To comment or view community discussion of this article, see the cross-posted version on the EA Forum.

Summary: it seemed interesting to reframe a variation of eca’s original EA Forum question – What would have happened every country got nukes in 1950? – as a way of interpolating between the real world and Bostrom’s “Vulnerable World” paper. Bostrom notes that an extreme worldwide police/surveillance state might be our only option to contain a sufficiently dangerous technology. In this post I explore the spectrum of potential real-world responses to a toned-down version of the extreme threat posed by “easy nukes” in Bostrom’s original thought experiment. In particular, I try to interpolate between our existing multi-polar world of many nation-states, and an imagined maximum-control world of unified government, imagining what a return to cold-war-esque or neo-imperial dynamics would look like.

---

Bostrom’s Vulnerable World Hypothesis

In his paper The Vulnerable World Hypothesis, Nick Bostrom invokes a thought experiment about “easy nukes” – what if it were possible to create incredibly destructive weapons with extreme ease? Considering a “counterfactual history in which Leo Szilard invents nuclear fission and realizes that a nuclear bomb could be made with a piece of glass, a metal object, and a battery arranged in a particular configuration,” the situation seems hopeless – although the information could perhaps be kept secret by classifying and policing the relevant academic fields, eventually the knowledge would get out. In a world where anyone with the right instructions can create a nuclear weapon, the imminent collapse of civilization seems almost inevitable. Even the best-case scenarios involve zealously purging the world of glass and electrical equipment, totalitarian surveillance, and a quickly-growing count of irradiated craters.

Bostrom uses this thought experiment to illustrate the general principle that, since the capabilities of future technologies are unknown, at any point we could be about to stumble upon a discovery that “makes the devastation of civilization extremely likely, unless civilization sufficiently exits the semianarchic default condition” of limited preventative policing, limited global governance, and diverse motivations.

In this post, I want to use a modification of Bostrom’s thought experiment to explore the specific dynamics of nuclear proliferation. What if nuclear weapons were intermediate in difficulty between metal/glass/battery “easy nukes”, and real nukes with their highly-enriched uranium encased in shaped charges of exotic, perfectly-timed high-explosives? What if all you needed for a nuclear bomb was to pack some ordinary, unenriched uranium-238 into a box of TNT?

Semi-Vulnerable: U-238 Is All You Need

Following the EA Forum question by eca that originally prompted my thoughts, we’ll imagine that the secret of semi-easy nukes was revealed to the world in 1950. Disclaimer: I don’t have any kind of special insight into nuclear strategy, policy, or technology – I’m just a typical internet layperson trying to think things through based on what I know. Corrections and comments are welcome!

In the real world, the primary focus of non-proliferation efforts has been to strictly control access to the highly-enriched uranium-235 and plutonium necessary to construct a bomb. If nuclear weapons could get by with only uranium-238, there would be no need for the large, high tech enrichment facilities and centrifuges that arduously concentrate natural uranium (99% U-238 and 1% U-235 by volume) to a level of over 80% U-235. Besides the question of obtaining enough fissile material, any would-be nuclear power must also solve the engineering challenges involved in creating a bomb. These engineering challenges are significant and expensive, but are considered within the reach of normal industrialized nations. But, if a nuclear explosion could be set off using ordinary construction explosives piled casually around the uranium, the engineering challenge of creating a nuclear bomb would disappear. In a world with semi-easy nukes, our civilization would be deprived of its most important anti-proliferation strategies. The power to create a nuclear weapon would be extended, not just beyond the club of rich industrialized nations, but to literally any person or group who came into possession of a few kilograms of uranium.

Unlike in Bostrom’s original thought experiment, however, this seems like a potentially survivable event for civilization. The world in 1950 was not already full of easily-accessible uranium the way it was full of glass and batteries, and uranium is naturally rare. So the most obvious focus for nonproliferation efforts would be a complete and zealously enforced ban on uranium in any form, combined with government seizure of all uranium ore mines, and a permanent ban on atomic power.

Unfortunately, uranium is widely distributed – there are large deposits in North America, in the USSR, in China and India, and in several sites within Africa. Smaller deposits are also present around the mediterranean and in south america. Even if the world gave its best effort, it would be much easier for determined countries to acquire unenriched uranium for semi-easy nukes than it is for them to acquire highly-enriched fissile material in the real world. Unlike the metal/glass/batteries scenario, the world’s uranium supply chains would likely remain pretty secure most of the time, so nuclear weapons would usually remain the purview of nations rather than lone individuals. But we would be living in a world where any nation that wanted it could have the A-bomb.

