That makes sense! After SBF, I’ve wondered about the causality around charitable giving and criminality. If GP signatories are more criminal than other billionaires, that would suggest that individuals may be accepting criminality that they can justify with altruism. But if they’re similarly criminal to other rich people, and rich people are more criminal than other groups, then that suggests a broader societal problem that we should adjust our legal system to properly disincentivize criminality.
andrew_richardson
Karma: 275
The base rate seems tricky to interpret without a real comparison. 3 out of 5 points in the summary say that the rate is high, but it doesn’t seem like the evidence exists to support that claim. The implied reference class is “non EA super-rich people”, but that class isn’t actually examined in the data.
The rates do seem subjectively very high. Way fewer than 10% of people I know have been convicted of financial crimes! But I wonder if founders and CEOs are being blamed for financial crimes that their companies commit, and approximately all successful companies commit financial crimes, defined broadly.
One interesting comparison is that the percent who have ever spent a night in jail is similar to the rate for the overall American population.
Overall great data though! It raises real concerns.
I’m curious, do you think the Giving Pledge should have higher standards? My hot take is that we don’t want to help launder reputations, but that I would personally be very happy if e.g. the Sacklers donated all their money to a real charity.
I’m concerned that the control group doesn’t seem like a representative comparison. It seems like Giving Pledge signatories are older, richer, more famous, and more successful than the sample of YC recipients.
A better comparison group might be the world’s 100 richest people, would it be possible to redo this analysis for them?
Hey @Fernando_MG, is this effort still ongoing? I’ve noticed that artists are paying a lot more attention to AI recently, because of the impact of generative visual AI on the industry. Are you working on ways to leverage that interest?
You should note in the intro that LMIC stands for “Low and Middle Income Countries”. I love the images.
Yeah I personally find it very hard to do ML interviews for that reason. So far I’m doing a mix of theory/conceptual questions and practical ML coding questions. It helps if the conceptual questions include some unusual setups, or ask about unusal tweaks.
It depends a lot on what I’m hiring for and what the candidate’s background is. If it’s an ML job, I ask ML programming and design questions, but if I’m hiring someone to do networking, I’ll ask a question about distributed algorithms or something. This is in contrast to how Google hires, because they’re hiring generic SWEs so they don’t care a lot about the particulars.
If I had to hire someone more senior, I would reach out to other more senior people who I trust to ask them for help.
For a small team, generalism can be more important than it seems.
If there’s too much demand for this, I’d also be willing to do these.
When I was interviewing at big tech companies, it helped that I had 3 roommates who worked for Google. In general it seems like EAs who want to work for big tech but don’t have existing connections to it are disadvantaged. That seems suboptimal. I’m glad you made this post.
It’s entirely possible to go into alignment work 10 or 15 years from now. If you spend the next decade working at tech companies, working with AI, or starting a company, you’re going to end that time with a ton of useful skills that you can bring to alignment research if you want to. But you’ll also have flexibility to work on other problems too.
My advice: pursue profit-driven work when you’re young, come back to direct EA work when you’re more experienced.
What are some effective charities that are working to address the situation?
The GiveDirectly campaign in Yemen seems similar, across the Gulf of Aden. They budget $192/person/4 months for basic needs.
Most vegetarians and vegans aren’t taking the vitamins that would be good for them, and that can be a health problem. What about subsidizing vitamins somehow? Maybe we could get vegetarian meal kits to include vitamins.
“Veg*n” encompasses both vegetarians and vegans.
I don’t think the math checks out here. I do a quick cost benefit analysis here, and meat vs. climate interventions don’t seem like comparable orders of magnitude to me. Not eating meat appears to be a much more cost effective sacrifice.
(Math: Responding to Robi’s comparison between not eating 1kg chicken (meat for 3 days) vs not using AC for 3 days. I’m not sure where the 10^(-14) °C number came from, but let’s assume it draws ~1,000 watts, or about ~100,000 watt-hours/3days. US power generation apparently produces about 0.85 pounds of CO2 emissions per kWh, so that’s 85 pounds of carbon. At about 2 trillion tons of carbon per degree Celcius, that is indeed about 10^(-14) °C of warming.)
Given a value of 10^8 deaths per degree of warming (notes), this comparison implicitly values human to chicken lives at about 1,000,000 : 1. (10^-6 = 10^8 * 10^-14) (1 chicken produces ~1kg meat)
Reasonable EAs disagree, but I guess I value chickens at more like a 10,000:1 ratio to humans, especially given how much suffering factory farming entails. Given that not eating meat for 3 days and not having AC for 3 days feel like comparably small hardships, the impact of not eating meat seems to be about 100 times higher for the same cost.
This sort of cost-benefit analysis makes me feel like personal sacrifices for the climate are not very high value compared to veganism.
Are there personal climate sacrifices that are more worthwhile than this, or is this line of reasoning incorrect somehow?
Could you describe a little bit more about the fight against malnutrition? Do you buy food for people, and if so what kind? What other sorts of program do you run? What is your budget like?
Pay for all the productivity apps that you’re thinking of using. Task managers like Amazing Marvin or note taking apps like Evernote are all slightly better in the paid version, and at least for me it’s always been worth it.
Around the house, spend money on shelving. Having plenty of storage space makes it easier to stay organized.
I’m worried that in some cases it might be the case that grant makers and grant receivers are friends who actively socialize with each other, and that might corrupt the grantmaking process.
Can you elaborate on your endnote that “this discussion won’t work”? What would success look like?
Effective Alternatives to Street Protests in Russia
The War in Ukraine has led many people to protest in the streets. In the west, this slightly affects the world by showing support for the war. But within Russia and Belarus, protesters are often arrested, and it’s not clear that they affect public opinion very much.
This seems like a lost opportunity, because people living in Russia need to be involved in any successful liberalization of the country’s government, which affects the rest of the world through the threat of nuclear war.
It’s clear that there’s a large number of people who live under a repressive government, who want to change it, and who have gotten the message that marching in the streets with a sign is a good way to do that. They’re willing to risk arrest and imprisonment to do it.
Assuming that there are more impactful things that those people can do, it would be good if someone figured out what those were and wrote them down clearly, ideally in Russian.
I don’t know what the answer is, which is why this is filed under Shortform. I don’t have the background to answer a social question like this well.
EA ideas suggest that most people who go to street protests on the weekend could better serve their goals by pursuing an impactful career, and street dissidents in Russia are probably no exception.
Thank you for your responses! I added edits to the essay to reflect this.
Overall, as I noted in the edits, this exercise has made me shift from being skeptical about all climate change interventions to considering shifting some donations from global poverty to climate change interventions. Not entirely convinced, but it seems a lot more plausibly effective than I first suspected.
Some things I don’t understand though:
It makes sense that with a convex harms curve, marginal harms will be worse than this back of the envelope linear calculation suggests. But it’s surprising that they’re 10 times higher. I guess it’s just very nonlinear, as you say, but that’s surprising to me.
The $1/ton estimate comes from CATF, which is a lobbying organization. Their cost effectiveness calculations account for money they spend lobbying, but not deadweight loss caused by taxes and regulations. How reasonable is it accept that sort of accounting?
Thank you for collecting the data in these surveys! Do you have an idea for the ETA for the 2023 data?
My motivation is that when I talk to people about EA, I often hear the criticism that EA is all about AI risk, and I respond by saying that actually EAs in practice mostly prioritize and donate to global health causes. I’ve been linking the 2019 survey for a few years now and I’m starting to notice that it’s out of date.