In the early Cold War, weapons were not very accurate and intelligence collection often wasn’t very good or up to date. If you were a war planner, the advent of the hydrogen bomb solved a lot of problems for you: even if you didn’t know exactly where a target was, and even if your weapon could miss by hundreds of meters, with a large enough nuclear yield, you could still guarantee the destruction of whatever targets you hoped to hit.

But with giant yield weapons, comes massive collateral damage, likely far beyond what is required to deter decision makers. The difference between a TSAR bomb (or its modern equivalent) and the lowest settings of a mini-nuke is still an order of magnitude larger than the difference between the conventional “mother of all bombs” and a hand grenade. The Beirut explosion last year was the size of the hand grenade blast in this analogy… but if you feel the need to visualize the differences more precisely, you can use NUKEMAP.

Massive yield weapons aren’t just big explosions: they produce more radioactive nuclear fallout that can spread over thousand miles and kill people for years, larger (though less efficient) electromagnetic pulses that can disable electronic grids and large electronic devices over millions of square miles, and mushroom clouds that can reach the stratosphere, contributing to nuclear winter.

With the proliferation of sensors, precision weapons, and fusion of information by narrow artificial intelligence, giant weapon yields may no longer necessary to assure deterrence. In my view, this presents the opportunity the reduce the risk of nuclear winter, but how would you achieve that?

In “Winter-Safe Deterrence” Seth Baum argues that limiting the global nuclear arsenals to ~50 weapons may be a path that allows a degree of deterrence while not threatening enough cities to threaten the global climate. To dig in to more detail and slightly contest the paper: it is not a limit of 50 weapons per say that assures the climate, but rather how large the weapons are and what they are aimed at. Since large yield weapons can loft dust straight to the stratosphere, they don’t even have to produce firestorms to start contributing to nuclear winter: once you get particles that block sunlight to an altitude that heating by the sun can keep them lofted, you’ll block sunlight a very long time and start harming crop yields. To get soot high enough otherwise, certain kinds of cities with large enough fire loads/fuel density would have to be hit to produce firestorms that loft soot high enough that it won’t just quickly fall or be rained back out of the sky. If a bomb knocks over buildings, a lot of the fuel often won’t be available to burn as collapse prevents oxygen from getting to it. With large yield hydrogen bombs where the radius of burning can significantly exceed the radius at which buildings will be knocked over, fires can do much better at sustaining firestorms and burning all available fuel. Lastly, for rural and wild lands, fuel density is normally too low for firestorms: meaning that missiles landing in the middle of nowhere trying to hit silos, submarines, air bases, and mobile launchers probably won’t wreck the global climate, but if they did, it would require very large yields.

Overall, my argument is that nuclear winter is uncertain with current arsenals and targeting plans, but that it could be made extremely unlikely without getting rid of nukes, or even shifting to very small numbers of nukes. Countries with lots of weapons are more likely to employ counterforce targeting to limit damage in the event of nuclear war (aiming at military forces, command and control, etc.) while if you have fewer weapons you are more likely to do counter value targeting for pure deterrence value: neglecting precision, increasing weapon yield, and targeting cities (how you produce nuclear winter). Low yield weapons are still orders of magnitude worse than conventional weapons and can provide plenty of deterring power. If you thought current nuclear deterrence was insufficient, would your solution be to replace all warheads with Tsar Bombs? Probably not, you could increase the yield of some weapons selectively if you have intelligence problems, or instead get better intelligence, increase precision, and get more low yield weapons.

While arms races are undesirable, arms races between superpowers raise the cost of military competition for everyone else, deterring some other arms races and attempts to get nuclear weapons in the first place. The fewer actors there are in military competitions, the fewer security dilemmas you have, and the fewer points there are to initiate conflict. If you think mutually assured destruction works, but that eventually someone will make a mistake or irrationality will change the calculus, then the last thing you want is for everyone to have nukes and magnify the odds that such mistakes will happen. Likewise, if you want to negotiate to reduce the risks of escalation or catastrophic damage then the last thing you want is to multiply the number of parties that have to negotiate. If early attempts at building nuclear arsenals can be halted even fairly late with low collateral damage, that may serve as a last line in deterring nuclear proliferation beyond diplomatic measures and sanctions. In general, I think it is good for the world that nukes are excessively expensive: thought experiments about what happens when they aren’t have fairly dystopian implications. In the grand scheme of things, even through the Cold War the world kept spending less of its wealth on weapons, and even with the Reagan build-up, the costs were highly concentrated on the Soviet Union, bringing the competition to a close for sometime without war. Cost imposition strategies that disproportionately punish totalitarian regimes, advance technology for democracies, and let everyone else enjoy more economic growth sound good to me (when they work).

