Summary: With Russia's invasion and Putin's gestures towards using nuclear weapons, we seemed to have passed a threshold where it makes sense to spent at least a couple of days thinking about this. Me and a couple of friends are thinking about what might be reasonable approaches from a personal decision-making perspective and would love your input.

There's a war between Russia and Ukraine and Putin (who has control over the largest amount of nuclear warheads) appears to be threatening with nuclear war if "The West" interferes. Specifically, he says "Whoever tries to hinder us should know that Russia's response will be immediate, and it will lead you to such consequences that you have never encountered in your history" (source).

Additionally, he appears to be paranoid and believes (or at least want his people to believe) that "Western Patriots" will bring weapons of mass destruction to Ukraine and "help it acquire these weapons to create yet another threat to our country.". Given this apparent paranoia, thirst for power, and his actual power, the risk of some form of nuclear attack seems to pass a threshold to think about it for at least one or two days.

From a personal decision-making perspective, it seems reasonable to optimize for personal safety (in the case of nuclear war) and the ability to contribute to the long-term flourishing of our civilization - including this threat.

Concretely, a couple of friends and me are considering changing location away from the UK and mainland Europe in a easily reversible and low cost way (e.g., home office from Morocco) while recurringly reassessing the situation for signals of escalation or de-escalation.

The reasoning behind moving away Europe is due to the likelihood of a first-strike on states with nuclear weapons (UK and France) and the subsequent outbreak of panic with an increased difficulty to leave the country and questionable ability to maintain a robust infrastructure for living (e.g., food supply).

PS. There's also a lot of interesting predictions on metaculus.

What do you think about the reasoning and the tentative personal implications?

Also, do you have any input on how a highly skilled personal and professional coach can contribute to the mitigation of this risk?


New Comment
38 comments, sorted by Click to highlight new comments since: Today at 7:36 AM

Putin seems to have ordered deterrence forces (which include nuclear arms) to be on high alert, roughly an hour ago.

Can someone weigh in about how unprecedented this is? Some media coverage has compared the severity of the current situation to the Cuba Crisis, which would be extremely alarming if remotely true. 

I'm not overly concerned with the news from this morning. In fact I expected them to raise the nuclear force readiness prior to or simultaneously to commencing the invasion, not now, which is expected going into a time of conflict/high tension from normal peacetime readiness. I had about a 5% chance this will escalate to a nuclear war going into it, and it's not much different now, certainly not above 10% (For context, my odds of escalation to full countervalue exchange in a US intervention in a Taiwan reunification campaign would be about 75%). Virtually all that probability is split between unfavorable developments dragging in NATO and accidents/miscalculation risk, which is elevated during tense times like this (something like, if the Russians had misinterpreted the attack submarine which entered their territorial waters last week as being a ballistic missile submarine sneaking up close to launch a first strike, or an early warning radar fluke/misidentification being taken seriously when it would've been dismissed during peacetime, either of which could've caused them to launch on warning).

Unintentional nuclear exchange will have no preceding signs, but unfavorable developments will, for example a NATO shootdown of a Russian plane or Russian fire straying over the border killing NATO troops which begins an escalation spiral. If we start seeing such incidents being reported, I would tell all my LW/EA friends to get the fuck out of NATO cities they're living in immediately.

Finally, here's a good thread on the heightened alert news. The US side still seems quite wary of nuclear escalation and hasn't even announced a reciprocal raising of DEFCON, which contributes to my relatively sanguine assessment of the situation at present. Keep in mind that this means I still have at least 5% on it, and that I don't feel comfortable living in a NATO (especially US) downtown core at ANY time, not just now, due to the significant and everpresent risk of accidental/sudden nuclear attack. For example, I'd really like it if MIRI would move out of Berkeley where they'd be instantly vaporized whenever nuclear war broke out, and now might be an even better time for them to take a temporary vacation outside the city (see my other comment in this thread), but I don't think the situation is far more alarming than usual just yet.

That seems way too high to me: are you willing to bet at 5%? (For epistemic purposes only. I hope no one reading will be offended.) If so confirm here and PM me on Forum to figure out the details.

Thank you! 

5% does sound very alarming to me, and is definitely a lot higher than I would have said at the beginning of the crisis (without having thought about it much, then). 

While 5% is alarming, you should notice that abukeki did not update much because of the crisis (if I understand it correctly), and so if your prior is lower than it should possibly stay lower.

I am starting to worry that the possibility of Russia using conventional or perhaps more likely tactical nuclear weapons in the Ukraine conflict is real.   My concern is one largely based in this article by Francesca Giovannini : A hurting stalemate? The risks of nuclear weapon use in the Ukraine crisis

For Mr. Putin, any kind of losing the war with Ukraine seems like a non-option, given his domestic situation and the possibility that his regime could come to an end. 

