RE: why aren't there as many EAs giving this much money: I'm (obviously) not Jeff, but I was at Alphabet for many of the years Jeff was. Relevantly, I was also involved in the yearly donation matching campaigns. There were around 2-3 other folks who donated similar amounts to Jeff. Those four-ish people were the majority of EA matching funds at Alphabet.
It's hard to be sure how many people actually donated outside of giving campaigns, so this might undercount things. But to get to 1k EAs donating this much money, you'd need like 300 companies with similarl...
Legal or constitutional infeasibility does not always prevent executive orders from being applied (or followed). I feel like the US president declaring a state of emergency related to AI catastrophic risk (and then forcing large AI companies to stop training large models) sounds at least as constitutionally viable as the attempted executive order for student loan forgiveness.
I agree that this seems fairly unlikely to happen in practice though.
At the time of me posting this there are 3 duplicate posts on this:
https://forum.effectivealtruism.org/posts/ycCXDofXr89DyxKqm/wealth-redistribution-are-we-on-the-same-page-1
https://forum.effectivealtruism.org/posts/3Gp3mKF4mg3aXqEKC/wealth-redistribution-are-we-on-the-same-page-5
https://forum.effectivealtruism.org/posts/7iAkFkaj7cHJbMZWd/wealth-redistribution-are-we-on-the-same-page-6
While this is in some ways poetic about a post asking whether or not we are all on the same page, I'm guessing you want to delete the duplicates.
I deeply appreciate the degree to which this comment acknowledges issues and provides alternative organizations that may be better in specific respects. It has given me substantial respect for LTFF.
This feels like a "be the change you want to see in the world" moment. If you want such an event, it seems like you could basically just make a forum post (or quick take) offering 1:1s?
I think that basically all of these are being pursued and many are good ideas. I would be less put off if the post title was 'More people should work on aligning profit incentives with alignment research', but suggesting that no one is doing this seems off base.
This is what I got after a few minutes of Google search (not endorsing any of the links beyond that they are claiming to do the thing described).
AI Auditing:
https://www.unite.ai/how-to-perform-an-ai-audit-in-2023/
Model interpretability:
https://learn.microsoft.com/en-us/azure/machine-learning/how-to-...
I agree that 'utilitarianism' often gets elided into meaning a variation of hedonic utilitarianism. I would like to hold philosophical discourse to a higher bar. In particular, once someone mentions hedonic utilitarianism, I'm going to hold them to the standard of separating out hedonic utilitarianism and preference utilitarianism, for example.
I agree hedonic utilitarians exist. I'm just saying the utilitarians I've talked to always add more terms than pleasure and suffering to their utility function. Most are preference utilitarians.
I feel like 'valuism' is redefining utilitarianism, and the contrasts to utilitarianism don't seem very convincing. For instance, you define valuism as noticing what you intrinsically value and trying to take effective action to increase that. This seems identical to a utilitarian whose utility function is composed of what they intrinsically value.
I think you might be defining utilitarianism such that they are only allowed to care about one thing? Which is sort of true, in that utilitarianism generally advocates converting everything into a common scale, b...
This comment came across as unnecessarily aggressive to me.
The original post is a newsletter that seems to be trying to paint everyone in their best light. That's a nice thing to do! The epistemic status of the post (hype) also feels pretty clear already.
Yeah, I hear you. [Edit: well, I think it was the least aggressive way of saying what I wanted to say.]
(I note that in addition to hyping the post is kinda making an ask for funding for the three projects it mentions--"Some of our favorite proposals which could use more funding"--and I'm pretty uncomfortable with one-sided-ness in funding-asks.)
As someone who went through the CEA application process, I wholeheartedly endorse this. I was also really impressed with CEA's approach the process, and their surprising willingness to give feedback & advice through it.
[It ended up being a mutually bad fit. I've spent my whole career as a C++ backend engineer at a FAANG and I like working in person, and that doesn't align super well with a small remote-first org that has a lot of frontend needs.]
It feels weird to me to hear that something is terrible to think. It might be terrible that we're only alive because everyone doesn't have the option to kill everyone else instantly, but it's also true. Thinking true thoughts isn't terrible.
If everyone has a button that could destroy all life on the planet, I feel like it's unrealistic to expect that button to remain unpressed for more than a few hours. The most misanthropic person on Earth is very, very misanthropic. I'm not confident that many people would press the button, but the whole thing is that it...
If AI + a nontechnical person familiar with business needs can replace me in coding, I expect something resembling a singularity within 5 years.
I think that software engineering is a great career if you have an aptitude for it. It's also way easier to tell if you are good at it relative to most other careers (ie, Leetcode, Hackerrank, and other question repositories can help you understand your relative performance).
So my answer is that either AI can't automate software engineers for a while, or they'll automate every career quite soon after software engin...
IANAL: I view 'effective altruism' to not be owned, and if any organization claims to own the term I'm going to ignore them. I expect most folks to share my opinion here.
Oh for sure, I wasn't thinking you were implying making it a requirement. I was trying to say that even a nudge towards explaining downvotes is a nudge towards evil (for me).
