Ben Millwood

The illu­sion of con­sen­sus about EA celebrities

• Ben Millwood 18 Jun 2024 13:34 UTC

  123 points

  53 ∶ 21

  on: My ex­pe­rience at the con­tro­ver­sial Man­i­fest 2024

  One aspect of the framing here that annoyed me, both in the OP and in some of the comments: the problem is not controversial beliefs, it is exclusionary beliefs. Here are some controversial beliefs that I think would pose absolutely no problem at this event or any other:

  • The many-worlds interpretation of quantum mechanics.

  • Virus gain-of-function research creates more risk than it prevents.

  • Nuclear energy is a necessary part of the transition away from fossil fuels.

  The problem with racism and transphobia is not that people disagree about them! The problem is that these beliefs, in their content on the object level, hurt people and exclude people from the discussion.

  Let’s avoid using “controversial” as a euphemism for “toxic and exclusionary”. Let’s celebrate the debate and discussion of all controversies that threaten no-one and exclude no-one. Suggesting any of that is at stake is totally unnecessary.

  What links here?

  • Ben Millwood's comment on Why so many “racists” at Man­i­fest? by Austin (18 Jun 2024 16:04 UTC; 63 points)

Florida and Alabama ban cul­ti­vated meat

• Ben Millwood 8 Oct 2021 19:44 UTC

  89 points

  0 ∶ 0

  on: The Cost of Rejection

  It’s a little aside from your point, but good feedback is not only useful for emotionally managing the rejection—it’s also incredibly valuable information! Consider especially that someone who is applying for a job at your organization may well apply for jobs at other organizations. Telling them what is good or bad with their application will help them improve that process, and make them more likely to find something that is the right fit for them. It could be vital in helping them understand what they need to do to position themselves to be more useful to the community, or at least it could save the time and effort of them applying for more jobs that have the same requirements you did, that they didn’t meet—and save the time and effort of the hiring team there rejecting them.

  A unique characteristic of EA hiring is that it’s often good for your goals to help candidates who didn’t succeed at your process succeed at something else nearby. I often think we don’t realize how significantly this shifts our incentives in cases like these.

• Ben Millwood 18 Jun 2024 10:45 UTC

  73 points

  10 ∶ 2

  on: BenMill­wood’s Shortform

  AI Safety Needs To Get Serious About Chinese Political Culture

  I worry that Leopold Aschenbrenner’s “China will use AI to install a global dystopia” take is based on crudely analogising the CCP to the USSR, or perhaps even to American cultural imperialism /​ expansionism, and isn’t based on an even superficially informed analysis of either how China is currently actually thinking about AI, or what China’s long term political goals or values are.

  I’m no more of an expert myself, but my impression is that China is much more interested in its own national security interests and its own ideological notions of the ethnic Chinese people and Chinese territory, so that beyond e.g. Taiwan there isn’t an interest in global domination except to the extent that it prevents them being threatened by other expansionist powers.

  This or a number of other heuristics /​ judgements /​ perspectives could change substantially how we think about whether China would race for AGI, and/​or be receptive to an argument that AGI development is dangerous and should be suppressed. China clearly has a lot to gain from harnessing AGI, but they have a lot to lose too, just like the West.

  Currently, this is a pretty superficial impression of mine, so I don’t think it would be fair to write an article yet. I need to do my homework first:

  • I need to actually read Leopold’s own writing about this, instead of making impressions based on summaries of it,

  • I’ve been recommended to look into what CSET and Brian Tse have written about China,

  • Perhaps there are other things I should hear about this, feel free to make recommendations.

  Alternatively, as always, I’d be really happy for someone who’s already done the homework to write about this, particularly anyone specifically with expertise in Chinese political culture or international relations. Even if I write the article, all it’ll really be able to be is an appeal to listen to experts in the field, or for one or more of those experts to step forward and give us some principles to spread in how to think clearly and accurately about this topic.

  I think having even like, undergrad-level textbook mainstream summaries of China’s political mission and beliefs posted on the Forum could end up being really valuable if it puts those ideas more in the cultural and intellectual background of AI safety people in general.

  This seems like a really crucial question that inevitably takes a central role in our overall strategy, and Leopold’s take isn’t the only one I’m worried about. I think people are already pushing national security concerns about China to the US Government in an effort to push e.g. stronger cybersecurity controls or export controls on AI. I think that’s a noble end but if the China angle becomes inappropriately charged we’re really risking causing more harm than good.

