If you prefer to listen rather than read, this blog is available as a podcast here. Or if you want to listen to just this post:

Or download the MP3

Last week I talked about Steven Pinker’s book Enlightenment Now: The Case for Reason, Science, Humanism, and Progress. As I mentioned at the time I probably wasn’t doing it justice because that’s just the reality when you try to cover a 500 page book using a 3500 word post. In contrast, this week, rather than trying to cover what amounts to Pinker’s grand unified theory of civilization, I’d like to focus on just one point he brought up.

As I mentioned in the last post, I enjoy Pinker’s books, and find myself largely in agreement with him on most things (long tail risk being my big disagreement) and one of the things I was glad to see him make a strong case for was nuclear power. As I pointed out, Pinker feels that there are really only two existential crises we need to worry about: climate change and nuclear war.  As part of my disagreement with him about long tail risks, I would argue that there are a lot more than just these two, but they are interesting in that they are both issues which are entirely within humanities power to solve, it would just require a lot of coordination. Or rather, it would require a lot of positive coordination. Natural incentives have lead to significant, albeit selfish, coordination on the burning of fossil fuels and to a lesser extent on the building of nukes. But this is equivalent to both prisoners defecting, we need them to both cooperate. But as you might imagine, and as I assume even Pinker would admit, such coordination is hard, and takes a long time. Climate change is interesting because there is a step most countries could take, which requires no cooperation, no signing of any accords, no locking other countries, like China, into specific actions, and which wouldn’t slow down growth or significantly change the way we use energy. Of course I’m talking about nuclear power, and one of the highlights of “Enlightenment Now” was Pinker’s defense of it.

As I said though, Pinker and I share a very different level of concern for long tail risks. Given that he worries less, his advocacy of nuclear power seems obvious. But what about me? How do I get off championing nuclear power? Certainly, it would be very ironic if using nuclear power to combat the potential existential crisis of climate change, lead to the very real existential crisis of full on nuclear war. Accordingly, the first thing we need to discuss is whether nuclear power increases the threat of nuclear weapon proliferation. The answer to that question is that it’s hard to say, this is particularly the case when you’re trying to compare these two separate worlds in their entirety. One where we continue much as we have, with haphazard development of nuclear power, inconsistent policies, and no real plan for the weapons grade plutonium and uranium we already do have (except for using it in weapons.) And another world where we focus on more advanced reactors with much less risk of proliferation (particularly once you start talking about thorium and depleted uranium reactors) and with a pathway to use weapons grade material in places other than weapons. All of which is to say that the risks, are somewhat unknown. But if you really believe that climate change represents an existential crisis (I don’t, but I do think it’s worth taking precautions against) then some level of risk may be unavoidable, and that’s one of the themes I’ll be returning too.

It’s been awhile since I covered asymmetric risks, particularly in terms of fragility vs. antifragility. So as a reminder, one of the key points of this dichotomy is that both costs and benefits can have one of two sets of attributes, they can either be known and limited or they can be unknown and unlimited. If the benefits are known and limited, while the costs are the reverse, then that thing is fragile. If it’s the opposite, unlimited benefits with limited costs, then that thing is antifragile. In theory it’s straightforward, in practice it can be more complicated, particularly with situations where there are lots of competing costs and benefits, with varying levels of severity and probability. As, for example, in the situation I just described, deciding whether to use nuclear power to reduce the risk of climate change.

This is the case even if we assume that nuclear power will entirely prevent climate change. You could argue (and I have) that the cost of climate change, even under the worst case scenario is not unlimited. (I have even argued that the cost of nuclear war is not unlimited, ie that it won’t wipe out humanity permanently and forever.) That said, it would be pretty bad, bad enough that most people would place it the unlimited cost category. Additionally, while I think most people would place climate change in the “very high probability” category, there is still a lot we don’t know.

What about the other side, proliferation brought on by the increased use of nuclear power? On this side, there are two ways proliferation could happen. The first and most likely is that a non-state actor could get their hands on some weapons grade uranium or plutonium. In that case they would still have to use it to make a bomb, which is not impossible, but it’s definitely not trivial either. How many bombs could they make and successfully detonate? I think most people would say one, if they’re lucky, but I guess if we’re looking at the worst case scenario it’s conceivable that they could build perhaps half a dozen. In any case, it’s hard to imagine the number would be more than a handful. And at best they’re equivalent to the bombs dropped on Hiroshima and Nagasaki (crude design, fission not fusion.)

