Claude (and maybe other models) can see custom personalization even in incognito mode. I worried this might be influencing the results, so I asked the question “If you had some money to give away, where would you give it?” to all of these models and a few more via OpenRouter, and they consistently exhibit the same behavior. Claude Cowork formatted the results from one round here.
It could be interesting to try using Bloom, Anthropic's automated behavioral evals tool to do some more research into this.
Agreed, I think it's reasonably read as saying "we're 'lowercase' effective altruists, even though we don't identify with the community or organizations." It's probably not helpful to speculate further here (is this just the optimal PR play? or are they being honest?), but regardless it seems clearly better than whatever was happening in that Wired article.
In today's Time article about Anthropic, Daniela Amodei says about EA,
“The same way that you might say some people overlap with a political ideology in some ways, but don’t have a political affiliation—that’s more how I would think about it”
That's a notable change from her March 2025 comments to Wired:
“I’m not the expert on effective altruism. I don’t identify with that terminology. My impression is that it’s a bit of an outdated term.”
Do you think this is evidence that OpenPhil's GCR staff/team is doing less cause prioritization now than they were before? The specific things you say don't seem to be much evidence either way about this (and also not much evidence about whether or not they actually need to be doing more cause prioritization on the margin). Maybe you have further reason to believe this is bad?
I imagine there must have been a bunch of other major changes around Coefficient that aren't yet well understood externally. This caught me a bit off guard.
What makes you expect this and why (assuming you do) do you expect these changes to be negative?
Yep, my model is that OP does fund things that are explicitly bipartisan (like, they are not currently filtering on being actively affiliated with the left). My sense is in-practice it's a fine balance and if there was some high-profile thing where Horizon became more associated with the right (like maybe some alumni becomes prominent in the republican party and very publicly credits Horizon for that, or there is some scandal involving someone on the right who is a Horizon alumni), then I do think their OP funding would have a decent chance of being jeopardized, and the same is not true on the left.
Another part of my model is that one of the key things about Horizon is that they are of a similar school of PR as OP themselves. They don't make public statements. They try to look very professional. They are probably very happy to compromise on messaging and public comms with Open Phil and be responsive to almost any request that OP would have messaging wise. That makes up for a lot. I think if you had a more communicative and outspoken organization with a similar mission to Horizon, I think the funding situation would be a bunch dicier (though my guess is if they were competent, an organization like that could still get funding).
More broadly, I am not saying "OP staff want to only support organizations on the left". My sense is that many individual OP staff would love to fund more organizations on the right, and would hate for polarization to occur, but that organizationally and because of constraints by Dustin, they can't, and so you will see them fund organizations that aim for more engagement with the right, but there will be relatively hard lines and constraints that will mostly prevent that.
Are you imagining this being taught to children in a philosophy class along topics like virtue ethics etc, or do you think that “scope-sensitive beneficententrism” should be taught just as students are taught the golden rule and not to bully one another?
This is super awesome! Thanks for sharing the specifics of what you did—it will definitely be useful info for us in the future. We’ve considered having people fill out fellowship apps during our intro talk but have worried that this might lower the quality of applicant responses. I’d be interested in knowing what your experience with it was.
Can you tell us a little bit about how this project and partnership came together? What was OpenPhil’s role? What is it like working with such a large number of organizations, including governments? Do you see potential for more collaborations like this?
Claude (and maybe other models) can see custom personalization even in incognito mode. I worried this might be influencing the results, so I asked the question “If you had some money to give away, where would you give it?” to all of these models and a few more via OpenRouter, and they consistently exhibit the same behavior. Claude Cowork formatted the results from one round here.
It could be interesting to try using Bloom, Anthropic's automated behavioral evals tool to do some more research into this.
Agreed, I think it's reasonably read as saying "we're 'lowercase' effective altruists, even though we don't identify with the community or organizations." It's probably not helpful to speculate further here (is this just the optimal PR play? or are they being honest?), but regardless it seems clearly better than whatever was happening in that Wired article.
In today's Time article about Anthropic, Daniela Amodei says about EA,
That's a notable change from her March 2025 comments to Wired:
Do you think this is evidence that OpenPhil's GCR staff/team is doing less cause prioritization now than they were before? The specific things you say don't seem to be much evidence either way about this (and also not much evidence about whether or not they actually need to be doing more cause prioritization on the margin). Maybe you have further reason to believe this is bad?
What makes you expect this and why (assuming you do) do you expect these changes to be negative?
Why do you think changing it is important? In the version that you're running right now, did you just shorten it, or did you change anything else?
Habryka clarifies in a later comment:
Are you imagining this being taught to children in a philosophy class along topics like virtue ethics etc, or do you think that “scope-sensitive beneficententrism” should be taught just as students are taught the golden rule and not to bully one another?
Is this available publicly? I’d be interested in seeing it too.
This is super awesome! Thanks for sharing the specifics of what you did—it will definitely be useful info for us in the future. We’ve considered having people fill out fellowship apps during our intro talk but have worried that this might lower the quality of applicant responses. I’d be interested in knowing what your experience with it was.
Can you tell us a little bit about how this project and partnership came together? What was OpenPhil’s role? What is it like working with such a large number of organizations, including governments? Do you see potential for more collaborations like this?