Tactical Details

What would happen in this world? Fortunately, in order to deliver a bomb to the target, nations want a variety of missiles, airplanes, etc, which must be bought or built. If every nation in our scenario had been magically gifted an arsenal to match the USA’s current nuclear triad of bombers, subs, and ICBMs to accompany their new warheads, the world would probably end in fire pretty quickly. But in 1950, not a single nation had invented ICBMs, and few countries even have long-range bombers.

When everyone who wants already has nukes, these other parts of a nation’s nuclear capabilities become much more important. If you are a poor African country in 1950, you don’t have intercontinental missiles ready to rain death anywhere on earth at a moment’s notice. More likely, you just have some loose warheads that you can hope to sneak into another nation’s port in a disguised container ship. Just giving everyone nukes would not put countries on an even, mutually-assured-destruction playing field. On the plus side, this disadvantage in delivery systems would quickly become a second pillar of nonproliferation efforts. Unfortunately, the resulting strategic instability as different countries jockeyed to gain advantages over their neighbors – more accurate missiles, stealthier submarines, hair-trigger alert systems, second-strike capability, etc – would probably become one of the main drivers of conflict.

Lack of second-strike capability, in particular, could prove very harmful, since nuclear strategy was much apparently more combustible before the introduction of stealthy submarine fleets. As Bostrom notes in his paper,

It is widely believed among nuclear strategists that the development of a reasonably secure second-strike capability by both superpowers by the mid-1960s created the conditions for ‘strategic stability’ (Colby and Gerson, 2013). Prior to this period, American war plans reflected a much greater inclination, in any crisis situation, to launch a preemptive nuclear strike against the Soviet Union’s nuclear arsenal. The introduction of nuclear submarine-based ICBMs was thought to be particularly helpful for ensuring second-strike capabilities (and thus ‘mutually assured destruction’) since it was widely believed to be practically impossible for an aggressor to eliminate the adversary’s boomer fleet in the initial attack.

The stabilizing technology of nuclear submarines would not be an option for many poor (or landlocked!) nations, thus lengthening the perilous early era of first-strike-friendly nuclear strategies.

With easy-to-make nukes but tightly-controlled missile technology, we would probably also see greater tactical emphasis on very small bomb designs for local, tactical battlefield use – this would further erode the “nuclear taboo” boundary between nuclear and non-nuclear weapons.

In the hopes of avoiding retaliation, anonymous attacks might become common – sneak weapons into the enemy’s cities and blow them up without warning. To counter this, powerful nations would become much more paranoid about their ports/borders, and accept commercial trade only from within their empire and a few trusted allies. (It’s almost surprising that anonymous nuclear attacks haven’t ever occurred in real life, since they’d be so easy to pull off. But in real life, the list of suspects is very short – with semi-easy-nukes, it would be much easier to hide.) International trade was only 8% of world GDP in 1950 anyways, versus 14% in 1913 and 24% today, so it wouldn’t have hurt too badly to pull up the drawbridges.

---

A New Geopolitical Equilibrium?

Thanks to the lack of ICBMs and rarity of long-range bombers in 1950, society would have a bit of time to settle into the new geopolitical equilibrium.

I expect that some dynamics would be similar to nuclear strategy in the real world: nuclear weapons would make wars less common, but the few wars that did happen would have a risk of being much more devastating. If you’re Saddam Hussein and you’re pondering whether to start the Iran-Iraq war, maybe you hold back for fear of a nuclear exchange, and start your own local cold war instead. I’d expect that maybe we have only a small fraction as many wars, but some of the remaining wars would involve nuclear weapons.

While interstate wars might decrease dramatically, I’m not sure if the frequency of civil wars would change as much. I could easily imagine a dark alternate history where maybe half of civil wars involve some small group (either rebels or an embattled remnant of government) getting their hands on a stray nuke and setting it off in the middle of a city.

Of course, culture would change worldwide to value stability and safety more, which you’d think would reduce civil wars, but this might have complicated effects – like, if people have higher tolerance for totalitarianism because they fear civil war and crave stability, maybe the totalitarianism just keeps getting worse until they rebel anyways, so the rate of civil wars only goes down a little?

