| It should be easier to focus directly on building international institutions.
Can you clarify this? As a new EA, I am unfamiliar with what this might be in reference to. Are you arguing for focusing on the creation of a second nation-state promoting liberal values and technological advancement (to backstop US leadership)? Or just circumventing politics and going straight at the underlying risks (AI, nuclear proliferation)? It seems without the air-cover of a liberal superpower, pursuing such ventures will be significantly harder.
Well for Trump increasing x risks, the potential downside is that there is going to be less international cooperation and less prospect for coordination between nations on issues like AI. That’s what I thought you had in mind and it implies that we should try to build more credible and effective groups like the UN and IAEA. It’s easier to work on this directly rather than doing things that make people less likely to support populists who are against international involvement. I guess there are other ways Ttump could be bad for x risks, so that might change things.
Gotcha. I was going the other way; considering what automation and globalization have done to low-skilled workers over the last 30 years (which is only going to get a lot worse), some sort of significant restructuring of our economic system (both wealth distribution and cultural self-actualization via employment) seems necessary to prevent full-scale upheaval (e.g. revolution, civil war, etc). Trump seems to be the first major indication that this problem is getting so bad that it could literally end America or at least American leadership and the promotion of liberal values.
The case is for defending the conditions under which it’s even possible to have a group of privileged people sitting around worrying about while the world is burning.
I do agree this is a very indirect effect but I also failed to see how more direct action would be more effective.
Well that is different but I don’t think it’s a likely threat.
considering what automation and globalization have done to low-skilled workers over the last 30 years
Low-skilled workers have seen general increases in standard of living over the last 30 years and no great increase in unemployment. Globalization generally increases wealth for all income brackets in the long run and automation has not historically increased long-term unemployment.
(which is only going to get a lot worse),
We may be able to have trucks that steer themselves on the road, but actually having unmanned vehicles take themselves from A to B reliably enough to be unmanned and without requiring a human in the cabin to do anything is years away.
In any case, the lack of truck drivers is a significant constraint on our economy alongside the fact that they can only drive 10 or so hours a day. Automation here might be unusually good.
some sort of significant restructuring of our economic system (both wealth distribution and cultural self-actualization via employment) seems necessary to prevent full-scale upheaval (e.g. revolution, civil war, etc).
It seems to me like wealth is a pretty straightforward question of welfare to resolve. As for cultural self-actualization… we’ve had retirees, part-time employees, and permanently unemployed people for a long time, but they never seem to be more troublesome or more dissatisfied than the rest of the electorate.
Trump seems to be the first major indication that this problem is getting so bad that it could literally end America or at least American leadership and the promotion of liberal values.
I don’t think so. We’ve had similar issues in the past with populist leaders and their movements, e.g. William Jennings Bryan (barely failed to get elected) and Huey Humphrey (assassinated). But our democracy has only improved over time.
Low-skilled workers have seen general increases in standard of living over the last 30 years and no great increase in unemployment. Globalization generally increases wealth for all income brackets in the long run and automation has not historically increased long-term unemployment.
Is this true in the US? My personal circle of concern certainly includes all of humanity (and beyond), but Trump’s election would appear to be an indication that displaced Americans don’t think that way.
We may be able to have trucks that steer themselves on the road, but actually having unmanned vehicles take themselves from A to B reliably enough to be unmanned and without requiring a human in the cabin to do anything is years away.
Do you have some info you can share on that? Everyone I’ve talked to in the autonomous vehicle area has indicated that adoption, not technology, will be the long pole here. And as you indicated, there are massive economic forces that will push for adoption.
And while I totally agree automation will be good for GDP, it will continue to displace a set of people who can vote.
we’ve had retirees, part-time employees, and permanently unemployed people for a long time, but they never seem to be more troublesome or more dissatisfied than the rest of the electorate.
Retirees already have cultural approval to not work: they’ve “earned” it. Part-timers and permanently unemployed are viewed as lazy or moochers by many folks in the US, where there’s a culture of self-improvement via working hard. That’s what I’m arguing will need to change.
I don’t think so. We’ve had similar issues in the past with populist leaders and their movements, e.g. William Jennings Bryan (barely failed to get elected) and Huey Humphrey (assassinated). But our democracy has only improved over time.
What metric are you using for improvement? On the axis of existential risk prevention, I would argue we’ve just taken a big step back.
Yes, average income when adjusted by PPP has increased (the standard narrative is that CPI-adjusted income has not increased, but the CPI is flawed—it overestimates inflation by focusing on cities and it ignores goods substitution), and nonwage compensation has increased (health benefits etc).
