A common heuristic argument I’ve seen recently in the effective altruism community is the idea that existential risks are low probability because of what you could call the “People really don’t want to die” (PRDWTD) hypothesis. For example, see here,
People in general really want to avoid dying, so there’s a huge incentive (a willingness-to-pay measured in the trillions of dollars for the USA alone) to ensure that AI doesn’t kill everyone.
(Note that I hardly mean to strawman MacAskill here. I’m not arguing against him per se)
According to the PRDWTD hypothesis, existential risks shouldn’t be anything like war because in war you only kill your enemies, not yourself. Existential risks are rare events that should only happen if all parties made a mistake despite really really not wanting to. However, as plainly stated, it’s not clear to me whether this hypothesis really stands up to the evidence.
Strictly speaking, the thesis is obviously false. For example, how does the theory explain the facts that
When you tell most people about life extension, even probably billionaires who could do something about it, they don’t really care and come up with excuses about why life extension wouldn’t be that good anyway. Same with cryonics, and note I’m not just talking about people who think that cryonics is low probability: there are many people who think that it’s a significant probability but still don’t care.
The base rate of a leader dying is higher if they enter a war, yet historically leaders have been quite willing to join many conflicts. By this theory, Benito Mussolini, Hideki Tojo and Hitler apparently really really wanted to live, but entered a global conflict anyway that could have very reasonably (and in fact did) end in all of their deaths. I don’t think this is a one-off thing either.
I have met very few people who have researched micromorts before and purposely used them to reduce the risk of their own deaths from activities. When you ask people to estimate the risks of certain activities, they will often be orders of magnitudes off, indicating that they don’t really care that much about accurately estimating these facts.
As I said two days ago, few people seemed concerned by the coronavirus. Now I get it: there’s not much you can do to personally reduce your own death, and so actually stressing about it is pointless. But there also wasn’t much you could do to reduce your death after 9/11 and that didn’t stop people from freaking out. Therefore, if the theory you appeal to is that people don’t care about things they have no control over then your theory is false.
Obesity is a common concern in America, with 39.8% of adults here being obese, despite the fact that obesity is probably the number one contributor to death besides aging, and it’s much more controllable. I understand that it’s really hard for people to lose weight, and I don’t mean to diminish people’s struggles. There are solid reasons why it’s hard to avoid being obese for many people, but the same could also be true of existential risks.
I understand that you can clarify the hypothesis by talking about “artificially induced deaths” or some other reference class of events that fits the evidence I have above better. My point is just that you shouldn’t state “people really don’t want to die” without that big clarification, because otherwise I think it’s just false.
Yeah similar to obesity, people seem quite willing to cave into their desires. I’d be interesting in knowing what the long-term effects of daily alcohol consumption are, though, because some sources have told me that it isn’t that bad for longevity. [ETA: The Wikipedia page is either very biased, or strongly rejects my prior sources!]
A common heuristic argument I’ve seen recently in the effective altruism community is the idea that existential risks are low probability because of what you could call the “People really don’t want to die” (PRDWTD) hypothesis. For example, see here,
(Note that I hardly mean to strawman MacAskill here. I’m not arguing against him per se)
According to the PRDWTD hypothesis, existential risks shouldn’t be anything like war because in war you only kill your enemies, not yourself. Existential risks are rare events that should only happen if all parties made a mistake despite really really not wanting to. However, as plainly stated, it’s not clear to me whether this hypothesis really stands up to the evidence.
Strictly speaking, the thesis is obviously false. For example, how does the theory explain the facts that
When you tell most people about life extension, even probably billionaires who could do something about it, they don’t really care and come up with excuses about why life extension wouldn’t be that good anyway. Same with cryonics, and note I’m not just talking about people who think that cryonics is low probability: there are many people who think that it’s a significant probability but still don’t care.
The base rate of a leader dying is higher if they enter a war, yet historically leaders have been quite willing to join many conflicts. By this theory, Benito Mussolini, Hideki Tojo and Hitler apparently really really wanted to live, but entered a global conflict anyway that could have very reasonably (and in fact did) end in all of their deaths. I don’t think this is a one-off thing either.
I have met very few people who have researched micromorts before and purposely used them to reduce the risk of their own deaths from activities. When you ask people to estimate the risks of certain activities, they will often be orders of magnitudes off, indicating that they don’t really care that much about accurately estimating these facts.
As I said two days ago, few people seemed concerned by the coronavirus. Now I get it: there’s not much you can do to personally reduce your own death, and so actually stressing about it is pointless. But there also wasn’t much you could do to reduce your death after 9/11 and that didn’t stop people from freaking out. Therefore, if the theory you appeal to is that people don’t care about things they have no control over then your theory is false.
Obesity is a common concern in America, with 39.8% of adults here being obese, despite the fact that obesity is probably the number one contributor to death besides aging, and it’s much more controllable. I understand that it’s really hard for people to lose weight, and I don’t mean to diminish people’s struggles. There are solid reasons why it’s hard to avoid being obese for many people, but the same could also be true of existential risks.
I understand that you can clarify the hypothesis by talking about “artificially induced deaths” or some other reference class of events that fits the evidence I have above better. My point is just that you shouldn’t state “people really don’t want to die” without that big clarification, because otherwise I think it’s just false.
People clearly DO want to die - $2.2 billion dollars of actual spending (not theoretical “willingness to pay”) on alcohol in the US in 2018.
Yeah similar to obesity, people seem quite willing to cave into their desires. I’d be interesting in knowing what the long-term effects of daily alcohol consumption are, though, because some sources have told me that it isn’t that bad for longevity. [ETA: The Wikipedia page is either very biased, or strongly rejects my prior sources!]