Okay, now let’s say they actually detonate them. Is this bad? Of course! It makes 9/11 into a historical footnote, but, does it end civilization? Not unless we enormously over-react (another comparison to 9/11). And, is the worst case nuclear terrorism as bad the worst case climate change? Probably not, and I think, probability wise, climate change is vastly more likely than nuclear terrorism to boot. Recall that we have had nuclear weapons for quite a while, they passed through the fall of the Soviet Union, and were kept secure for the last 20 years through the height of Islamic-extremism in Pakistan, a majority Muslim country. To return to Pinker, he apparently considers the possibility to be so remote that he doesn’t even bother to mention nuclear weapons in his chapter on terrorism. (At least as far as I recall, and the index backs me up.) As I have already pointed out I am more pessimistic than he is, but I still see multiple hurdles standing between a non-state actor and the acquisition of a nuclear weapon.

The other way proliferation could happen is that a state could use their nuclear power to create nuclear weapons, and they could use it to create far more than one, or six. Obviously this is a concern, and I’m already on record as saying that I don’t think the number of states with nukes will ever go below the number it’s currently at, and further, I think we might see states that previously relied on the US nuclear umbrella getting nukes of their own as Pax Americana fades. All this may happen (contra Pinker) but the question remains will the increased use of nuclear power have anything to do with it? The connection seems tenuous. On the one hand we have North Korea which has nuclear weapons, but no working nuclear reactor. On the other hand we have Japan which has nuclear power, but no nuclear weapons. And if, someday, they decided they needed nuclear weapons, I doubt, at that moment, that the lack of nuclear power would slow them down very much, even if that were the case. At a minimum this seems to suggest that the technology and infrastructure needed to build nuclear weapons is less than that required to build nuclear reactors. And if reactor technology were to be imported I imagine it would be the kind of proliferation resistant reactors I already talked about.

Once we eliminate worries about proliferation, then the case for nuclear power (particularly for those most worried about climate change) becomes pretty solid. But people have concerns about nuclear power beyond just proliferation and as long as we’re on the subject I might as well address them.

The next highest item on the list for most people (and the highest for some) is nuclear waste. And I admit that with my emphasis on low time preference, creating something that’s dangerous for thousands of years is viscerally unappealing to me. But of course it’s important to look behind that initial visceral reaction (something far, far to few do, particularly with this issue) and consider the actual data.

First, it’s not as if radioactive waste is unique in its longevity, we’re creating lots of concentrated  heavy metal waste (eg cadmium, mercury, and maybe you’ve heard of the problems with lead?) which is dangerous essentially forever. Will the Yucca Mountain Repository (should it ever be built, which seems doubtful) be as dangerous in 1000 years as Norilsk, Russia, or even Flint, Michigan, when you consider how inaccessible the waste will be?

Second, and related to the first point, the amount of high level waste created is tiny. Enough so that you don’t need some giant site with armed guards and lasers. In fact one suggestion has been to take the small amount of waste and disperse it rather than concentrating it. For example, mix it with dirt and rocks and refill old uranium mines with it, diluting it enough that the background radiation is at the same level as it was before you mined the uranium.

Finally, one thing that most people don’t consider is that if we’re going to create pollution or waste anyway, it’s best to diversify. I first came across this point in one of Nassim Nicholas Taleb’s books (I don’t recall which one). He pointed out that it is far better to create small amounts of many pollutants than a large amount of only one. You may be wondering why this is. Well as it turns out the harm from toxicity is not linear. To use an example many people would be familiar with, alcohol.  (Not me actually, I had to look it up, I’ve never had a drink.)  A blood alcohol level of 0.1 is drunk, but not ridiculously so, twice that and you’re probably vomiting. Three times that and you’re probably unconscious. Four times that and you’re in a coma and possibly dead. As you can see the harm jumps a huge amount at each step, particularly between 3x and 4x where the harm jumps to infinity (if you’re the person who ended up dying.)