Do you have some info you can share on that?
Sure, class 8 trucks are a couple hundred thousand dollars each and the cargo is worth a lot too. They have mechanical problems. They have expensive insurance. They have complex routes. They have to reliably reach their destination and meet with the proper people to load and unload. With all this responsibility and risk it’s an easy decision to put a driver at the wheel, especially when even ordinary cars aren’t ready to go around without drivers.
Everyone I’ve talked to in the autonomous vehicle area has indicated that adoption, not technology, will be the long pole here.
Well there you have it. If it will take a long time to be adopted then there won’t be an unemployment problem for a long time.
And while I totally agree automation will be good for GDP, it will continue to displace a set of people who can vote.
They will vote for welfare, which they will get.
Retirees already have cultural approval to not work: they’ve “earned” it. Part-timers and permanently unemployed are viewed as lazy or moochers by many folks in the US, where there’s a culture of self-improvement via working hard. That’s what I’m arguing will need to change.
I don’t think any of this takes away from my point.
What metric are you using for improvement?
None of the democracy and freedom indices have been around for that long, so there is no metric. But we can look at the corruption and civil values of 100 years ago, and comparing it today, and note that it’s pretty clear that things are better all round now than they were back then. For instance, we don’t prevent women and blacks from voting anymore, and we don’t put people in prison for speaking out against the war. We don’t hold racist views about who is eligible to hold public office.
On the axis of existential risk prevention, I would argue we’ve just taken a big step back.
That’s a different axis, and it has taken many steps back, perhaps with the elections of GWB, Ronald Reagan, Eisenhower, and Truman.
Yes, average income when adjusted by PPP has increased
The distribution is what would matter for my argument. Do you have any links to data there? Where is this data from? Thanks.
Sure, class 8 trucks are a couple hundred thousand dollars each… If it will take a long time to be adopted then there won’t be an unemployment problem for a long time.
I have first hand information from those knowledgable on the topic that the technology will be ready in the next few years, and I haven’t heard a refutation of that. So I said the long pole was adoption. I didn’t say how long the pole was. You already acknowledged that the economic incentives for adoption are huge. I posit that adoption will be faster than you suspect. But we have to wait and see on this one, unless you have more information to share on why you disagree.
They will vote for welfare, which they will get.
They didn’t vote for welfare this time. Which is my point.
I don’t think any of this takes away from my point.
Maybe you can explain… you seemed to be saying that the fact we have part-timers and unemployed proves that we don’t have any cultural stigma against not working. I said that people do have such a stigma, and then you said that doesn’t take away from your point. Not sure what I missed.
That’s a different axis, and it has taken many steps back, perhaps with the elections of GWB, Ronald Reagan, Eisenhower, and Truman.
Great, that’s the axis I was trying to focus on, which is why I asked what metric you were using. And yes, it’s taken steps back before. That’s doesn’t preclude it getting worse.
I’m simply pointing out that if automation and globalization disrupt the US middle class enough, political instability could increase x-risks.
They didn’t vote for welfare this time. Which is my point.
You’re talking about what would happen in a world of high structural unemployment. But we don’t have high structural unemployment, so you can’t say that future elections will have the same dynamics. In this election cycle, Trump voters weren’t primarily motivated by the economy, they were motivated by immigration and culture issues.
Maybe you can explain… you seemed to be saying that the fact we have part-timers and unemployed proves that we don’t have any cultural stigma against not working.
I said that it shows that unemployed people aren’t likely to start a revolution or anything just because they’re unemployed.
People don’t start upheavals just because they don’t have cultural acceptance.
I’m simply pointing out that if automation and globalization disrupt the US middle class enough, political instability could increase x-risk.
If you take a broad view then half of all things will increase x risk in the long run and half of them will decrease it. I think that the issue here is too crowded and intractable and the effects are too complicated and uncertain for it to be a big goal for us to put effort on, though I agree that it would be good if the nation had a robust social system going forward.
In this election cycle, Trump voters weren’t primarily motivated by the economy, they were motivated by immigration and culture issues.
I think we need to see more data on this one. I’d put forth the idea that in fact the underlying issues were in fact economic, and that the cultural and immigration issues were mostly just utilized by Trump as things to blame for the underlying economic woes.
The terminology here can be confusing. When progressives talk about the working class, typically they mean the poor. But the poor, in the bottom 30% of American families, are very different from Americans who are literally in the middle: the middle 50% of families whose median income was $64,000 in 2008. That is the true “middle class,” and they call themselves either “middle class” or “working class.”
