rowborg

Posts

Sorted by New

Comments

What does Trump mean for EA?

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).

Thanks for this!

What does Trump mean for EA?

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.

Here's an article I found instructive on the perspective and here's a relevant passage:

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?

What does Trump mean for EA?

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.

What does Trump mean for EA?

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.

Interested to hear your thoughts. Thanks!

What does Trump mean for EA?

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:

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.

What does Trump mean for EA?

| 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.

What does Trump mean for EA?

Given Trump's high uncertainty and variance, one might imagine that the chances of him doing something that increases x-risks significantly is at least one or two orders of magnitude higher than Clinton's.

Given that, restructing the US economy to better deal with the shocks of automation and globalization, which might prevent future instances of this type of intentional systemic upheaval, suddenty seems like a potentially important new category of x-risk by proxy.

By that logic, does something like UBI become an important potential area of focus?