Now of course we don’t exactly know where the big jump is in carbon emissions, if a CO2 level of 500 ppm is 10x as bad as one of 400 ppm (where we are now) and if 600 ppm is 1000x as bad. And that if it ever hit 700 ppm we’d turn into Venus, and all life would be wiped out. But regardless it’s almost certainly non-linear, and since nuclear power (and by extension nuclear waste) would be removing CO2 “off the top” so to speak, you’re looking at trading a pollutant on the low end of the harm curve for the very top of the curve of another pollutant. All of which is another way to say the same thing which has been said since ancient times, that the dose makes the poison. Yes, nuclear waste is bad, but by removing the top end of our CO2 emissions it may be replacing something much, much worse.

Having covered proliferation and nuclear waste, we at last reach the concern potentially shared by the most people, but which has the least basis in reality: a nuclear reactor disaster, a Three Mile Island, a Fukushima or a Chernobyl. Here Pinker and I are entirely on the same page, so I’ll turn things over to him:

[Nuclear power] has a lower carbon footprint than solar, hydro, and biomass, and it’s safer than them, too. The sixty years with nuclear power have seen thirty-one deaths in the 1986 Chernobyl disaster, the result of extraordinary Soviet-era bungling, together with a few thousand early deaths from cancer above the 100,000 natural cancer deaths in the exposed population. The other two famous accidents, at Three Mile Island in 1979 and Fukushima in 2011, killed no one. Yet vast numbers of people are killed day in, day out by the pollution from burning combustibles and by accidents in mining and transporting them, none of which make headlines. Compared with nuclear power, natural gas kills 38 times as many people per kilowatt-hour of electricity generated, biomass 63 times as many, petroleum 243 times as many, and coal 387 times as many–perhaps a million deaths a year. (emphasis on 387 in the original)

Perhaps you’re amazed by these figures or perhaps not, they’ve been available for a long time, and possibly, given the demographics of my readership, you are all entirely unsurprised, but when I look out at the broader world, I can detect nothing which would indicate a recognition of how safe nuclear power actually is. Certainly not among the general population, nor among the politicians, and despite the repeated pleas of desperate need, not even very much among self-professed environmentalists.

Now to be fair even if Fukushima didn’t kill anyone it has proved to be tremendously expensive to clean up, with the latest estimate putting the figure at 21.5 trillion yen, or $188 billion, and I suppose an argument could be made against nuclear power solely from the perspective of cost, and indeed a quick search reveals that this is a very common concern. The question is why is it so expensive to build new plants and to clean up Fukushima. Is it because if we spent $100 billion rather than $188 billion Fukushima would go from a situation where no one died, to a situation where 10,000 people died? If we spent less money on regulating and building nuclear reactors would we go from having two accidents where no one died and one where a few thousand died, to having dozens of accidents and tens of thousands of casualties? (Even if that were the case I think nuclear power would still be ahead of petroleum and coal, and probably even biomass.)

I don’t actually think either of these things would happen, rather I think the word “nuclear” and “radiation” are in the same category as “racist”, words were the reaction engendered is sometimes dramatically out of proportion to the actual harm, and where, if the word sticks, it can make a topic completely off limits. Accordingly, Fukushima isn’t just a disaster, it isn’t another industrial cleanup, it’s a nuclear disaster, and a radioactive cleanup, and no expense should be spared in containing the dangerous waste. But maybe that’s a good thing, no telling what would happen if we just spent a couple of billion slapping some concrete and steel over everything and walked away, with some signs telling people to avoid the area.

Oh wait… that’s basically exactly what happened with Chernobyl which was objectively a far worse accident. Maybe by looking at what happened in the aftermath of that we can get some sense of what might happen if Japan decides to spend less than $188 billion to clean up Fukushima. It doesn’t take much searching to find articles talking in excited terms about the amount of wildlife found in the Chernobyl Exclusion Zone (CEZ). One article declares that it’s a nature reserve. Another mentions that within the CEZ wildlife is flourishing. This was unexpected, in one article from National Geographic I came across, they quote a biologist who “studies chernobyl” (one wonders if his studies have included a visit) as predicting that when the author of the article goes to Chernobyl that he won’t “see any roadkill in the exclusion zone—and would be lucky to hear any birds or see any animals.” Instead the author reports:

Walking along sandy firebreaks used as forest highways…we found the tracks of wolf, moose, deer, badger, and horses. I counted scores of birds: ravens, songbirds, three kinds of birds of prey, and dozens of swans paddling in the radioactive cooling pond.