“The thing that really gets me is that Democrats try to offer policies (paid sick leave! minimum wage!) that would help the working class,” a friend just wrote me. A few days’ paid leave ain’t gonna support a family. Neither is minimum wage. WWC men aren’t interested in working at McDonald’s for $15 per hour instead of $9.50. What they want is what my father-in-law had: steady, stable, full-time jobs that deliver a solid middle-class life to the 75% of Americans who don’t have a college degree. Trump promises that. I doubt he’ll deliver, but at least he understands what they need.
I thought this analysis was interesting.
I said that it shows that unemployed people aren’t likely to start a revolution or anything just because they’re unemployed.
Ah, I understand your point now. Sure, that’s true, but I’m imagining what happens if the lower-middle class gets squeezed out, as mentioned above. I need to go learn more about economic trends of the middle class in American history before I can make this case, though.
If you take a broad view then half of all things will increase x risk in the long run and half of them will decrease it.
Wait, I’m new to x-risk analysis and TDT… but is that true? Why would the distribution of the space of all possible actions split 50⁄50 towards increasing or decreasing x-risks? That doesn’t seem right. Am I missing something?
“The thing that really gets me is that Democrats try to offer policies (paid sick leave! minimum wage!) that would help the working class,” a friend just wrote me. A few days’ paid leave ain’t gonna support a family. Neither is minimum wage. WWC men aren’t interested in working at McDonald’s for $15 per hour instead of $9.50. What they want is what my father-in-law had: steady, stable, full-time jobs that deliver a solid middle-class life to the 75% of Americans who don’t have a college degree. Trump promises that. I doubt he’ll deliver, but at least he understands what they need.
@rowborg/@kbog, this is a great thread, and I’m only going to dip my toe in, here, minimally, to +1 this point that rowborg has made. Employment isn’t just about exchanging labour for a wage, it’s a major source of meaning in people’s lives. The assertion that standards of living in the US have risen is probably correct, even for people at the bottom of the income distribution. But this standard of living increase can’t explain the incredibly sour mood of a huge chunk of people. Obviously it’s going to take a lot of time for researchers to unpack that, but the meaningless-employment story seems like a pretty plausible component: if people feel what they do doesn’t have any meaning, and particularly if they feel their children don’t have prospects for meaning-ful employment then despair and anger seem like totally to-be-expected reactions.
I think we need to see more data on this one. I’d put forth the idea that in fact the underlying issues were in fact economic, and that the cultural and immigration issues were mostly just utilized by Trump as things to blame for the underlying economic woes.
I saw it based on data from exit polls. I don’t know where I saw it. If there is any other evidence, sure that would be neat to see. Though it is objectively the case that our underlying economic woes are smaller now than they were in 2012 and 2008.
Wait, I’m new to x-risk analysis and TDT… but is that true? Why would the distribution of the space of all possible actions split 50⁄50 towards increasing or decreasing x-risks? That doesn’t seem right. Am I missing something?
Well think about it this way, if the majority of political positions increased x-risk, we would make the world safer just by doing things in the opposite way all the time. But that’s absurd.
The more general point is that you can construct a complicated chain of reasoning for lots of things to increase x risk, so it’s not particularly interested just that something increases it. What matters is how much and how tractable it is.
Though it is objectively the case that our underlying economic woes are smaller now than they were in 2012 and 2008.
On aggregate, yes. I’m talking about the economic situation for a certain group, which is the lower-middle class. I don’t know that this is true for them. I’ll go try and dive in on this.
Well think about it this way, if the majority of political positions increased x-risk, we would make the world safer just by doing things in the opposite way all the time.
If I’ve parsed your meaning correctly, you are saying every action has a counterfactual non-action, and if you include not just the space of possible actions but the impact of non-action, then risks break 50/50? That makes more sense, but doesn’t really seem to be helpful in deciding between actions to take. The space of possible actions to take is what should concern us, yes?
Regardless, I take your larger point, which is that we should put efforts where you can maximize impact and you are skeptical on the effective value of this one as opposed to others. I would certainly need to be more quantitative in my analysis of this and back up my ideas/claims for sure.
The more general point is that you can construct a complicated chain of reasoning for lots of things to increase x risk, so it’s not particularly interested just that something increases it. What matters is how much and how tractable it is.
Totally agreed. I have yet to hear an argument that 1) invalidates my chain of reasoning or 2) indicates there are other inflection points along the chain that are better. Whether or not this compares well with other options for x-risk reduction is an exercise I have yet to perform (as I mentioned above).
| More like x-risk by proxy by proxy.
Fair enough.
| It should be easier to focus directly on building international institutions.
