Geoffrey Miller's Quick takes

Just on whether this is the most serious PR crisis in EA history: it might be, but the most accurate measure of salience / interest I have is Wikipedia pageviews, where it seems smaller than the OpenAI board drama, and much smaller than FTX.

https://x.com/bellaforristal/status/2102305645328208316

I definitely not the "worst" PR crisis. The DOW account is a random poster for the Office of the Under Secretary of War for Research and Engineering, not Hegseth. 

But I agree that it would be best for Republicans not to be polarized against AI safety. They aren't so far; thanks in part to Speaker Johnson. I suggest not letting random trolls goad us into abandoning courtesy and the presumption of good will. It is counterproductive to respond to much of this stuff, as that imbues it with more seriousness than it deserves. 

To be fair, MAGA is basically the antithesis of all that Effective Altruism stands for, so it would not be surprising that we would be natural enemies and this would eventually happen.

MAGA is about contracting the moral circle of concern to "real Americans". This is literally the opposite direction as EAs wanting to expand the moral circle to animals. MAGA is flagrantly anti-intellectual. A stereotypical example of a MAGA is a gun-loving hunter who shoots animals for sport and brags about being a red-meat eating tough guy.

Like, the socialists disagree with us about the methods (we advocate individual action like charity and career change, while they advocate systemic change) but ultimately we share the ends of equality and wanting a better future. We compete in terms of recruiting for the idealistic altruists in the world. We both tend towards internationalism and some form of the common good for all.

MAGA doesn't share any of that. They harken to past national glory, see the people we donate bednets to as subhuman, see the elites we listen to as inherently untrustworthy, denigrate empathy and glorify egoism, etc. I just don't see why we should expect anything else from them than pure disdain.

In theory, AI x-risk is more important than all of that, and we do have a single common cause in not wanting to be dead, but I see little else we can even attempt to agree on.

I think I agree with the assessment, but what proposals have you thought about to stem this tide?

Trump tweeting remarks is less of a PR crisis than FTX collapsing was. Trump says wild stuff constantly. Are we actually gonna invade Greenland? Probably not.

It is for sure sad that this is likely going to increase the polarization of AI Safety.

But its not a scandal, and FTX very much was.

I agree this moment presents an opportunity for some truth-tracking comms, and I'm sad that the administration isn't into AI safety, but I disagree with your vibe. EAs don't need to "defend" EA on social media. They can just go on doing good. Winning hearts and minds is nice but not super important; I feel unthreatened by the "trash-talk." And the truth is convergent; just doing good is a great way to improve your reputation in the long run.

And yet on Bluesky I’m getting threads trashing EA as doing the bidding of Elon Musk and big tech, being a cult, and not caring about poor people. (Threads citing this post: https://www.realtimetechpocalypse.com/p/effective-altruism-is-a-dangerous )

Sigh

I'm not sure what a reasonable response to this would be. I can see engaging with conservatives and even maybe with people who identify as MAGA.

At the risk of stating the obvious/strawman, I think interacting with the famous tweets and threads coming from Trump and Hegseth seems unlikely to be fruitful.

I think it's only a slight exaggeration to say that Trump basically does not respond to evidence or logic, and they tend to lash out vindictively against criticism. The only approach that seems to reliably get Trump on your side is to give him and his family money or to praise him.

I guess someone could make the case that those latter approaches would be the right way forward, but they have obvious downsides:

• Incompatible with EA and rationalist epistemics and approaches;
• Feeding a tiger that eventually eats you;
• Aligning AI safety very concretely with an extremely unpopular, controversial, and corrupt president would alienate reasonable people on the other side, and lead to a loss of trust and respect. (I think this point is the most important one.)

But perhaps this is also a straw man. I don't know if anyone here is advocating that AI safety advocates should donate to and praise Trump

It's not actually clear that we actually should respond. There's a lot of aura in deciding that attacks are beneath you and you're not going to respond to them.

This is definitely not going to help:

https://forum.effectivealtruism.org/posts/yPi9W46rurF7MPfTY/about-the-altruists-the-netflix-series-on-ftx

One thing I notice is that a lot of people are talking about "we" as members of the "EA community", but these people seem to have very different interpretations of what that means.  It seems EA has outgrown itself.

At this stage, for me personally, the EA label feels like a liability.  Membership to a club that I don't really want to be a part of.  I want to be altruistic, effectively.  But there are so many other things that seem to have become associated with EA that I really don't want to be.

I appreciate being part of this forum, though, because one characteristic of many EA individuals seems to be "thoughtful".  I can't help but think there might be others who lurking around EA that feel the same way.

(Not a solution, but a general observation about people who engage in bashing EA.)

The "dot connectors" will always connect the dots, infer or invent nefarious motivations, and try to bucket you as they like. The problem is that you can't neatly map EAs onto the political spectrum -- yes, there are dominant trends, but the variance in views is sufficiently high that commentators have genuinely no clue where EAs belong. This makes sense because most major movements in history have been political ones, so when assessing EA, most people pull out their internal political philosophy detector and you end up with a mess like the chart below! 

