To be clear, I was a lot more surprised when I was told about some of what OpenPhil did in DC, once starting to facepalm really hard after two sentences and continuing to facepalm very hard for most of a ten-minute-long story. It was so obviously dumb, that even me, with basically zero exposure to American politics or local DC norms and only some tangential experience running political campaigns in a very different context (an authoritarian country), immediately recognized it as obviously very stupid. While listening, I couldn’t think of better explanations than stuff like “maybe Dustin wanted x and OpenPhil didn’t have a way to push back on it”. But not having anyone who could point out how this would be very, very stupid, on the team, is a perfect explanation for the previous cringe over their actions; and it’s also incredibly incompetent, on the level I did not expect.
As Jason correctly noted, it’s not about “policy”. This is very different from writing papers and figuring out what a good policy should be. It is about advocacy: getting a small number of relevant people to make decisions that lead to the implementation of your preferred policies. OpenPhil’s goals are not papers; and some of the moves they’ve made that their impact their utility more than any of the papers they’ve funded more are ridiculously bad.
A smart enough person could figure it out from the first principles, with no experience, or by looking at stuff like how climate change became polarized, but for most people, it’s a set of intuitions, skills, knowledge that are very separate from those that make you a good evaluator of research grants.
It is absolutely obvious to me that someone experienced in advocacy should get to give feedback on a lot of decisions that you plan to make, including because some of them can have strategic implications you didn’t think about.
Instead, OpenPhil are a bunch of individuals who apparently often don’t know the right questions to ask even despite their employer’s magic of everyone wanting to answer their questions.
(I disagree with Jason on how transparent grant evaluations are ought to be; if you’re bottlenecked by time, it seems fine to make handwavy bets. You just need people who are good of making bets. The issue is that they’re not selected for making good bets in politics, and so they fuck up; not with the general idea of having people who make bets.)
Yep, it doesn't give us P(x | do(event that involves randomness)). The two are indeed not equivalent.
(For the context, all of these are entirely AI-generated, and I have not submitted them to MIRI)
I want to signal-boost this LW post.
I long wondered why OpenPhil made so many obvious mistakes in the policy space. That level of incompetence just did not make any sense.
I did not expect this to be the explanation:
THEY SIMPLY DID NOT HAVE ANYONE WITH ANY POLITICAL EXPERIENCE ON THE TEAM until hiring one person in April 2025.
This is, like, insane. Not what I'd expect at all from any org that attempts to be competent.
(openphil, can you please hire some cracked lobbyists to help you evaluate grants? This is, like, not quite an instance of Graham's Design Paradox, because instead of trying to evaluate grants you know nothing about, you can actually hire people with credentials you can evaluate, who'd then evaluate the grants. thank you <3)
2i made a discord server if people want to collaborate on designing ads for If Anyone Builds It! Including sharing ideas that are not finished designs, etc.
i think we can just agree to donate the prizes to MIRI (or maybe someone proposes some attribution mechanism, but it seems great if people just laser-focus on making an awesome thing for the sake of it and celebrate the results and everyone who contributed).
Whoever can set up a shared workspace, please also do that!
A good idea might be to come up with, think a bit about, & share various:
- things (ideas, vibes, thoughts, information, calls to action) the ads might try to communicate
- design ideas/styles/etc.
- specific features that can be used in the ads
- etc.
One way this can be bad is that people might anchor on existing ideas too much, but I wouldn't be too worried, people normally have lots of very personal preferences and creative ideas and styles, and iteration on something in collaboration might be great.
i really like the general idea and the two square metro versions!
Thanks!
What are the approximate % of “sell more books”, “show what the current situation in AI is”, “make this the Overton window”, etc. in what you want the ads to optimize for?
PSA: if you're looking for a name for your project, most interesting .ml domains are probably available for $10, because the mainstream registrars don't support the TLD.
I bought over 170 .ml domains, including anthropic.ml (redirects to the Fooming Shoggoths song), closed.ml & evil.ml (redirect to OpenAI Files), interpretability.ml, lens.ml, evals.ml, and many others (I'm happy to donate them to AI safety projects).
1some ideas for inspiration
more
Everyone should do more fun stuff![1]
I thought it'd just be very fun to develop a new sense.
Remember vibrating belts and ankle bracelets that made you have a sense of the direction of north? (1, 2)
I made some LLMs make me an iOS app that does this! Except the sense doesn't go away the moment you stop the app!
I am pretty happy about it! I can tell where’s north and became much better at navigating and relating different parts of the (actual) territory in my map. Previously, I would remember my paths as collections of local movements (there, I turn left); now, I generally know where places are, and Google Maps feel much more connected to the territory.
If you want to try it, it's on TestFlight: https://testflight.apple.com/join/kKKfMuDq
It can vibrate when you face north; even better, if you're in headphones, it can give you spatial sounds coming from north; better still, a second before playing a sound coming from north, it can play a non-directional cue sound to make you anticipate the north sound and learn very quickly.
None of this interferes with listening to any other kind of audio.
It’s all probably less relevant to the US, as your roads are in a grid anyway; great for London though.
If you know how to make it have more pleasant sounds, or optimize directional sounds (make realistic binaural audio), or make react-native do nice vibrations when the app is in the background instead of bzzzz, and want to help, please do! The source code is on GitHub: https://github.com/mihonarium/sonic-compass/
unless it would take too much time, especially given the short timelines