Can you clarify this? As a new EA, I am unfamiliar with what this might be in reference to. Are you arguing for focusing on the creation of a second nation-state promoting liberal values and technological advancement (to backstop US leadership)? Or just circumventing politics and going straight at the underlying risks (AI, nuclear proliferation)? It seems without the air-cover of a liberal superpower, pursuing such ventures will be significantly harder.
Well for Trump increasing x risks, the potential downside is that there is going to be less international cooperation and less prospect for coordination between nations on issues like AI. That’s what I thought you had in mind and it implies that we should try to build more credible and effective groups like the UN and IAEA. It’s easier to work on this directly rather than doing things that make people less likely to support populists who are against international involvement. I guess there are other ways Ttump could be bad for x risks, so that might change things.
Gotcha. I was going the other way; considering what automation and globalization have done to low-skilled workers over the last 30 years (which is only going to get a lot worse), some sort of significant restructuring of our economic system (both wealth distribution and cultural self-actualization via employment) seems necessary to prevent full-scale upheaval (e.g. revolution, civil war, etc). Trump seems to be the first major indication that this problem is getting so bad that it could literally end America or at least American leadership and the promotion of liberal values.
As Qloachu said in a different thread on this page:
I do agree this is a very indirect effect but I also failed to see how more direct action would be more effective.
Well that is different but I don’t think it’s a likely threat.
Low-skilled workers have seen general increases in standard of living over the last 30 years and no great increase in unemployment. Globalization generally increases wealth for all income brackets in the long run and automation has not historically increased long-term unemployment.
We may be able to have trucks that steer themselves on the road, but actually having unmanned vehicles take themselves from A to B reliably enough to be unmanned and without requiring a human in the cabin to do anything is years away.
In any case, the lack of truck drivers is a significant constraint on our economy alongside the fact that they can only drive 10 or so hours a day. Automation here might be unusually good.
It seems to me like wealth is a pretty straightforward question of welfare to resolve. As for cultural self-actualization… we’ve had retirees, part-time employees, and permanently unemployed people for a long time, but they never seem to be more troublesome or more dissatisfied than the rest of the electorate.
I don’t think so. We’ve had similar issues in the past with populist leaders and their movements, e.g. William Jennings Bryan (barely failed to get elected) and Huey Humphrey (assassinated). But our democracy has only improved over time.
Is this true in the US? My personal circle of concern certainly includes all of humanity (and beyond), but Trump’s election would appear to be an indication that displaced Americans don’t think that way.
Do you have some info you can share on that? Everyone I’ve talked to in the autonomous vehicle area has indicated that adoption, not technology, will be the long pole here. And as you indicated, there are massive economic forces that will push for adoption.
And while I totally agree automation will be good for GDP, it will continue to displace a set of people who can vote.
Retirees already have cultural approval to not work: they’ve “earned” it. Part-timers and permanently unemployed are viewed as lazy or moochers by many folks in the US, where there’s a culture of self-improvement via working hard. That’s what I’m arguing will need to change.
What metric are you using for improvement? On the axis of existential risk prevention, I would argue we’ve just taken a big step back.
Interested to hear your thoughts. Thanks!
Yes, average income when adjusted by PPP has increased (the standard narrative is that CPI-adjusted income has not increased, but the CPI is flawed—it overestimates inflation by focusing on cities and it ignores goods substitution), and nonwage compensation has increased (health benefits etc).
Sure, class 8 trucks are a couple hundred thousand dollars each and the cargo is worth a lot too. They have mechanical problems. They have expensive insurance. They have complex routes. They have to reliably reach their destination and meet with the proper people to load and unload. With all this responsibility and risk it’s an easy decision to put a driver at the wheel, especially when even ordinary cars aren’t ready to go around without drivers.
Well there you have it. If it will take a long time to be adopted then there won’t be an unemployment problem for a long time.
They will vote for welfare, which they will get.
I don’t think any of this takes away from my point.
None of the democracy and freedom indices have been around for that long, so there is no metric. But we can look at the corruption and civil values of 100 years ago, and comparing it today, and note that it’s pretty clear that things are better all round now than they were back then. For instance, we don’t prevent women and blacks from voting anymore, and we don’t put people in prison for speaking out against the war. We don’t hold racist views about who is eligible to hold public office.
That’s a different axis, and it has taken many steps back, perhaps with the elections of GWB, Ronald Reagan, Eisenhower, and Truman.
The distribution is what would matter for my argument. Do you have any links to data there? Where is this data from? Thanks.