Maybe the net advantage of explaining downvotes would be good, but I personally should probably be discouraged from explaining my downvotes.
I disagree and I downvoted this because explaining why you downvoted something is disproportionately likely to end up with me arguing with someone on the internet. I find this really unpleasant.
I'm happy to have a rule for giving an explanation to you if I downvote your posts. I've talked with you as a person outside of internet arguments, so I'm not as worried about getting into a protracted argument.
But as a general rule, I think I should be discouraged from explaining my downvotes so that I keep up my mental health.
Separately, if this was a thread that ...
Fair point! I was assuming that by collective decision making you meant much closer to 1 person 1 vote, but if it's well defined term I'm not sure of the definition.
I haven't heard much discussion on a market-based feedback system, and I'd be very interested in seeing it tried. Perhaps for legal or technical reasons it wouldn't work out super well (similar to current prediction markets), but it seems well worth the experiment.
I think that this incorrectly conflates prediction markets and collective decision making.
(Prediction) markets are (theoretically) effective because folks that are able to reliably predict correctly will end up getting more money, and there are incentives in place for correct predictions to be made. It seems that the incentives for correct decision making are far weaker in collective decision making, and I don't see any positive feedback loop where folks that are better at noticing impactful projects will get their opinions weighted more highly.
I think tha...
While I agree with this question in the particular, there's a real difficulty because absence of evidence is only weak evidence of absence with this kind of thing.
This post makes it harder than usual for me to tell if I'm supposed to upvote something because it is well-written, kind, and thoughtful vs whether I agree with it.
I'm going to continue to use up/downvote for good comment/bad comment and disagree/agree for my opinion on the goodness of the idea.
[EDIT: addressed in the comments. Nathan at least seems to endorse my interpretation]
Thanks, I think this is an excellent response and I agree both are important goals.
I'm curious to learn more about why you think that steelmanning is good for improving one's beliefs/impact. It seems to me that that would be true if you believe yourself to be much more likely to be correct than the author of a post. Otherwise, it seems that trying to understand their original argument is better than trying to steelman it.
I could see that perhaps you should try to do both (ie, both the author's literal intent and whether they are directionally correct)?
[EDI...
I'd be interested to see some of those tried for sure!
I imagine you'd also likely agree that these proposals tradeoff against everything else that the EA orgs could be doing, and it's not super clear any are the best option to pursue relative to other goals right now.
I agree. I think that it's incredibly difficult to have civil conversations on the internet, especially about emotionally laden issues like morality/charity.
I feel bad when I write a snotty comment and that gets downvoted, and that has a real impact on me being more likely to write a kind argument in one direction rather than a quick zinger. I am honestly thankful for this feedback on not being a jerk.
Do you think that group bargaining/voting in EA would be a good thing for funding/prioritization?
I personally like the current approach that has individual EAs and orgs make their own decisions on what is the best thing to do in the world.
For example, I would be unlikely to fund an organization that the majority of EAs in a vote believed should be funded, but I personally believed to be net harmful. Although if this situation were to occur, I would try to have some conversations about where the wild disagreement was stemming from.
I think there's probably a bunch of different ways to incorporate voting. Many would be bad, some good.
Some types of things I could see being interesting:
In the interests of taking your words to heart, I agree that EAs (and literally everyone) are bad at steelmanning criticisms.
However, I think that saying the 'and literally everyone' part out loud is important. Usually when people say 'X is bad at Y' they mean that X is worse than typical at Y. If I said, 'Detroit-style pizza is unhealthy,' then there is a Gricean implicature that Detroit-style pizza is less healthy than other pizzas. Otherwise, I should just say 'pizza is unhealthy'.
Likewise, when you say 'EAs seem particularly bad at steelman...
[aside, made me chuckle]
This is an inevitable issue with the post being 70 pages long.
I think online discussions are more productive when its clear exactly what is being proposed as good/bad, so I appreciate you separately commenting on small segments (which can be addressed individually) rather than the post as a whole.
Thanks for including this! I really liked the shrimp sticker, and partly I liked it because it simply came across as friendly. I honestly didn't know that live shrimp have different ordinary posture and color compared to cooked shrimp, and that makes the sticker feel a lot less friendly to me!
I'd ideally like a sticker with what looks like a happy shrimp. A live shrimp in a circle with something like 'expanding the moral circle' feels like almost exactly the vibe I'd love to send out, for what it's worth.
Separately, I get that making merch/art/anything like this is difficult, so I appreciate the work that has already gone into putting the store together.
I wanted to mention that I went through the first week's lectures and exercises and I was really impressed at the quality!
Also a software engineer, and this also is a pretty spot on description for me. 25 hours of productive work is about my limit before I start burning out and making dumb mistakes.
I instead read this point as saying "assume that if we persuaded 100 folks to give up carp, then 1 of those would replace their carp consumption with salmon." So it's talking about the replacement effect, rather than the number persuaded (the latter gives magnitude, as you say).