It could be tempting for many countries to aim for a North Korea strategy – acting crazy and threatening everyone around them in exchange for concessions. Perhaps, after a bunch of initial chaos, the world would eventually enforce a much stronger norm against these rogue nations – imagine a kind of worldwide NATO who wouldn’t have any qualms preemptively nuking dictators who seemed to be going down the North Korea path.

Overall, the idea of a mostly peaceful world where “mutually assured destruction” logic extends without incident to a multipolar mexican standoff seems incredibly unlikely to me. But I am also optimistic that semi-easy-nukes is not so terrible as to be an automatic game-over for civilization. Instead, I think this world will fall somewhere between the following two extremes, either finding a way to live with low-level nuclear conflict or transitioning towards stricter global control to suppress it:

In 1950, the world starts out looking mostly similar to our own history, except wars are less common, but those wars often involve nuclear weapons, so everyone is just nuking each other occasionally, especially during civil wars, in poorer countries with less nuclear capability, and among dictatorships rather than democracies. (Cold-war-like dynamics mostly prevail among rich and powerful nations, just like IRL) This is a pretty terrible state of affairs because many more people die, and over time a larger part of the earth’s territory, atmosphere, etc, is being irradiated. Also, if nothing big is done, I’d expect the average rate of nuclear detonations per year to increase as technology continues to advance.

In order to prevent continual low-level nuclear war as described above, powerful nations might try to coordinate much more to eliminate the threat posed by unstable and rogue nations. This transition towards world government, if it happens at all, might be peaceful but might alternately involve a very severe, full-scale thermonuclear war – either as the crisis that prompts the decision to centralize, or as a result of the decision, when the winning coalition must now seize power by potentially obliterating all the objecting countries.

It would be really depressing if we repeatedly had giant full-scale worldwide nuclear wars, getting worse and worse as technology advanced, and ALSO failing to change world governance to put a stop to it. But I guess humanity has disappointed me before, so it could definitely happen – this would be especially likely if each nuclear war has an equalizing effect, making the most powerful countries disproportionately more fractured and broken and chaotic, so after the first big war there’s no alliance powerful/functional/responsible enough to impose order and stop the next round.

The Spectrum of Global Coordination

In The Vulnerable World Hypothesis, Bostrom mentions three techniques for dealing with destructive technology: preventative policing, coordinated global governance, and reducing the diversity of motivations among mankind. In our scenario, I’m taking for granted that policing works decently well, if imperfectly. Unenriched uranium is more difficult to police than enriched uranium, but vastly easier than batteries and glass – more important than questions of “how well can uranium production be policed” will be questions of “which states will allow uranium mining and export to what other countries”, which is a governance question. On the other hand, 1950s society is in no position to start mind-controlling or genetically-engineering its citizens, so “diversity of motivations” isn’t a lever we have access to. (Also, to be effective, a mind-control strategy would have to be adopted by many countries – another governance question). This leaves us with global governance as the primary axis of variation. Society will naturally attempt to coordinate and centralize more in response to the threat, but at the same time it’s possible that nuclear proliferation naturally creates divisions between nations, on net giving us a more fractured society. To sketch out the space of possibilities, here is a sliding scale of global-governance intensity:

1. 100% literal single world government with totalitarian surveillance (likely established after a devastating WW3)

2. Closely cooperative alliance of all major governments (perhaps but not necessarily established after a devastating WW3)

3. Colonial/imperial system where there’s plenty of competition between empires, but no great power is deliberately supplying nukes to rebel groups, and rouge nations are reliably punished.

4. Proxy-war system where the two superpower teams are simultaneously suppressing and encouraging proliferation as they fight over declining european empires (like the coups and revolutions of the real Cold War but with lots of actual detonations)

5. Multi-polar world where the superpowers play defense and the developing world is defined by regional tensions (like the India-Pakistan standoff but everywhere, between eg Turkey & Greece or South Africa & Angola) and the most damage is done by medium-scale nuclear exchanges (like a war between Iran and Iraq / Saudi Arabia)

6. Anarchic, hyper-fragmented world where independence movements succeed everywhere and random nukes are going off all over whenever local mexican standoffs break down (eg Maoist China collapses into multiple warring states, Northern Ireland experiences nuclear terrorism, etc)

Note that one can be optimistic vs pessimistic about amount of deaths and conflict that happen in each scenario, so this spectrum of governance coordination is skew from the axis of good vs bad outcomes. You can be optimistic-imperialist (world government would probably happen, and it would be great), pessimistic-imperialist (world government would happen, but it would be a catastrophe and would only succeed at shutting down small wars in exchange for making the biggest wars more likely), optimistic-anarchist (everyone will be in a tense-but-peaceful standoff), or pessimistic-anarchist (semi-easy-nukes would for sure give us a chaotic slow-rolling nuclear war until there’s no civilization left, like in Bostrom’s original story).