I have first hand information from those knowledgable on the topic that the technology will be ready in the next few years, and I haven’t heard a refutation of that. So I said the long pole was adoption. I didn’t say how long the pole was. You already acknowledged that the economic incentives for adoption are huge. I posit that adoption will be faster than you suspect. But we have to wait and see on this one, unless you have more information to share on why you disagree.
They didn’t vote for welfare this time. Which is my point.
Maybe you can explain… you seemed to be saying that the fact we have part-timers and unemployed proves that we don’t have any cultural stigma against not working. I said that people do have such a stigma, and then you said that doesn’t take away from your point. Not sure what I missed.
Great, that’s the axis I was trying to focus on, which is why I asked what metric you were using. And yes, it’s taken steps back before. That’s doesn’t preclude it getting worse.
I’m simply pointing out that if automation and globalization disrupt the US middle class enough, political instability could increase x-risks.
Haven’t seen that data specifically but it’s inferred:
http://www.quickmba.com/econ/macro/cpi/
https://youtu.be/8sf3kt1KduY
https://fred.stlouisfed.org/series/MEFAINUSA672N (I guess family income increased even with cpi)
You’re talking about what would happen in a world of high structural unemployment. But we don’t have high structural unemployment, so you can’t say that future elections will have the same dynamics. In this election cycle, Trump voters weren’t primarily motivated by the economy, they were motivated by immigration and culture issues.
I said that it shows that unemployed people aren’t likely to start a revolution or anything just because they’re unemployed.
People don’t start upheavals just because they don’t have cultural acceptance.
If you take a broad view then half of all things will increase x risk in the long run and half of them will decrease it. I think that the issue here is too crowded and intractable and the effects are too complicated and uncertain for it to be a big goal for us to put effort on, though I agree that it would be good if the nation had a robust social system going forward.
I think we need to see more data on this one. I’d put forth the idea that in fact the underlying issues were in fact economic, and that the cultural and immigration issues were mostly just utilized by Trump as things to blame for the underlying economic woes.
Here’s an article I found instructive on the perspective and here’s a relevant passage:
I thought this analysis was interesting.
Ah, I understand your point now. Sure, that’s true, but I’m imagining what happens if the lower-middle class gets squeezed out, as mentioned above. I need to go learn more about economic trends of the middle class in American history before I can make this case, though.
Wait, I’m new to x-risk analysis and TDT… but is that true? Why would the distribution of the space of all possible actions split 50⁄50 towards increasing or decreasing x-risks? That doesn’t seem right. Am I missing something?
@rowborg/@kbog, this is a great thread, and I’m only going to dip my toe in, here, minimally, to +1 this point that rowborg has made. Employment isn’t just about exchanging labour for a wage, it’s a major source of meaning in people’s lives. The assertion that standards of living in the US have risen is probably correct, even for people at the bottom of the income distribution. But this standard of living increase can’t explain the incredibly sour mood of a huge chunk of people. Obviously it’s going to take a lot of time for researchers to unpack that, but the meaningless-employment story seems like a pretty plausible component: if people feel what they do doesn’t have any meaning, and particularly if they feel their children don’t have prospects for meaning-ful employment then despair and anger seem like totally to-be-expected reactions.
I saw it based on data from exit polls. I don’t know where I saw it. If there is any other evidence, sure that would be neat to see. Though it is objectively the case that our underlying economic woes are smaller now than they were in 2012 and 2008.
Well think about it this way, if the majority of political positions increased x-risk, we would make the world safer just by doing things in the opposite way all the time. But that’s absurd.
The more general point is that you can construct a complicated chain of reasoning for lots of things to increase x risk, so it’s not particularly interested just that something increases it. What matters is how much and how tractable it is.
On aggregate, yes. I’m talking about the economic situation for a certain group, which is the lower-middle class. I don’t know that this is true for them. I’ll go try and dive in on this.
If I’ve parsed your meaning correctly, you are saying every action has a counterfactual non-action, and if you include not just the space of possible actions but the impact of non-action, then risks break 50/50? That makes more sense, but doesn’t really seem to be helpful in deciding between actions to take. The space of possible actions to take is what should concern us, yes?
Regardless, I take your larger point, which is that we should put efforts where you can maximize impact and you are skeptical on the effective value of this one as opposed to others. I would certainly need to be more quantitative in my analysis of this and back up my ideas/claims for sure.
Totally agreed. I have yet to hear an argument that 1) invalidates my chain of reasoning or 2) indicates there are other inflection points along the chain that are better. Whether or not this compares well with other options for x-risk reduction is an exercise I have yet to perform (as I mentioned above).
Thanks for this!