But EA is a moral philosophy movement, and the chain of thinking is genuinely different. Instead of thinking how to organize society and labor, EAs unanimously agree on beneficentrism and deal with questions like, "What morally matters? To what degree? Which interventions are most effective? How do you even assess what is most effective?" When you organize a movement around these set of questions, you end up with:

• Some people who want to automate software engineering, some who want to pause it entirely, and others who think we should defensively accelerate progress
• At least two frontier AI labs: let's not forget OpenAI received $30 million in philanthropic money during its inception!
• Some EAs who think that AI will be a big deal for {their cause area}, others who are skeptical of the whole AI bundle
• Some EAs passionately dislike AI writing, some are fine with methodical use of AI in writing, and some are even more liberal about it
• One particular EA who is the loudest voice combatting the data center water usage myth
• (At least) one person from the EA-sphere who has large holdings in AI infrastructure
• And conservative AI Safetyists like you and liberal long timeline accelerationists like me

I don't know what the best solution for combatting EA bashing is, but spreading the idea that EA is more politically and intellectually diverse than people think should help. 

This is  a slow-burn solution, but the most effective support and rebuttals will come from people who aren't EAs, but are just fair/principled, and have had enough exposure to EA to know when attacks are unfair. E.g. See Dean Ball this week. So the more surface area EAs can create with those sorts of people, the better the position EA is in. For example, I think Andy Masley's datacenter water use posts created a lot of surface ara with such people and has been better for the EA 'brand' than any specific rebuttal.

(A part of this strategy involves, as a general principle, "behave with as much dignity, integrity and fairness as possible, even when others aren't). (Admittedly my own responses usually involve gently poking fun, but I do try to stay good natured).

I think we could use a documentary series where we just go follow around orgs or individual EAs for a couple days and see how they talk, live and act. It would be pretty cheap at the very least. 

You mention having “an ambition, even a prayer for [the AI developers]” (~12 min). You might mean this figuratively, but many viewers of that channel will probably take it literally. When you admit elsewhere that you don't believe in god and don't practice any religion, they likely see this as a contradiction and suspect you’re not being genuine once you become more popular.

My guess is that it's better to be upfront about not being a Christian in order to retain authenticity. You'll probably always be regarded as an outsider by the conservative right (just for another example: your twitter handle includes 'poly', and you've spoken at length online about being polyamorous), but you could hope to be perceived as 'the outsider who gets us'. This kinda worked for a while for Sam Harris or Milo Yiannopoulos.

I haven't seen the series, but am currently halfway through the second book.

I think it really depends on the person. The person I imagine would watch three-body problem, get hooked, and subsequently ponder about how it relates to the real world, seems like they also would get hooked by just getting sent a good lesswrong post?

But sure, if someone mentioned to me they watched and liked the series and they don't know about EA already, I think it could be a great way to start a conversation about EA and Longtermism.

I completely agree Geoffrey! I originally read Liu Cixin's series before I became involved in EA, and would highly recommend to anyone who's reading this comment.

I think the series very much touches on themes similar in EA thought, such as existential risk, speciesism, and what it means to be moral.^[1]

I think what makes Cixin's work seem like it's got EA themes is that a lot of the series challenges how humanity views its place in the universe, and it challenges many assumptions about both what the universe is, and our moral obligations to others in that universe, which is quite similar to how EA challenges 'common-sense' views of the world and moral obligation.

1. ^{^}

   (I also referenced it in this reply to Matthew Barnett)

PS: Fun fact: after my coauthor Peter Todd (Indiana U.) and I read '3 Body Problem' novel in  2015, we were invited to a conference on 'active Messaging to Extraterrestrial Intelligence' ('active METI') at Arecibo radio telescope in Puerto Rico. Inspired by Liu Cixin's book, we gave a talk about the extreme risks of active METI, which we then wrote up as this journal paper, published in 2017:

PDF here

Journal link here

Title: The Evolutionary Psychology of Extraterrestrial Intelligence: Are There
Universal Adaptations in Search, Aversion, and Signaling?

Abstract
To understand the possible forms of extraterrestrial intelligence (ETI), we need not only astrobiology theories about how life evolves given habitable planets, but also evolutionary psychology theories about how intelligence emerges given life. Wherever intelligent organisms evolve, they are likely to face similar behavioral challenges in their physical and social worlds. The cognitive mechanisms that arise to meet these challenges may then be copied, repurposed, and shaped by further evolutionary selection to deal with more abstract, higher-level cognitive tasks such as conceptual reasoning, symbolic communication, and technological innovation, while retaining traces of the earlier adaptations for solving physical and social problems. These traces of evolutionary pathways may be leveraged to gain insight into the likely cognitive processes of ETIs. We demonstrate such analysis in the domain of search strategies and show its application in the domains of emotional aversions and social/sexual signaling. Knowing the likely evolutionary pathways to intelligence will help us to better search for and process any alien signals from the search for ETIs (SETI) and to assess the likely benefits, costs, and risks of humans actively messaging ETIs (METI).

The book in my opinion is better, and relies so much on vast realizations and plot twists that it's better to read it blind—before the series and before even the blurb at the back of the book! So for those who didn't know it was a book, here it is: https://www.amazon.fr/Three-Body-Problem-Cixin-Liu/dp/0765377063 

I didn't know about this, now I think I have a new netflix shot to watch! thanks!

On the topic, I hear season 7, episode 5 of young sheldon is abput a dangerous AI. Edit: I watched the episode, it's not.