You can see how that the lack of details is basically asking me to... trust you without evidence?
Edit: to use less 'gotcha phrasing, anonymously claiming that another organization is doing better on feedback, but not telling me how, is asking for me to blindly trust you for very little reason.
I don't think feedback practices are widely considered secrets that have to be protected, and if your familiarity is with the UK Civil Service, that's a massive organization where you can easily give a description without unduly narrowing yourself down.
This thread is the kind of tiring back and forth I'm talking about. Please, try organizing feedback for 5k+ rejected applicants for something every year and then come back to tell me why I'm wrong and it really is easy. I promise to humbly eat crow at that time.
Phone calls for me are socially awkward and I generally want some time to privately process rejection rather than immediately need to have a conversation about it. Also I generally keep my phone at home during business hours so it's quite likely I'd need to spend half an hour playing phone tag.
Was going to say the same. I've only ever been rejected over email (or ghosted entirely). I would also find it off-putting to get a phone call rejection. I guess organizations can choose to call if they wanted, but I wouldn't personally encourage it.
A hypothetical example that I would view as asking for trust would be someone telling me not to join an organization, but not telling me why. Or claiming that another person shouldn't be trusted, without giving details. I personally very rarely see folks do this. An organization doing something different and explaining their reasoning (ex. giving feedback was not viewed as not a good ROI) is not asking for trust.
Regarding why giving feedback at scale is hard, most of these positions have at best vague evaluation metrics which usually bottom out in "help t...
I guess I read that as a description of what they're doing rather asking me to trust them. CEA can choose the admission criteria they want, and after attending my first EAG earlier this year I felt like whatever criteria they were using seemed to broadly make for a valuable event for me as an attendee.
I think you're really underestimating how hard giving useful feedback at scale is and how fraught it is. I would be more sympathetic if you were running or deeply involved with an organization that was doing better on this front. If you are, congrats and I am appreciative!
Separately regarding trust, I don't feel obligated to trust senior EAs. I sometimes read the analyses of senior EAs and like them, so I start to trust them more. Trust based on seniority alone seems bad, could you give some examples where you feel senior EAs are asking folks to trust them without evidence?
I am telling you what Google told me (and continues to tell new interviewers) as part of its interview training. You may believe that you know the law better than Google, but I am too risk averse to believe that I know the law better than them.
I'm sorry to hear about your negative experience with GiveWell's hiring cycle.
I think that it's easy to under-estimate how hard it is to hire well though. For comparison, you can honestly give all the same complaints about the hiring practice of my parent company (Google).
It is slow, with many friends of mine experiencing up to a year of gap between application and eventual decision.
Later interviewers have no context on your performance on earlier parts of the application. This is actually deliberate though, since we want to get independent signal at each ...
An explicit part of interviewer training is noting that we shouldn't say anything about a candidate's performance (good or bad) to the candidate, for fear of legal repercussions.
Legal repercussions for interview feedback have been discussed on the EA Forum in the past, e.g. in the comments of this post. The consensus seems to be that it's not an issue either in theory or in practice. Certainly if your feedback is composed of lawyer-approved boilerplate, and only offered to candidates who ask for it, I think your legal risk is essentially nil. [Edit: ...
I worry that this post is claiming that EAs are uncommonly likely to recommend rules violations in order to achieve their goals (ie, ends justify the means). I don't think that's true, and I generally see EAs as trying very hard to be scrupulous and do right by all involved parties.
Concretely, I believe that if you went to an EA conference or a similar gathering and presented people with prisoners dilemma issues, or just lost wallets, they would behave more pro-socially than average for the country.
I think that the FTX collapse is a very salient example of EA folks committing crime (perhaps in the belief that the ends justified it?), but that doesn't mean that EA increases the probability of crime.
As someone who has recently been in the AI Safety org interview circuit, about 50% of interviews were traditional Leetcode style algorithmic/coding puzzle and 50% were more practical. This seems pretty typical compared to industry.
The EA orgs I interviewed with were very candid about their approach, and I was much less surprised by the style of interview I got than I was surprised when interviewing in industry. Anthropic, Ought, and CEA all very explicitly lay out what their interviews look like publicly. My experience was that the interviews matched the public description very well.
Thanks! Current volume is reasonable but I will totally forward some your way if I get overwhelmed.
I forgot to mention in the body, but I should thank Yonatan for putting a draft of this together and encouraging me to post it. Thanks! I've been meaning to do this for a while.
This is a touchingly earnest comment. Also is your ldap qiurui? If those words mean nothing to you, I've got the wrong guy :)
There are a lot of 'lurkers', but less than 30 folks would be involved in the yearly holiday matching thread and sheet. Every self-professed EA I talked to at Google was involved in those campaigns, so I think that covers the most involved US Googlers.
Most people donated closer to 5-10% than Jeff or Oliver's much higher amounts, that is for sure true.
So I think both your explanations are true. There are not that many EAs at Google (although I don't think that's surprising), and most donate much less than they likely could. I put myself in that bucket, as ... (read more)