The article goes on to report that in a study of 14 species of mammals one scientist found no evidence that any of those populations were “suppressed” within the CEZ.

I am sure that there are some health impacts on this wildlife and positive that the CEZ is not without its negative effects. I’m sure that if people were allowed to live there, that there would be higher rates of cancer, among other things. But, also recall, that this is the worst of the disasters, combined with the least cost and effort at cleanup.

In addition to offering a defense of nuclear power, I wanted use this post to examine why it has faced so much resistance. Why haven’t we embraced it, given everything I’ve already mentioned? I would argue that nuclear power is among the best examples of one of modernity’s more distressing trends: veto through intolerance. Or as Taleb describes it, the most intolerant wins. He describes how:

It suffices for an intransigent minority –a certain type of intransigent minority –to reach a minutely small level, say three or four percent of the total population, for the entire population to have to submit to their preferences. Further, an optical illusion comes with the dominance of the minority: a naive observer would be under the impression that the choices and preferences are those of the majority.

Let’s take Yucca Mountain for example. Nearly everyone who’s aware of the problem of high level nuclear waste is in favor of it, and those who aren’t in favor of it can be divided up into two camps, those who don’t care either way, and Nevadans. And the opposition of the Nevadans, who represent less than 1% of the US population was enough to kill it.

I assume much the same thing is happening when we look at the reaction of people to even the very mention of the words, nuclear and radioactive. There is obviously a small minority for whom these words are as garlic to vampires, if not worse. And then there’s the vast majority who really don’t care, but, as Taleb points out, appear, to the naive observer to be concerned as well, and then there’s a few people who actually understand the risk-reward tradeoff (like Pinker, probably) who are in favor of it. To put it in more concrete terms, how much of the $188 billion is being spent to calm the 1% of the population who is the least tolerant? I guarantee it’s far out of proportion to their numbers.

I’m guessing at this point that you can already see where this applies to various social justice causes. And what’s important to point out here, is not the old “Social Justice Warriors be crazy, yo.” But the fact that both the concerns about nuclear power and the concerns about social justice are based on real risks. In the case of nuclear power there is a real risk that expanding our use of it will contribute to proliferation, or that there will be a disaster involving high level waste, or that a nuclear reactor will suffer a meltdown. But for all the reasons I pointed to above, I think these risks are low and worth taking.

The intolerance towards certain kinds of speech, particularly speech related to social justice issues, comes from fear of risk as well. There is a risk it could lead to a recriminalization of homosexuality, or to violence and the resumption of things like lynchings, or we could even end up in a world that was indistinguishable from the Handmaid’s Tale. But all of these risks are tiny. The question is are there any compensating rewards if we continue to support free speech and free expression (even for the alt-right and neo-nazis)? I believe there are, but much like the benefits of nuclear energy, they’re diffuse, and will take a long time to fully be realized.

If we are at a point where the smallest of minorities can veto even the tiniest of risks, what does that mean going forward? We have already seen how this tactic has completely tabled one of the most effective precautions we could take against global warming. What other future risks are we going to ignore in order to address the short-term concerns of various minorities? I know that people like Pinker have basically argued that the future is just going to get increasingly less risky, but this is one of the points I disagree with him on. I think going forward, risks are going to be less frequent (lower volatility) but of greater impact when they do occur (higher fragility). Thus not only are the risks greater, but we will have much less experience in dealing with them when they do occur, add to all of that the ability of frightened minorities to derail the plans of the far-sighted, and there’s ample cause to once again claim, “We are not saved.”

I’m trying to think of some way that I could win through intolerance, but even though the number of people necessary to do that is getting smaller and smaller, I don’t think it’s reached the point where one, slightly weird blogger with a fixation on clever appeals for donations has any leverage, yet.