Content warning: nuclear doom

Previously on last week's episode: A Few Terrifying Facts About The Russo-Ukrainian War

You have probably thought about prepping at some point.

You have probably also not prepped as much as you'd like.

Neither have I.

And neither of us likely will be prepared "enough", in this world, if the worst happens.

You've probably seen those prepper checklists and balked at completing every item on them.

Maybe you've told yourself it's not worth living after a nuclear war.

Maybe you've convinced yourself that preparing for the set of things that includes normal disasters like an earthquake or hurricane will be all that is reasonable to do.

I've long yearned for a proper collective response to Global Catastrophic Risks In Your Local Neighborhood.

I'm coming to accept that there's simply no reaching adequacy. I'm still likely to fuck up on some significant axis and still pretty likely to die after a nuclear war, though less than before each preparation I took.


If you believe in dying with dignity, then you can die with more or less dignity after a nuclear war.

You can die from radiation after a couple weeks, and because you had emergency water, someone else had access to water and didn't die of thirst, and you were able to help a couple friends get to safety before you succumbed.

You can have potassium iodide pills, and not die of  radiation but then still get cancer after a year. You have emergency water and two weeks of food, but you somehow get stuck in your urban area because all your friends with cars took off and you forgot to download cash from an ATM to pay the people who want money to drive you out of the nuke zone.

Part of a nuclear war is likely to be EMPs, which result from atmospheric detonations. This means your electronics probably get fried. And the banks. You lose access to the Internet, to your phone, to most of your money... Probably? But you can protect your electronics with a Faraday Cage bag! That's a whole lot of dignity to have your phone or a radio working, even if it's just a 4-hour charge and all the value it brings is an offline map and the ability to reminisce over your photos before you die. Maybe you find a source of electricity and can extend your battery life. Maybe some internet works since the internet was originally partly designed to survive a nuke I think.

You have a car or a friend with a car, you have emergency food and water and enough gas, you have your bugout bag, your trauma kit, you have camping supplies and a thick wad of cash. You go somewhere out in the boonies with your friends. You run out of your food in a month and have to hunt and forage... except you're a bunch  of nerds who don't know how to do that and slowly starve to death and get sick as you eat things you're not supposed to eat and don't eat enough.

Anything else I say is too much speculation if it isn't already. I'm not going to research this for my handwavey survival stories to be more accurate nor am I your source for emergency preparedness checklists. There is probably some really bad assumptions in my model of how things would go but I don't think it's worse than most, in fact I want to expose it. It's important to stay humble and not think that you're a smart person who needs to know everything about what will happen vs maintaining the ability to act under uncertainty and admit this is just a bad situation to be in where every little bit helps.

I can't decide for you what is the Pareto optimal set of actions between, say, increasing AI dignity vs nuclear dignity. Maybe you should do nothing and hope your company or someone else does it for you. Maybe you should just spend 1 hour stocking up on food and water. Maybe you should have a plan to move if a nuke is detonated. Maybe you're not working on anything important and helping EAs prep could be a really valuable thing you could contribute right now.

The shape of the utility curve may reach quickly diminishing returns after you buy some basics and make a quick evac plan. 

Or it may be a curve where you only gain days, then weeks, then months, then finally if you put a lot of effort in you get years of survival (i.e. increasing utility towards the end; my math is rusty, I'm more of a wordcel at this part of my life). And maybe at an individual level, that effort is Just Not Worth It, and this is a coordination failure that could be helped with an assurance contract or some other equilibrium-shifting thing like Elon Musk ostentatiously hiding in a vault. 

Because survival is a public good. We have collectively failed to not have a vault space for each person. We have collectively failed even more by having Damoclean weapons hanging over us in the first place, except for whatever actions we took that seem anthropically favored so far.

Actually taking this shit seriously looks weird. It can look like wearing a Hazmat suit in January 2020. Even just having one N95 mask then though would have been a good 80/20. Then again, there's a lot of assurance you can get out of complete safety. At throwing enough dakka at it to solve the problem for yourself.

You're not by yourself, though. You're reliant on a web of people around you and still will be after a nuclear war.

This is why I've decided, instead of focusing on myself, to ask many online friends what they're doing, to help them. I figure if I help others this will flow back to me and help me figure out my own situation. Making it social is more fun that way. I think most people have a blindspot around this topic and if we look out for each other in ways not typical to the Western individualist ethos this will do more total good.

This is definitely not an act of procrastination on my part.

I don't mean any of the doomy vibes as some kind of pessimism or emotional abuse. What matters is what you in fact manage to get done. Even if you lose, you can play better or worse based on some trivial inconveniences, so don't give up and don't get overwhelmed. I want you to survive if WW3 happens. And it's good to think about this now because if not this year, then maybe it's a confrontation over Taiwan in the next ten years. Now is the best time to make the distant close.

You're allowed to cheat. You don't need to pass a trivia test about how to find 17 different nuts and berries in some forest. You're allowed to have attachment to your current way of life and be pragmatic. You're allowed to, hypothetically, fly to New Zealand, build a hobbit house, fill it with Soylent and set up a solar-powered Playstation. But it's probably better to stay close to other people though and not give into the myth that survivalism means holing up from the horde of zombies. A lot of people would starve, though, so don't worry about looking weird if you store up a literal decade's worth of food for yourself and for barter.