Bringing this all together, I think a good path for nuclear modernization would be to generally reduce nuclear weapon yields while increasing precision: this makes the weapons more credible that you will use them, and enhances deterrence in that manner while decreasing the odds of global nuclear winter if something ever goes wrong somehow. For tiny states, this creates a far more credible threat of counter force nuclear targeting: dis-incentivizing proliferation, while for large states targeting problems would become far more difficult due to the number of weapons, and thus counterforce targeting would be much more difficult. I don’t think smaller weapons do any less good of a job at deterring decision makers: at point blank range these weapons produce absurdly high overpressure that will crush any bunker, removing the need for extreme yield weapons to take out out hardened bunkers while missing by hundreds of meters. I think it is better if deterrence shifts to deterring decision makers and militaries rather than inherently threatening entire societies. I don’t hold these ideas with extreme certainty, but they should at least be debated, and if wrong, thoroughly debunked.

I think there are two reasons this debate doesn’t really happen: the first is political warfare, and the second is the asymmetries in attachment to reality and interests between military strategists and activists.

In my view, political warfare and active measures created a lot of problems with anti-nuclear activism during the Cold War: activism can get weaponized, hijacked, or coopted in a naïve risk increasing directions, by those seeking to increase their own power, and once groupthink gets started, it can keep going with its own force. Deterrence held up by neutron bombs would have posed far less long-term radiation and nuclear winter risk, but Soviet hijacked peace activism ended that option by rebranding neutron bombs “capitalism bombs.” By bolstering anti-nuclear campaigns against neutron bombs (asymmetric U.S. advantage vs. USSR), and nuclear power (more NATO energy independence) Soviet influenced activism seems to have directly contributed to both planetwide nuclear risk and climate change. Maybe the peace/green movements would have gone that way on their own… but it seems strange when the outcome achieved aligns with higher risk/worse environment and more relative power for an authoritarian state. Beyond these efforts, the Soviet Union also pursued active measures to shift peace group messaging from “no missiles” to “no more missiles” to lock in their advantages in Europe. This made negotiating weapons out far more difficult, and to this day Russia has a huge number of tactical nuclear weapons that people usually don’t count/ignore because they aren’t counted by arms control treaties (I have made this mistake before). Overall, influence operations and disinformation campaigns do target existing rifts in society, and sometimes rely on locals blowing them out of proportion to have effect… but anti-nuclear activism easily could have focused on other directions and doing so would likely have been much better for the environment.

Why did political warfare work at all in hijacking peace activism when often it is so ineffective? Though some peace activists had sympathy to communism, many just were good people that wanted to reduce the risks of millions or billions of people dying. This is where I think asymmetry of interest and attachment to reality comes in: those that enter the military are more likely to have a competitive mindset on military subjects, while peace activists are far less likely to. The military mind may miss opportunities for de-escalation, while the pacifist mind will miss the entire game because it has little interest in plotting out in detail how to invade a country, win a war, or thoroughly imagine the motivations of someone who does such things. Both types of minds will miss many opportunities: the peace activists because they aren’t wrestling with the concrete details of military competition, and the military planners because they have a psychological disposition toward winning with their preferred tools. If you want things to improve, the sort thought leaders you’d want in peace activism would be able to keep their larger goals in mind while thinking much more deeply about the specifics of technology and competition.

Overall, the particular form of arms control I’m arguing for here may be extremely difficult or infeasible due to enforcement issues, but that same problem doesn’t stop activists from pushing for arms control on subjects where enforceability is even harder yet and where the benefits aren’t as clear (e.g. lethal autonomous weapons). This post isn’t terribly detailed or rigorous, but it’s pretty far ahead of most conversations I have had on arms control in the past 5 years: I think it’s time for a new look for ways to improve nuclear arms control, and for the best arguments to get fleshed out and win.