Petrov Day 2024: What is a Petrovian virtue?

On the 26th of September, each year, we celebrate Petrov day.

Stanislav Yevgrafovich Petrov was a lieutenant colonel in the Soviet Air Defence Force, working at Oko, the Soviet missile defence program. On this day in 1983, during a night shift, Petrov was on duty when his monitor alerted him of a missile inbound from the US. He froze.

"The siren howled, but I just sat there for a few seconds, staring at the big, back-lit, red screen with the word 'launch' on it," he says. (source: BBC News interview)

More missile warnings came in. It was Petrov’s job to trust the computer system and relay its message up the command chain. If he had, the USSR would likely have started a retaliatory strike on the US. 

But Petrov knew the computer system, and doubted that a true missile warning would have passed through the levels of verification so fast. He decided not to call his superiors. 23 minutes after the first missile alert, he was relieved. If the warning hadn’t been a false alarm, nuclear missiles would already have hit the USSR.

You can read more about Petrov's story here. If you are a LessWrong user, you can opt in to their Petrov Day event by pushing a large red button...

We celebrate Petrov day, on its surface, because of all the people who are alive today due to his actions. But we also discuss the man himself, and his story. What are we trying to learn from him?

There isn’t a very obvious career-related lesson in his story. We don’t celebrate him for joining the Soviet Air Defence Force. We celebrate what he did on one particular day. But, if we try a little bit, we can pull out some general lessons, or virtues that Petrov exhibited.

While I was drafting this post, I saw that LessWrong beat me to it and did a poll to find the most important Petrovian virtues last year — their winner was “Avoiding actions that noticeably increase the chance that civilization is destroyed”. This is a pretty objectively valuable virtue, but I’m especially interested in exploring virtues which are more frequently useful to us. In other words, virtues we can make part of our characters. Below are some attempts:

Scope-sensitivity: Decisions which look superficially similar can have wildly different stakes, and it takes a persistent effort to be sensitive to them. In my job, for example, most of my communications are with only a few people, and its fine/ expected to be very casual. However, some go out to tens of thousands, and can plausibly be framed as representing Effective Altruism in general. I set up systems like copy-editors and sign-offs so that I stay sensitive to that difference.

Or, for example, maybe you’re in a job where you often have to make quick and accurate decisions to pass on to your soviet military superiors. Many decisions will have fairly high stakes, but perhaps, a year off from your retirement, you might be given a decision to make that, if wrong, could lead to the deaths of tens or hundreds of millions. Best to be scope sensitive then.

Steadfastness: Avoiding social pressure— being ready to be the one blocker in the room. 

Perhaps this is more exemplified by the story of Vasily Arkipov, another Soviet military man, this time in the Navy, who was on a nuclear submarine during the cuban missile crisis. When depth charges went off around the submarine he was on, the first and third in command wanted to launch a nuclear torpedo, thinking that a war must have begun. But Arkipov, the second-in-command, disagreed. Without unanimous approval, they couldn’t launch the weapon. His colleagues were angry, but he held firm.

It’s extremely difficult to be the one individual unilaterally blocking an extremely high stakes action. It’s also really hard to be steadfast in a virtuous way, i.e. without also picking up the vice of stubbornness. The key to really holding this virtue is to combine it with scope-sensitivity — if the stakes aren’t too high, be less stubborn. But maybe there is another virtue we need in order to make these work in our lives. After all, Arkipov and Petrov wouldn't have been as virtuous if they were wrong. 

Epistemic Humility <- X -> Epistemic Pride. X marks the virtue. What should we call it? An epistemically humble Petrov would have deferred to his orders, an epistemically humble Arkipov would have deferred to the first-in-command. If either had been too epistemically proud, they may have fared worse in a world where they should have fired the weapon^[1].  

Perhaps we should call X "Epistemic Responsibility": the ability to recognise when the grounds of your own beliefs are likely better than those of your interlocutor, and to argue for them.  

What next?

• Add more virtues in the comments below, or disagree with mine. 
• Write your own Petrov themed post! You can tag them "Petrov Day". 
• Don't destroy the world.

1. ^{^}

   If there are such worlds.