Having looked at warhead stocks, nuclear winter research, etc, I think nuclear war isn’t an x-risk either.
I’m also rather doubtful that climate change significantly increases the probability of nuclear war. Regional conflicts and insurgencies in certain places, sure. But the pathway from there to nuclear war is very unclear. You can point to the Indo-Pakistani dyad as a possible flashpoint, but both of them have few nuclear weapons. And their historical conventional conflicts did not escalate to involve other countries.
But it is outcomes that are morally close to extinction, the loss of most of humanity’s capacity and potential. Nuclear winter of a few degrees would not impact agriculture so adversely to cause this to happen. At this point you are multiplying so many small probabilities in series that you cannot call climate change a real x-risk without doing the same for so many other things that are equally likely to set a chain of bad events in motion.
I think it is useful to discuss qualifies as an X-risk. Asteroid/comet impact is widely regarded as an X-risk, but a big one that could cause human extinction might only have a one in a million probability in the next 100 years. This is a 0.0001% reduction in humanity’s long term value. However, if you believe 80,000 Hours that nuclear war might have a ~3% chance in the next 100 years and this could reduce the long term future potential of humanity ~30%, that is a ~1% reduction in the future of humanity this century. So practically speaking, it is much more of an X-risk than asteroids are. Similarly, if you believe 80k that extreme climate change has a ~3% chance in the next 100 years and it reduces the long run potential by ~20%, that is a 0.6% reduction in the long term future of humanity. This again is much larger than asteroids. I personally think the nuclear risk is higher and the climate risk is lower than these numbers. It is true that some of the long-term impact could be classified as trajectory changes rather than traditional X risk. But I think most people are interested in trajectory changes as well.
I don’t think the 80k estimate on climate change is based on a thorough investigation of the science. I just don’t see how from the impacts estimated in the next 100 years, extreme climate change could be thought to be a greater than 0.1% ex risk. The heat stress of >4 degrees would be bad but if things started going that badly, I think the world would take action. In a few decades it will be much cheaper to abate GHGs and everyone will have an interest in doing so
I generally agree. The question is whether we should call something an X-risk by the impact if it happens alone or by the impact*probability. If the latter, and if comets are an X-risk, then we should call extreme climate change (and definitely nuclear war) an X-risk.
By that logic you are turning the idea of an x-risk into anything that really matters in the long run. So poverty is an x-risk too in this definition. That makes it not a useful definition and is also very different from how most people think about the term.
Extinction (or something just as bad): x-risk. I go by that.
“Normal” nuclear war could be only only a first stage of multistage collapse. However, there are some ideas, how to use exiting nuclear stockpiles to cause more damage and trigger a larger global catastrophe—one is most discussed is nuking a supervolcano, but there are others. In Russian sources is a common place that retaliation attack on US may include attack on the Yellowstone, but I don’t know if it is a part of the official doctrine.
Future nuclear war could be using even more destructive weapons (which may exist secretly now). Teller has been working on 10 gigaton bomb. Russians now making Poseidon large torpedo system which will be probably equipped with 100 Mt cobalt bombs.
However, there are some ideas, how to use exiting nuclear stockpiles to cause more damage and trigger a larger global catastrophe—one is most discussed is nuking a supervolcano,
Absurd. Why would anyone do that?
retaliation attack on US may include attack on the Yellowstone, but I don’t know if it is a part of the official doctrine.
Future nuclear war could be using even more destructive weapons
Even the most destructive historical weapons (e.g. Tsar Bomba) have not been deployed. Warheads have gotten smaller over recent decades. No reason for this trend to reverse.
I’ve read that US has an instrument to attack hardened underground facilities by multiple heavy nuclear strike in one place, which allows creating much deeper crate than a single nuclear explosion and destroy targets around 1 km deep. The same way an volcanic caldera cover could be attacked, and such multiple strikes could weaken its strength until it blow up by internal pressure—so you don’t need to go through the whole caldera’s cover. no new weapons for it is needed—just special targeting of already exiting.
Having looked at warhead stocks, nuclear winter research, etc, I think nuclear war isn’t an x-risk either.
I’m also rather doubtful that climate change significantly increases the probability of nuclear war. Regional conflicts and insurgencies in certain places, sure. But the pathway from there to nuclear war is very unclear. You can point to the Indo-Pakistani dyad as a possible flashpoint, but both of them have few nuclear weapons. And their historical conventional conflicts did not escalate to involve other countries.