I think this spectrum is also interesting for illustrating what “increased global coordination” might look like in practice. In philosophy papers about the far future, we can imagine 100% unified planetary governance, but what does it mean to be halfway between where we are now and that hypothetical unified state? I think people imagine it means weak agreement among the various nations of the earth (like the rough international consensus around climate change), or weak supranational governments (like the United Nations). But it could also mean a world of stronger alliances and larger empires, where there are effectively fewer sovereign powers on earth. This might not look very coordinated if those rival empires are at each other’s throats, but it is indeed a form of more unified global governance.

A Median-Case Story of Increased Coordination

I think the most likely (and perhaps best-case) scenario is that the world eventually makes a serious attempt at the neo-imperial system of 2.0, although this scenario only really works if the USSR and China decide to play along. It would obviously be a more totalitarian world than the real world, and it still wouldn’t stop a lot of nukes from going off. But I think in a neo-imperial system, although it world would be much more fragile than the real world, could still have a basically-normal future that is not guaranteed-doomed. Versus anything below 1.5 feels like it might work well for a few decades, but would slowly drift farther and farther off the rails as each nation’s capabilities advance.

This sudden increase in the importance of strictly-controlled empires might mean that decolonization of africa happens much later, if at all. Britain and France thus remain important world powers, as USA-supported policemen of the third world. Likely the USA could extend a similar nonproliferation umbrella over South America – not overtly colonial, but a more intensive presence than the meddling of the real-world cold war. (Would this be a workable plan? To what extent did European colonialism end thanks to egalitarian, anti-colonial sentiment and draining willpower to defend the empires – versus because they actually lost the military ability to exert control over their dominions?)

(Between the Soviet Union and the European empires, it’s amazing to me how few independent countries are on that map, and the power is even more concentrated when you think about how poor the rest of the world was. A much darker world, but being so centralized could have its perks against x-risk. Africa was still mostly colonized – maybe in our alternate world, Britain and France just keep it that way, the USA covers latin america, and the USSR & China exert similar nonproliferation pressure in their spheres of influence.)

I’d guess that the very first nukes to go off would probably be in Asia – either as part of ongoing Korean war, or something else involving Mao/China. They have the right combination of motive, lack of imperial oversight, and enough wealth to source lots of uranium quickly.

The danger would ramp up dramatically into the 60s, by which time many more groups have had time to acquire uranium. Practically any wars in the Middle East and Southeast Asia can now entail devastating medium-scale nuclear exchanges. In Africa and Latin America, perhaps we see a steady beat of occasional detonations (one every few years on average) by attempted revolutionaries / independence movements / coup attempts.

Earlier, I’ve talked in the abstract about how much the major powers would be willing to coordinate. In practice, I think those big questions would usually come down to how USSR and China decide to behave. Maybe they join the US and Europe as global-policing empires, keeping nuclear weapons out of the hands of allies/possessions, like North Korea and Warsaw Pact countries. Or for a darker scenario, in a world of easy proliferation, perhaps they actually decide to supply nuclear weapons to rebel communist groups – just as the USA and USSR supported coups and revolutions with conventional weapons during the real cold war.

If the USSR and China both behaved nicely, there could be a smooth path to a colonial/imperial system – this would be very repressive, meaning that there were really only a small handful of sovereign powers controlling humanity. But given the desperate circumstances, it might be the best way to keep the peace. If they don’t play along, however, western powers would face a stark choice – first-strike the communist powers in a full-scale thermonuclear war that would kill millions, and seek to establish a hegemonic system of global control in the aftermath? Or retreat into a defensive mode, which would start off looking like the IRL cold war (proxy battles in the third world, except with occasional nuclear detonations) and would slowly transition to a multi-polar world featuring occasional medium-scale nuclear wars and increasingly many threats (from rogue North-Korea-type nations) that the west would be unable to defend themselves from?