Don't waste effort. Be efficient. If you must read the doomscroll, keep a budget for that. Stay on track with your current work. Don't up and leave your apartment in a rush without thinking about your possessions, for instance.

Despite my general biases on the topic I am super uncertain what's the right amount of prepping to do for the marginal person. I do think it's nonzero. What I am saying is you have a choice between gradations of failure and that if you can optimize, then there's a lot you can optimize that can also help others, and if in all likelihood you can't, then satisfice with decisiveness as early as you can. Think more like a gamer with nothing to lose than an ape avoiding the crushing weight of total paralyzing ruination. I am very much writing this to tell myself this.

After I publish this I plan to not think about this for a week, for I will be at Schelling Point. This is a tradeoff I want to make even though there are alarm bells in my head telling me there is smoke coming from under the door. I want to devote myself to what is meaningful despite the geopolitical uncertainty. All we have to decide is what to do with the time that is given us.

If you want to win, then maybe come up with a game-theoretic proof for a scheme of Mutually Assured Cooperation that the world can switch over to. 

But if we don't solve the world's problems in time... Have you no dignity?

New Comment
8 comments, sorted by Click to highlight new comments since: Today at 5:28 AM

I don't think I've seen a full counter-value nuclear exchange scenario where most of the deaths come from the initial nuking itself. The highest risk of death actually comes from radioactive contamination of food supplies (America bases much of its strategic nuclear arsenal in sparsely populated farmland) and maybe nuclear winter reducing crop yields. This means that as long as you don't starve to death in the first few years, you've mostly made it.

Therefore, having a year's worth of food and water per person should be the cheapest way of hedging your bets against a nuclear holocaust. That comes to 100-200kg of grains and beans in a sealed container, which Amazon says is ~$200. Water may be a bigger issue, but fortunately, most people live near a well or river. Water purification tablets are something like $50 for 300 gallons, which should be enough for a year of drinking per person. Obviously, bottled water would be also be needed to wait out the first few weeks of fallout, but a lot of people already buy their caffinated drink of choice in bulk, and 100 Monster Energy in a corner works just as well.

That's most of what you need to survive hunkering down for a full year. Maybe also buy a gun and join a militia for your security needs? Honestly, it doesn't seem too hard to achieve, even for someone living in an urban apartment. Biggest concern may be keeping your provisions secret, since your starving neighbors will get jealous eventually and you can't feed your entire building.

What's the ~$200 food item you saw that would last a person a year, can you link?

(Edit: I see a 30-day container for about $200. It seems to me that $200 buys a month's worth, not a year's worth?)

Bulk rice and beans in vacuum-sealed bags, plus multivitamins. You won't be having fun, but you'll survive the year. It's pretty easy to stockpile if you're going to be eating those anyways. The main takeaway should be to buy non-perishables you normally consume in bulk when you think there is an elevated risk of nuclear war and draw down those provisions when you think the risk has passed. This means that the cost of preparing for nuclear war is the interest you would otherwise earn on <$500, which is pocket change per year. The needed storage space for a person-year of food and water is also under a cubic meter, so that shouldn't be too much of a concern.

You mentioned a specific Amazon price point that I found surprisingly low and is cheap enough to meaningfully affect strategic plans. The cheapest pasta I've seen is on WebstaurantStore but the shipping doubles the cost to $2/lb approximately. Did you have a link for the $200/year product(s)?


These are a bit more than $1 per pound for Amazon and half that for the bulk supplier. Dried rice is around 3700 calories per kg, and even a fit adult male on bed rest only needs around 1500 a day (I think). So 300 pounds should be enough for a year. My math was a bit off since I confused pounds for kilograms initially, haha.

If I survive for an extra two weeks, that's about 4% of a QALY, which is about $4,000. So if there's a 1% yearly risk of nuclear apocalypse, it's worth spending $40/year for that chance of extra time. I don't think there's any additional return from my death being dignified. If I try and die I'm still dead.

In case of nuclear apocalypse humanity appears to be best off if most people die quickly, because this will improve the ratio of food to mouths and reduce the risk of long-term problems where, for example, we eat all the fish and go extinct.

If you have a model where there are collective benefits from one more family surviving two more weeks I'm interested in hearing it.

I consider it collectively important that alignment researchers and their +1s survive, as well as other x-risk researchers and probably other cause areas.

So if there's a 1% yearly risk of nuclear apocalypse

Some think the number is much higher than priors due to current events. You're also not factoring in that that yearly percentage adds up, and a lot of preparations are a one-off action that benefits future you (assuming you don't dig into your backup food).

The 1% number was intended to be illustrative, not definitive. I'm not a nuclear risk expert. The QALY figure may also vary. A $40/yr cost could be a $400 investment that depreciates over ten years. In that case I would value it based on the projected risk over ten years. I'm not seeing additional value in "nuclear dignity points" above these admittedly hard-to-calculate figures.

To preserve x-risk research during civilizational collapse I think attempts to preserve information and insights would perform better than attempting to preserve individual researchers, especially since it could be done in parallel with preserving other information that aids recovery.