  (For the avoidance of doubt, I think the Chinese government is inhumane, and that all undemocratic governments are fundamentally illegitimate. I think exporting democracy and freedom to the world is a good thing, so I’m not against cultural expansionism per se. Nevertheless, assuming China wants to do it when they don’t could be a really serious mistake.)

• Ben Millwood 21 May 2022 15:41 UTC

  70 points

  0 ∶ 0

  on: Bad Omens in Cur­rent Com­mu­nity Building

  I feel a desire to lower some expectations:

  • I don’t think any social movement of real size or influence has ever avoided drawing some skepticism, mockery, or even suspicion,

  • I think community builders should have a solid and detailed enough understanding of EA received wisdom to be able to lay out the case for our recommendations in a reasonably credible way, but I don’t think it’s reasonable to expect them to be domain experts in every domain, and that means that sometimes they aren’t going to be able to seem impressive to every domain expert that comes to us.

  • To be frank, it isn’t realistic to be able to capture the imagination of everyone who seems promising even if we make the best possible versions of our arguments. Some people will inevitably come away thinking we “just don’t get it”, that we haven’t addressed their objections, that we’re not serious about [specific concern X] and therefore our point of view is uninteresting. Communication channels just aren’t high-fidelity enough, and people’s engagement heuristics aren’t precise enough, to avoid this happening from time to time.

  • When some people are weirded out by the way we behave or try to attract new members, it seems to me like sometimes this is just reasonable self-protective heuristics that they have, working exactly as intended. People are creeped out by us giving them free books or telling them to change their careers or telling them that the future of humanity is at stake, because they reason “these people are putting a lot into me because they want a lot out of me”. They’re basically correct about that! While we value contributions from people at a wide range of levels of engagement and dedication, the “top end” is pretty extreme, as it should be, and some people are going to notice that and be worried about it. We can work to reduce that tension, but I don’t think it’s going away.

  Obviously we should try our best on all of these dimensions, progress can be made, we can be more impressive and more appealing and less threatening and more welcoming. But I can’t imagine a realistic version of the EA community that honestly communicates about everything we believe and want to do and doesn’t alienate anyone by doing that.

• Ben Millwood 6 Feb 2023 1:54 UTC

  64 points

  28 ∶ 1

  on: EA’s weird­ness makes it un­usu­ally sus­cep­ti­ble to bad behavior

  This post presents as the only counterargument to expelling weirdness from our community that we might have less good ideas about how to make the world better. Obviously a big and serious concern! Seems important to include it. But I didn’t hear two other important reasons (in my view) to tolerate weirdness:

  • Dialling down weirdness is often difficult and stressful, and in some cases effectively amounts to excluding people entirely (whether intentionally or not),

  • Cultural designation of harmless lifestyles, or beliefs, or neurodivergence, as weird or inappropriate is immoral, and we should not tolerate it lightly.

  I agree you shouldn’t do weird bad things, but you shouldn’t do ordinary bad things either. A general campaign against weirdness will hurt people who do not deserve to be hurt, and that’s wrong regardless of whether it will overall be better or worse for the prospects of the movement as a whole.

  When the culture you live in is (e.g.) homophobic, or transphobic, or stigmatizes autism, being counter-cultural is a moral obligation.

• Ben Millwood 18 Jun 2024 16:04 UTC

  63 points

  20 ∶ 5

  on: Why so many “racists” at Man­i­fest?

  This is a bit self-indulgent of me, but I’m going to quote myself from the comments of the other post, because I think it’s relevant here too:

  One aspect of the framing here that annoyed me, both in the OP and in some of the comments: the problem is not controversial beliefs, it is exclusionary beliefs. Here are some controversial beliefs that I think would pose absolutely no problem at this event or any other:

  • The many-worlds interpretation of quantum mechanics.

  • Virus gain-of-function research creates more risk than it prevents.

  • Nuclear energy is a necessary part of the transition away from fossil fuels.

  The problem with racism and transphobia is not that people disagree about them! The problem is that these beliefs, in their content on the object level, hurt people and exclude people from the discussion.

  Let’s avoid using “controversial” as a euphemism for “toxic and exclusionary”. Let’s celebrate the debate and discussion of all controversies that threaten no-one and exclude no-one. Suggesting any of that is at stake is totally unnecessary.