The article outlines three assumptions that those who don't think Russia will use nuclear weapons make: "that Russia has a strong interest in not destroying Ukraine, because Putin wants to occupy it; that even though Putin is a thug, he is not a crazy enough thug to break a taboo against the use of nuclear weapons in war, a taboo that has held for 75 years; and that there are plenty of other options that the Russians can exercise in subduing Ukraine. "

However, I think all three of these assumptions are suspect. 

Regarding the first assumption, we have already seen Russia step up its attacks on civilian infrastructure in the last two days. For Putin, a victory of some kind (like regime change) is paramount, and the longer the invasion takes, with weapons and aid flowing to Ukrainian fighters, the harder that possibility becomes. 

Regarding the second assumption taboo, there are reasons to suspect Putin is willing to break the nuclear taboo. He has already threatened the possibility of not renewing treaties with the US to limit the number of nuclear weapons, he has sidelined many of his advisors, and he is an authoritarian clinging on to power and trying the restore Russia's status as a superpower. He may see all options as being on the table. 

Finally, with regard to the third assumption, it's true that Russia's conventional forces are vastly stronger than Ukraine's, but we have seen the fierce resistance of Ukrainian fighters, and a prolonged occupation would require many more troops. The use of a tactical nuclear weapon would test Ukraine's and NATO's resolve and signal his willingness to do whatever it takes to win the war.

I'm not an expert but it does seem highly alarming.

Also, beyond the purely personal, are there any actions that could be taken by individuals right now  that would have a positive impact on humanity's chances to recover, conditional on nuclear war? 

Some (probably naive) ideas: 

  • Downloading and printing vital information about how to rebuild food supply and other vital infrastructure (so that it can easily be accessed despite varying degrees of infrastructure collapse); I guess ALLFED's articles might be a good starting point (even though I could not quickly find any distilled strategy document/user guide for their research)?
  • Increase the likelihood that you will be able to distribute this information: Ensure survival, build local networks, practice leadership skills etc.
  • Make sure others do the same: While there already seem to be a lot of preppers, I do not know whether their culture emphasizes strategies for rebuilding a flourishing civilization over mere survival.  "Altruistic prepping" might be a relatively neglected niche (or not, these are just off the cuff thoughts...)

Also, see this 60 second advice article:

Some good reading for the bunker: Lewis Dartnell's The Knowledge: how to rebuild our world from scratch.

I'm reading this book now and finding it very good! I'm surprised because most books in this genre I've tried lately have been really bad and I couldn't bear to continue reading them. This one is fun because in addition to being practical it takes you on a fast route tour of just the essentials from the tech tree we've developed over human history, without the long winding arbitrary delays we had historically, which is interesting as well as potentially useful. Makes me think Cyberpunk is more likely in some ways (for some types of disaster) than I'd realised. It should be said that the book assumes technology is somewhat intact but most (but not all) of the population are gone.

I think some form of this could be valuable, noting Sebastian's point that decreasing risk should be the main priority. It struck me reading the main article that the tendency for EAs to congregate to some extent geographically poses a challenge from a long-term perspective. Oxford, the community's beating heart, is uncomfortably close to London (the obvious civilian target) and Portsmouth (home of the Royal Navy, probably second-top priority military target), meaning a large fraction of the community would be wiped out in a nuclear war. It might be prudent for EAs who can work remotely to set up 'colonies' in places unlikely to be devastated by a nuclear exchange, to provide resilience. 

I'm guessing that with the number of nukes Russia has, most cities in NATO above ~100k people will be targeted, including Oxford. In contrast, basically everywhere in the Southern Hemisphere is likely safe (New Zealand is a favourite hide out for elites I hear).

I appreciate the line of thought but it also seems important to focus on decreasing the risk and/or severity of nuclear war.

Seems worth mentioning Russian nuclear policy. (I believe my interpretations/translators are fairly accurate but might miss diplomatic/legal nuance.)

Russian war doctrine (#27) of 2014 prohibits the first strike unless WMD are used against it or its allies or when aggression with conventional weapons greatly endangers Russia's existence:

Российская Федерация оставляет за собой право применить ядерное оружие в ответ на применение против нее и (или) ее союзников ядерного и других видов оружия массового поражения, а также в случае агрессии против Российской Федерации с применением обычного оружия, когда под угрозу поставлено само существование государства. Решение о применении ядерного оружия принимается Президентом Российской Федерации.

This was further clarified in section III of the 2020 nuclear weapons policy. Conditions are (a) launch of ballistic missiles against Russia or allies; (b) use of nuclear or other WMD against Russian or allies; (c) infiltrating with critical infrastructure, which might disrupt the second-strike capabilities; (d) aggression with conventional weapons greatly endangers Russia's existence.