A final key question, of course, is the simple feasibility of establishing strict global governance. In the short term, the goal is to control proliferation bottlenecks like ICBM technology and uranium mines. But in the long-term, would it really be possible for a handful of nations to maintain colonial-era control over such broad empires? Semi-easy nukes would increase the desire of powerful countries to control the rest, but it would also make that task harder by somewhat levelling the military playing field. If the leading nations don’t have the state capacity to pull off global governance, then we’ll be stuck in a multipolar anything-goes world no matter what we think is preferable.

Nuclear “Contagion” Impacts the Desirability of Global Governance

One potential crux here is the importance of the “nuclear taboo” (which in the real world is intact, and in alt-1950 would have be broken almost immediately) and the idea of something like nuclear conflict contagiousness. In the real world, we have a vision that once even small nukes start flying, possibly things escalate extremely quickly, meanwhile drawing in more third parties until every country is sending missiles skyward as part of an omnicidal conflagration. I’m not sure how realistic this vision of nuclear contagion fits either our scenario or the real world – wouldn’t all third parties want to make very clear that they are staying totally out of any ongoing nuclear conflict? Obviously everyone is rightly terrified of ever finding out by testing the limits of the nuclear taboo, and I’m no nuclear strategist, so I’m uncertain about the strength nuclear contagion.

The “contagiousness” of nuclear conflict helps determine the value of strict global governance: if conflicts are extremely contagious (such that something like the real-world Syrian Civil War ends up with the superpowers at DEFCON 1), then small-scale wars are still extremely dangerous, and global policing is very desirable. If nuclear conflict isn’t contagious at all and it’s easy to stay out of a dispute, then it would be a lot more acceptable for the leading nations to just let nuclear wars happen, in the same way that the modern world often lets civil wars happen without intervening too much. Just play defense by being really paranoid about your ports/borders, and threatening to first-strike anyone who develops suspicious new long-range capabilities.

I really don’t know much about the question of contagiousness. Is there something special about fission weapons and the “nuclear taboo” that affects contagiousness? (Maybe nations feel like they have to “use or lose” their ICBMs before they are destroyed by opponents.) Or does all war seem contagious because it naturally erupts at the center of complex knots of geopolitical tensions and alliances? (Like the rapid domino-like declarations of war that set off WW1, or the agglomeration of seemingly disparate atrocities and conflicts centered around WW2.)

• If nuclear attacks are specifically and specially contagious, we should be most worried about something like a nuclear Israel-Iran or India-Pakistan conflict. Aside from the horrific direct cost of an India-Pakistan exchange, how likely would it be to eventually draw in the gigantic arsenals of the USA and Russia?

• If it’s more about the underlying geopolitical conditions and the universal logic of escalation, we should be most worried about small direct conflicts between the biggest nuclear powers getting rapidly out of hand. Maybe the USA feels pressured to confront China early over Taiwan since China’s power is only rising over time, and the vision of having a limited-casualties, mostly-naval battle in the South China Sea ends up being wrong, with geopolitical energy fueling rapid massive escalation between the two.

---

What does this tell us about the real world?

To keep things simple, I’ve been treating this question as purely an alternate history exercise – what would have happened by 2020 if things were different in 1950. Maybe from a long-term perspective of thousands of years, that much proliferation means you’re totally doomed. But on a timescale like that, it’s still early days for the real world’s nuclear proliferation dynamics, too.

But that brings up a key question: are we ourselves totally doomed? The spread of knowledge and the advance of technology makes real nuclear weapons accessible to more and more nations despite the difficulty. It’s possible that proliferation in the semi-easy-nukes scenario vs the real world will end up in essentially the same place, with one history just being a fast-forward version of the other.

As far as I can tell , the real world is indeed just doing universal proliferation in slow-motion. But, the slowness matters: it provides time enough to work out strategic implications, develop defensive technology, establish norms of behavior, and create international agreements, all of which help the cause of peace. In a best-case scenario, perhaps proliferation advances so slowly that it is essentially an irrelevant force compared to the many other changes the future holds.

Differential technology development is another key difference – in our world, submarines are easier for nations to build than nukes, so more nations in our world will be able to wield the stabilizing second-strike capability from the moment they get the bomb.

