question because Trent is actually talking about broad framing. Trent is asserting that if you follow the “always double down on 11” heuristic, in the long run, you’ll make money. Mikey is caught up on his granular experience and how he lost by following that heuristic. (Technically, Mikey is right. If you’re broad framing, you need to have enough capital to avoid going bust after a few initial losses. Plus, there is no successful heuristic in gambling, the house plans for it and always tips the odds in their favor.)

But the point is that Trent and Mikey, like my Rogan-loving friend and I, aren’t arguing about facts but about heuristics. And heuristics shouldn’t be judged on a case-by-case basis, but how often they’re right in the long run.

So my yet-to-be-proven belief, is that if you follow the boring fallacy you will be right more often than the evil fallacy or the confident pundit. I also have an anti-confident bias so I avoid individuals who are overconfident and arrogant, except for a few arrogant people I think are usually right (Nassim Taleb, Eliezer Yudkowsky). Instead I rely on a consensus of experts or pundits who give specific, measurable hot takes; bonus points if they use a confidence interval.

*If the UCSB study was a better example, then it would have shown that the Kafka/Lynch subjects found more false positives (i.e. identified patterns that weren't actually there) than the control group. But this is my favorite study ever and I've been working really hard to find a way to work it into a blog post, so deal with it.