LW Petrov Day 2022 (Monday, 9/26)
Next Monday is Petrov Day (September 26), an annually observed Rationalist/EA holiday inspired by the actions of Stanislav Petrov:
As a Lieutenant Colonel of the Soviet Army, Petrov manned the system built to detect whether the US government had fired nuclear weapons on Russia. On September 26th, 1983, the system reported five incoming missiles. Petrov’s job was to report this as an attack to his superiors, who would launch a retaliative nuclear response. But instead, contrary to the evidence the systems were giving him, he called it in as a false alarm.
It was subsequently determined that the false alarms were caused by a rare alignment of sunlight on high-altitude clouds and the satellites’ Molniya orbits, an error later corrected by cross-referencing a geostationary satellite.
In explaining the factors leading to his decision, Petrov cited his belief and training that any U.S. first strike would be massive, so five missiles seemed an illogical start.
Petrov underwent intense questioning by his superiors about his actions. Initially, he was praised for his decision. Petrov himself stated he was initially praised by Votintsev and was promised a reward, but recalled that he was also reprimanded for improper filing of paperwork with the pretext that he had not described the incident in the military diary.
He received no reward. According to Petrov, this was because the incident and other bugs found in the missile detection system embarrassed his superiors and the influential scientists who were responsible for it, so that if he had been officially rewarded, they would have had to be punished. He was reassigned to a less sensitive post, took early retirement (although he emphasized that he was not “forced out” of the army, as is sometimes claimed by Western sources), and suffered a nervous breakdown.
For more information see 1983 Soviet nuclear false alarm incident
On LessWrong, we find our own way to celebrate, generally involving a large red button that brings down the frontpage for the duration of Petrov Day.
What does Petrov Day celebrate?
There isn’t a canonical precise answer accepted by everyone. There’s a cluster of virtues and actions that people find worthy of remembering with different degrees of emphasis. These include:
Not doing things that would cause immense destruction (or the end of the world)
Avoiding the dangers of unrestrained escalation
Not taking unilateralist action
Resisting social pressures in order to do the right thing
Making the right decision even in the face of uncertainty
You might even say part of the Petrov Day tradition is debating which virtues Stanislav Petrov displayed and which ones we ought to celebrate. Personally, I like the underlying simple theme of “someone was in a high-stakes situation where they could have chosen a destructive path, and they didn’t” and “things were close, but we survived”. As far as the LessWrong celebration goes, each year I like the idea of exploring a different sub-element of surviving high-stakes scenarios and the virtues required to do so. This brings us to this year’s plans...
The 2022 Plan
This is what I’m thinking:
Every user with an existing LessWrong account (created before 2022-09-21) and non-negative karma is able to participate. We may manually exclude some known historical troublemakers.
Your actions will be anonymous, including to the LessWrong team. This is a major change from last year. If you act counter to what other people think you should do, you’ll only have to live with your own self-judgment and the mental simulation of others :P
There is a virtue in preventing yourself from ending up in tempting situations where you’d have to apply a lot of willpower (e.g. facing an alluring red button that calls out to you like a siren). Also a virtue of distancing yourself from things you think are wrong. Accordingly, you can opt out of LessWrong’s Petrov Day commemoration by checking the “opt out of Petrov Day” checkbox in your account settings. The checkbox will be hidden once Petrov Day starts to prevent people undoing their self-commitment in a moment of weakness.
We will place a Big Red Button on the frontpage, as is customary. Pressing the button and entering a valid launch code causes the frontpage of LessWrong to be taken down and replaced with a “Game Over” screen (other pages will remain up).
At the start of the 24-hour celebration, only those with 2,300 karma or more will be able to successfully launch. Every hour after that, the karma required to destroy the frontpage drops by 100 points.
I will post the launch code for the missiles in the comments below once The Button is displayed, starting at 8pm, Sunday, September 25 PST / 3am, September 26 UTC.
Should you press the button?
Personally, on net, I think you shouldn’t. I think it destroys some real value and is symbolic of destroying real and even greater value. Plus, I think there’s symbolic value in not pressing buttons that launch nukes. But that’s what I think...I may write more in the comments about what I think so as not to overly privilege my own view here. I do think there are not-crazy arguments in favor of pressing it.
Yet many argue that Petrov’s virtue lay in deciding for himself the correct thing to do, not going with what authority figures or social consensus dictated. To that end, this year’s LessWrong Petrov day has been designed to more readily let you come to your own conclusion and act on that without fear of a public shaming.
You’re free not to participate (see opt-out above) and you’re free to apply your own interpretation, different from mine. If you reason that it is in fact correct to press the button, earnestly believe that, and have reasoned well – then I commend you for pressing it.
A useful piece of info I think might help inform people’s judgments is the practical cost of bringing down the LessWrong frontpage for a number of hours.
On average, a bit over 3,000 unique people visit the LessWrong frontpage each day. Equivalently, 100-150/hour. If the frontpage goes down halfway through Petrov Day, that is 1500 people who wanted to check latest posts, etc., who were not able to do so.
We’ve got a few more days before Petrov Day and while our plates are pretty full, it’s not too late to make adjustments. I’d also be interested to hear alternative ideas, e.g. Ben Pace’s An Idea for a More Communal Petrov Day in 2022
Let me know your thoughts! I look forward to surviving Petrov Day with y’all!
This is the frame of Petrov Day presented on LessWrong in 2019, though I found it then and still now an odd one.
The first couple of years of Petrov Day on LessWrong involved selecting a group for ~100-200 trusted users and giving them launch codes that would let them bring down the site, kind of exploring “how large is the circle of people we can trust?”. Last year we played out the game theory of Mutually Assured Destruction by pairing up with the EA Forum.
To get more technical, there are two parts of launch: (1) pressing the red button, (2) entering a launch code and pressing “launch”. The code will not record any user-identifying information with the second event. We do track which users have pressed the red button at all in order to persist the animation state (for historical reasons it was written that way). My guess is that with effort it might be possible for us to forensically determine the identity of someone who made a successful launch event, but I commit that we will not do so barring extreme circumstances (that I haven’t yet imagined, but it’s hard to think of everything in advance).
If you wish to opt-out once Petrov day has begun, message us.
That isn’t to say I can’t promise you there are zero consequences for your choices. That’s not within my power. For example I can’t promise no one else will ever ask you “did you ever press the LessWrong Petrov day button?”, though you could glomarize.