  (To the extent you want to upvote /​ downvote this sentiment, maybe go do that on the original comment and not this one, unless that feels inappropriate somehow? idk, maybe I’m overthinking this; I removed my own upvote from this post)

• Ben Millwood 14 Feb 2022 1:13 UTC

  57 points

  4 ∶ 0

  on: Sim­plify EA Pitches to “Holy Shit, X-Risk”

  My main criticism of this post is that it seems to implicitly suggest that “the core action relevant points of EA” are “work on AI or bio”, and doesn’t seem to acknowledge that a lot of people don’t have that as their bottom line. I think it’s reasonable to believe that they’re wrong and you’re right, but:

  • I think there’s a lot that goes into deciding which people are correct on this, and only saying “AI x-risk and bio x-risk are really important” is missing a bunch of stuff that feels pretty essential to my beliefs that x-risk is the best thing to work on,

  • this post seems to frame your pitch as “the new EA pitch”, and it’s weird to me to omit from your framing that lots of people that I consider EAs are kind of left out in the cold by it.

• Ben Millwood 18 Jun 2024 11:07 UTC

  55 points

  15 ∶ 6

  in reply to: Habryka’s comment on: My ex­pe­rience at the con­tro­ver­sial Man­i­fest 2024

  I don’t think they were trying in any way to “cultivate a home for racists”.

  I think one way you can read this situation is: racists are looking for an “intellectual home” in some sense, and since they don’t find one in most of the mainstream, they look for places that they can parasitically occupy and use for their own ends. The warning here is: the forecasting community need not only to avoid cultivating a home for racists, but also to proactively defend against racists cultivating a home for themselves. And if the forecasting community can’t build walls against this kind of parasitism, then the rationalist community needs to protect themselves from the forecasting community. And if they can’t do that, then EA needs to protect itself from the rationalist community.

  The core of much of this thinking is that racists (and fascists, the alt-right generally) don’t play fair in the marketplace of ideas, and they will manipulate and exploit your welcome if you extend them one. I’m not sure how well I can defend this idea (might write a top-level comment about it if I can feel confident enough about it), but I think that’s often what people are getting at with these kinds of concerns.

• Ben Millwood 18 Feb 2023 18:28 UTC

  51 points

  16 ∶ 3

  on: EA, 30 + 14 Rapes, and My Not-So-Good Ex­pe­rience with EA.

  Hi there,

  I have been a funder of both CEA and Rethink Priorities in the past, to the tune of low 5-digit sums per year. I’m personally acquainted with Julia, Chana and Peter, and I trust their judgement to an extent, so I tend to believe that they have good reasons for their actions, even if I don’t know what they are yet.

  Despite this, if the basic claims you’re making (as I understand them) are true:

  • you’ve been instrumental in getting a double-digit number of abusers excluded from the community over the last 6 years or so,

  • this is a relatively uncontroversial claim that Julia et al would straightforwardly agree with,

  then I’d be willing to fund you to continue this work, even if CEA didn’t want to pay you themselves. Given your credentials and background, I don’t know if I could afford you by myself, but if not I imagine there might be other people interested in contributing too.

  I don’t want to sound like I’m promising anything. I’m generally nervous of being over-eager to fund things without fully understanding the situation and all perspectives on it. I think there’s a reasonable chance that when I ask the CH team about this they have very good reasons for why they haven’t had a better relationship with you, and perhaps there could be some reason that funding you to continue this work would be counterproductive somehow.

  I also understand that the fact you’re not paid to do this work is not your only complaint, and maybe not even in your top three complaints. But it is the complaint that seems easiest to address, especially as a comparative outsider.

Can the AI af­ford to wait?

• Ben Millwood 28 Mar 2024 18:54 UTC

  48 points

  13 ∶ 3

  in reply to: Jason’s comment on: Re­post: SBF sen­tenced to 25 years jail

  While I see what you’re saying here, I prefer evil to be done inconsistently rather than consistently, and every time someone merely gets what they deserve instead of what some unhinged penal system (whether in the US or elsewhere) thinks they deserve seems like a good thing to me.

  (I don’t personally have an opinion on what SBF actually deserves.)

• Ben Millwood 16 Jun 2024 15:49 UTC

  45 points

  22 ∶ 1

  on: Vlog Brothers (Hank and John Green) are pro­mot­ing re­duced beef consumption

  I like the spirit of the video, but my impression is that if this causes people to switch from beef to chicken, it’ll be good for climate change but bad for animal welfare, since chickens have much worse lives than farmed cows. I haven’t studied how much reduced beef consumption prompts increased chicken consumption, but my first guess would be that it’s enough to make the climate benefits not worth it (especially in comparison to other smaller-downside methods of fighting climate change).