  1. Условиями, определяющими возможность применения Российской Федерацией ядерного оружия, являются: а) поступление достоверной информации о старте баллистических ракет, атакующих территории Российской Федерации и (или) ее союзников; б) применение противником ядерного оружия или других видов оружия массового поражения по территориям Российской Федерации и (или) ее союзников; в) воздействие противника на критически важные государственные или военные объекты Российской Федерации, вывод из строя которых приведет к срыву ответных действий ядерных сил; г) агрессия против Российской Федерации с применением обычного оружия, когда под угрозу поставлено само существование государства.

Living in New Zealand, there appears to be a near-consensus that it's one of the best places in the world to be in the event of a nuclear war and/or nuclear winter. So from a personal perspective I'm pretty well-optimised.

From an EA perspective,  however,  I've wondered for a while about the cause area of an "NZ bolthole for x-risks".  If New Zealand is indeed likely to come out relatively well in the event of many disasters, then it seems there could be significant value in investments that have high value in some of the more likely scenarios.

 This hasn't got much beyond some idle chatter with other EAs, and I'm quite uncertain as to exactly what the highest value investments would be. I think exploratory research at least is warranted. 

However,  with the Ukraine situation the question (at least with respect to nuclear-risk) has suddenly become a lot more time sensitive. Rigorous research may produce answers too late. What low-cost (>$1,000) actions could we in New Zealand take that would have an outsized impact in the event of a nuclear war in the next 3 months? 

Nuclear winter is a very unlikely, highly conjunctive theory which requires many independent things to ALL happen perfectly, which are already individually suspect. E.g. that cities will all firestorm after being hit by airburst detonations (which itself relies on assumptions like adequate fuel loading per square meter, collapsed structures from the air blast not suffocating the oxygen, etc.), that this will burn in a way producing lots of black carbon, that this carbon will be nearly all lofted into the stratosphere, that this will block a high percentage of sunlight, that the carbon will persist there for many years, that it has to happen during the summer in warmer climates, etc.

Altogether, I think it's unlikely enough a possibility to be ignored in planning, which removes that utility of New Zealand, although it may still have value as a sheltered, unaffected place to escape the chaos and societal collapse of other countries following nuclear war.

I'm not sure if the bolthole idea is referring to an escape for EAs in particular or relocating as many people as possible in general, but the former is something I've considered which I think needs more discussion. Keeping valuable EAs (AI alignment researchers in particular) alive both through the initial exchange and the chaotic period which follows is extremely valuable to improve the world that follows a nuclear conflict, especially to improve the relative trajectories of AI alignment and capabilities by ensuring the built-up base of alignment knowledge & talent is not destroyed.

As such, people should be developing protocols to evacuate those researchers preemptively at times when nuclear war looks likely to ensure they're not killed in the initial detonations (or better yet, permanently relocate them to places with no or much lower risk of nuclear attack in the first place); as well as have preplanned long-term locations where they can ride out fallout and societal collapse/civil violence, stocked with enough food for years etc. I believe it's fine if such locations are in the US given no nuclear winter, and it would be difficult to travel to somewhere like NZ in the post-attack environment anyway.

What do you think about the conclusions of How bad would nuclear winter caused by a US-Russia nuclear exchange be? - EA Forum ( ?

I expect that a US first strike against Russia would involve around 1,300 nuclear weapons used exclusively on counterforce targets (90% confidence interval: 670 – 1,700), and that a Russian first strike against the US/NATO would involve around 1,100 nuclear weapons (90% confidence interval: 750 – 1,200), also exclusively used on counterforce targets.
I expect that a counterforce second strike would look pretty similar to a first strike, though it’s possible that a second strike by Russia would be somewhat smaller than its first strike […]
In effect, there’s a 90% chance that the actual change in temperature caused by a nuclear exchange between the US and Russia will be between -1.6 degrees Celsius and -4.2 degrees Celsius at its most severe. Similarly, there’s a 90% chance that the actual change in precipitation caused by a US-Russia nuclear exchange would be between -9.2% and -24% during the worst period of climate disturbances. […]
By my estimation, a nuclear exchange between the US and Russia would lead to a famine that would kill 5.5 billion people in expectation (90% confidence interval: 2.7 billion to 7.5 billion people).

Don't buy the stuff about expecting a famine that kills billions at all? Especially since she didn't seem to have dug into the actual criticisms of the nuclear winter theory in her post sequence, e.g. the independent components of the theory.  I think very likely (>90%) there won't be any change in temperature at all, which will be the case if any of those components fail. And as I understand it she has since updated towards being less bullish on it since those posts, and people who succeeded her at RP don't think nuclear winter is that likely either.