But remember, X-risk is not just extinction—there are many routes to long term future impacts from nuclear war—some are mentioned here.
But it is outcomes that are morally close to extinction, the loss of most of humanity’s capacity and potential. Nuclear winter of a few degrees would not impact agriculture so adversely to cause this to happen. At this point you are multiplying so many small probabilities in series that you cannot call climate change a real x-risk without doing the same for so many other things that are equally likely to set a chain of bad events in motion.
I think it is useful to discuss qualifies as an X-risk. Asteroid/comet impact is widely regarded as an X-risk, but a big one that could cause human extinction might only have a one in a million probability in the next 100 years. This is a 0.0001% reduction in humanity’s long term value. However, if you believe 80,000 Hours that nuclear war might have a ~3% chance in the next 100 years and this could reduce the long term future potential of humanity ~30%, that is a ~1% reduction in the future of humanity this century. So practically speaking, it is much more of an X-risk than asteroids are. Similarly, if you believe 80k that extreme climate change has a ~3% chance in the next 100 years and it reduces the long run potential by ~20%, that is a 0.6% reduction in the long term future of humanity. This again is much larger than asteroids. I personally think the nuclear risk is higher and the climate risk is lower than these numbers. It is true that some of the long-term impact could be classified as trajectory changes rather than traditional X risk. But I think most people are interested in trajectory changes as well.
I don’t think the 80k estimate on climate change is based on a thorough investigation of the science. I just don’t see how from the impacts estimated in the next 100 years, extreme climate change could be thought to be a greater than 0.1% ex risk. The heat stress of >4 degrees would be bad but if things started going that badly, I think the world would take action. In a few decades it will be much cheaper to abate GHGs and everyone will have an interest in doing so
I generally agree. The question is whether we should call something an X-risk by the impact if it happens alone or by the impact*probability. If the latter, and if comets are an X-risk, then we should call extreme climate change (and definitely nuclear war) an X-risk.
I see, yes good point.
By that logic you are turning the idea of an x-risk into anything that really matters in the long run. So poverty is an x-risk too in this definition. That makes it not a useful definition and is also very different from how most people think about the term.
Extinction (or something just as bad): x-risk. I go by that.
“Normal” nuclear war could be only only a first stage of multistage collapse. However, there are some ideas, how to use exiting nuclear stockpiles to cause more damage and trigger a larger global catastrophe—one is most discussed is nuking a supervolcano, but there are others. In Russian sources is a common place that retaliation attack on US may include attack on the Yellowstone, but I don’t know if it is a part of the official doctrine.
Future nuclear war could be using even more destructive weapons (which may exist secretly now). Teller has been working on 10 gigaton bomb. Russians now making Poseidon large torpedo system which will be probably equipped with 100 Mt cobalt bombs.
Absurd. Why would anyone do that?
I’m sure it isn’t. Also, scientifically speaking it doesn’t even seem possible to ignite a supervolcano with nukes: https://www.iflscience.com/environment/what-would-happen-if-a-nuclear-bomb-was-dropped-on-yellowstone-supervolcano/
Even the most destructive historical weapons (e.g. Tsar Bomba) have not been deployed. Warheads have gotten smaller over recent decades. No reason for this trend to reverse.
Theoretical reasons for Doomsday weapon was laid by Herman Khan in “On Thermonuclear war”. I scanned related chapter here: https://www.scribd.com/document/16563514/Herman-Khan-On-Doomsday-machine
The main idea is that it is ideal defence weapon, as no body will ever attack a country owning such a device.
The idea of attacking the Yellowstone is discussed very often in Russian blogosphere (like here https://izborskiy-club.livejournal.com/310579.html), and interest to the geophysical weapons was strong in the Soviet Union (details here: http://nvo.ng.ru/armament/2006-04-21/6_weapontheyfear.html) - this interest ended up in creating Poseidon artificial tsunami system which is now under testing.
I’ve read that US has an instrument to attack hardened underground facilities by multiple heavy nuclear strike in one place, which allows creating much deeper crate than a single nuclear explosion and destroy targets around 1 km deep. The same way an volcanic caldera cover could be attacked, and such multiple strikes could weaken its strength until it blow up by internal pressure—so you don’t need to go through the whole caldera’s cover. no new weapons for it is needed—just special targeting of already exiting.
Russian Poseidon system has 100-200 Mt bombs delivered by a very large torpedo and is in final stages of construction. https://en.wikipedia.org/wiki/Status-6_Oceanic_Multipurpose_System