A final contrast the semi-easy-nukes scenario: in the real world, it seems impossible that a lone-wolf madman or small group could ever build a bomb entirely on their own without attracting attention, almost regardless of how advanced technology gets. Nuclear terrorism might be rife in the semi-easy-nukes scenario, but it seems easy to control in the real world, and limited to the fear that someone could steal and activate a fully-constructed bomb. If nation-states are the only nuclear threat, this has many simplifying effects on strategy and diplomacy: the list of possible nuclear-attack motivations is shorter, anonymous attacks become less attractive, second-strikes are an effective form of deterrence, etc.

If we are headed for a world of universal proliferation (perhaps with both nuclear and other futuristic weapons), the imperial/neocolonial governance scenario might also have lessons for us. Truly unified global governance has never existed historically, and seems difficult to achieve with no clear path for getting from here to there (especially if we want to imagine imposing the new world order without using devastating force). Furthermore, as many including Nick Bostrom and Bryan Caplan have noted, global government seems dangerous and undesirable even if it’s stable. It would be a single point of failure for human civilization, posing threats like permanent cultural/economic stagnation, worldwide totalitarian lock-in, and other dystopian scenarios.

By contrast, imperialism has been the norm throughout history, with the past 100 years being a dramatic exception to the general rule. Historical imperialism and colonialism certainly deserve their dark reputation for exploitation and human rights abuses, but there are also more positive examples in the modern day – the European Union is a good example of nations giving up some sovereignty to a supranational organization in a controlled way in exchange for the benefits of solidarity and standardization. The NATO alliance offers a blueprint for a relatively light-touch form of global governance that restricts itself to military affairs only. In reality, a world of increased global coordination would probably look like a mixture of these types of voluntary supranational associations and the territory grabs / “spheres of influence” of the bad old days. (Personally, I thought that Vernor Vinge’s sci-fi novel “Rainbow’s End” had a good depiction of this scenario – kind of China + US + EU in cooperation versus the assorted forces of chaos and unexpected technological disruption.) If it’s true that our future is headed in a neo-imperial direction of increasingly coordinated governance, this suggests that it might be beneficial to help groups of smaller countries build EU-like alliances. If countries find themselves needing to join a team of some kind to stay safe and succeed, much better that they voluntarily join a multinational organization than fall into the bullying sphere of influence of a distant superpower.

What works for nukes might not work for A.I.

Bostrom’s original thought experiment forces us to realize that, given a sufficiently dangerous technology, human civilization might have no choice but to implement extreme draconian measures (unified global government, extreme surveillance with harsh penalties, and attempts at mind control) to avoid extinction. Over the course of this post, considering a toned-down version of his thought experiment has caused me to come off as almost wistful about the concentration of power during what were otherwise very dark times for humanity – colonialism and the Cold War. But the threat posed by AI is very different.

With nukes, you can create a peaceful world if you just suppress things well enough. That’s something that even evil, incompetent empires are pretty good at. With AI, a suppression strategy might buy time, but it seems doomed to fail eventually – instead, you need to actually engage with the technical challenge of building a solution in the form of a safe, aligned AI. Personally, I worry that the challenge is so great, so subtle and so easy to get wrong accidentally, that no existing government has the wisdom and capacity to navigate the issue with a good probability of success. As Gwern said, “If the worst-case AI x-risk happened, it would be hard for every reason that corona was easy… Don’t count on our institutions to save anyone: they can’t even save themselves.”

In short, for difficult but simple threats that our current civilization has the skills to handle, centralization and increased global governance is good, but for complex problems beyond our current civilizational ability to solve, I think we’re dead by default, so our best option may be to decentralize, increase governance experimentation (via things like charter cities), increase the variance of outcomes, and hope that either the fruits of governance experimentation percolate everywhere and helps society become wiser, or that the increased variance alone improves our odds of a good outcome.

Naturally, this inspires questions about the details, rather than just where the world should fall on an overall coordination spectrum – how can we acquire the benefits of experimentation, competition, and improvement in some dimensions –like creating more pressure and energy for reforms that increase institutional decision-making capability– while retaining the benefits of cooperation in other areas, like avoiding technology arms-races or being able to implement bans on dangerous research?

---

I hope you found this post useful. I’d be happy to read your comments or corrections by email at jacksonw.xyz@gmail.com. You can participate in the discussion of the cross-posted article on the EA forum, or find additional comments about related issues over at the original EA Forum question by eca. If on the other hand you are coming to this question less from an effective-altruist perspective and more out of a dark fascination with the Manhattan Project, Trinity test, etc, you might enjoy my commentary on the game Braid.