>I'm not sure if the bolthole idea is referring to an escape for EAs in particular or relocating as many people as possible in general,

Perhaps "bolthole" is not quite the term I'm looking for, at least in the sense of primarily being about relocating individuals. Rather, I'm using it as a catch-all term for all "post-apocalyptic" preparations. A seed bank and/or data bank located in New Zealand would be good examples. 

It should be mentioned that the border to NZ will be closed to most people (due to covid) until July, with a few exceptions opening up around March

  • Skilled workers earning at least 1.5x the median wage may be eligible to be granted an ‘other critical worker’ border exception.
  • [critical services include]
    • food production and its supply chain
    • key public services like health and emergency services
    • lifeline utilities such as power and water supplies
    • transport
    • critical financial services
    • news media
    • social welfare
    • human and animal health and welfare. (EA-relevant?)

I'm not super worried. Maybe this is because I am old enough that I grew up with a perception that nuclear war could happen at any time and unexpectedly kill us all. The current threat level feels like a return to the Cold War: something could happen, but MAD still works and Putin, like everyone else, doesn't really have anything to gain from all out nuclear war, but does have something to gain from playing chicken. So we should expect a lot of posturing but probably no real action, except by accident.

In think the largest risk of nuclear weapons comes from the use of tactical nukes being used in the conflict zones. I would expect Putin to use them if he felt desperate enough, especially since he would use them on Ukrainian soil. But presumably no nukes would be deployed on NATO countries or Russia itself since that would trigger all out nuclear retaliation. So most of the nuclear risk probably falls on people literally within Ukraine.

There is a theory that Putin is terminally ill and could therefore be open to taking the rest of the world with him. I don't know how much weight to put on it.

Even if he wants to do that, his power is not absolute. I'd expect/hope for his generals to step in if he tries something like that, perhaps using it as reason for a coup.

I feel like it makes a lot of sense to have a plan like yours but I am unsure about when will be the right time to go trough with it. It would probably be valuable to have some event x that would trigger that plan immediately. I just wonder what that could be. Declaration of war on any NATO state? Putin threatening to use nukes? (Which seems to be happening right now) The metaculus prediction of WW3 before 2050 rising above 30%? (Currently 22%) Any ideas?

Agreed, it seems to be escalating fast. Although it is debatable whether the attack on Ukraine was expected, what transpires is becoming more obvious, with Putin's thinly veiled threats of retaliation if NATO defends Ukraine. I have seen excellent arguments against No Fly Zone because enforcement of said NFZ necessitates actually shooting down planes.  

Which makes me wonder about the feasibility of a peaceful solution. Sam Bankman-Fried transferred money to the Ukrainian people, and Elon Musk established Starlink over Ukraine.  How effective would this modern guerilla-style be in supporting Ukrainian efforts?  It could provide support for Ukraine without the open state declaration of support which would incite escalation. 

My gut feeling is, that this is excessive. Seems to be a sane reaction though, if you agree with Metaculus on the 3% chance of Putin attacking the Baltics.

Do you agree that there is a 3% chance of a Russia-NATO conflict? Is Metaculus well enough calibrated, that they can tell a 3% chance from a 0,3% chance?

Miscalibration might cut both ways... 

On one hand, It seems quite plausible for forecasts like this to usually be underconfident about the likelihood of the null event, but on the other hand recent events should probably have substantially increased forecasters' entropy for questions around geopolitical events in the next few days and weeks. 

To me it seems that potential nuclear war is a lot more concerning than proximal country invasion as it would escalate much more slowly. 

While it might be excessive I'm more worried about status-quo bias. Most of us haven't experienced a serious war and will be overly focused on how good and safe things seem now plus the highly inconvenient short-term personal implications.

What do you put P(nukes used | Russia-NATO conflict)? Intuitively it seems high to me. There hasn't been a direct conflict between them yet; basically down to MAD.

Metaculus is (as of posting this comment) on 5% for Nuclear Detonation Fatality By 2024 .

Why was this prediction removed?

You have to sign-up for the nuclear risk tournament to see it (which doesn’t come with any commitments to participate afaict).

Moreover, Metaculus is currently on 10% for  >100 nuclear detonations given one fatality by 2024. Seems a bit of an underestimate. 

The Metaculus link suggests it may also be worth considering (especially if you are in the US), how a cyber attack might effect you.

This page, from Rob Wiblin, has been shared on Twitter recently. Contains some advice some minimal version of preparation (e.g. buy potassium iodide tablets):

Does anyone know where to buy potassium iodide tablets? I can't find any seller which is not out-of-stock and which works on the internet

In Germany at least you can get it at a pharmacy. I'm planning to get some this week.