• Ben Millwood 17 Dec 2023 13:08 UTC

  43 points

  14 ∶ 2

  on: BenMill­wood’s Shortform

  People often propose HR departments as antidotes to some of the harm that’s done by inappropriate working practices in EA. The usual response is that small organisations often have quite informal HR arrangements even outside of EA, which does seem kinda true.

  Another response is that it sometimes seems like people have an overly rosy picture of HR departments. If your corporate culture sucks then your HR department will defend and uphold your sucky corporate culture. Abusive employers will use their HR departments as an instrument of their abuse.

  Perhaps the idea is to bring more mainstream HR practices or expertise into EA employers, rather than merely going through the motions of creating the department. But I think mainstream HR comes primarily from the private sector and is primarily about protecting the employer, often against the employee. They often cast themselves in a role of being there to help you, but a common piece of folk wisdom is “HR is not your friend”. I think frankly that a lot of mainstream HR culture is at worst dishonest and manipulative, and I’d be really sad to see us uncritically importing more of that.

• Ben Millwood 27 Aug 2019 11:52 UTC

  42 points

  0 ∶ 0

  on: Move­ment Col­lapse Scenarios

  This is a question I consider crucial in evaluating the work of organizations, so it’s sort of embarrassing I’ve never really tried to apply it to the community as a whole. Thanks for bringing that to light.

  I think one thing uniting all your collapse scenarios is that they’re gradual. I wonder how much damage could be done to EA by a relatively sudden catastrophe, or perhaps a short-ish series of catastrophes. A collapse in community trust could be a big deal: say there was a fraud or embezzlement scandal at CEA, OPP, or GiveWell. I’m not sure that would be catastrophic by itself, but perhaps if several of the organizations were damaged at once it would make people skeptical about the wisdom of reforming around any new centre, which would make it much harder to co-ordinate.

  Another thing that I see as a potential risk is high-level institutions having a pattern of low-key misbehaviour that people start to see (wrongly, I hope) as an inevitable consequence of the underlying ideas. Suppose the popular perception starts to be “thinking about effectiveness in charity is all well and good, but it inevitably leads down a road of voluntary extinction /​ techno-utopianism /​ eugenics /​ something else low-status or bad”. Depending on how bad the thing is, smart thoughtful people might start self-selecting out of the movement, and the remainder might mismanage perceptions of them even worse.

• Ben Millwood 18 Jun 2024 11:12 UTC

  40 points

  6 ∶ 0

  in reply to: Barry Cotter’s comment on: My ex­pe­rience at the con­tro­ver­sial Man­i­fest 2024

  What principles unite effective altruism?

  [...]

  1. Open truthseeking: Rather than starting with a commitment to a certain cause, community or approach, it’s important to consider many different ways to help and seek to find the best ones. This means putting serious time into deliberation and reflection on one’s beliefs, being constantly open and curious for new evidence and arguments, and being ready to change one’s views quite radically.

  I think any other source of EA principles you can find will say the same thing.

When “hu­man-level” is the wrong thresh­old for AI

• Ben Millwood 21 May 2024 8:33 UTC

  37 points

  2 ∶ 0

  on: BenMill­wood’s Shortform

  I wonder how the recent turn for the worse at OpenAI should make us feel about e.g. Anthropic and Conjecture and other organizations with a similar structure, or whether we should change our behaviour towards those orgs.

  • How much do we think that OpenAI’s problems are idiosyncratic vs. structural? If e.g. Sam Altman is the problem, we can still feel good about peer organisations. If instead weighing investor concerns and safety concerns is the root of the problem, we should be worried about whether peer organizations are going to be pushed down the same path sooner or later.

  • Are there any concerns we have with OpenAI that we should be taking this opportunity to put to its peers as well? For example, have peers been publically asked if they use non-disparagement agreements? I can imagine a situation where another org has really just never thought to use them, and we can use this occasion to encourage them to turn that into a public commitment.

• Ben Millwood 6 Mar 2023 13:01 UTC

  37 points

  14 ∶ 0

  in reply to: Habryka’s comment on: Nick Bostrom should step down as Direc­tor of FHI

  Given the problems that FHI has run into here, do you think it’s likely to continue to do good work? My guess is that being able to hire is crucial to that work. Given that, we have to ask ourselves: how long should we wait for Nick to find a resolution to these problems before we conclude that he’s unable to solve them? My guess is that two years into a hiring freeze is long enough.

  I agree with you that we should update somewhat in the direction of cooperating with universities being difficult and costly. But we should also entertain the hypothesis that it’s good to collaborate with universities, but it’s not good to ask Nick to be the one who does it.