Why are OpenAI and apparently NVIDIA employees good at collectively mobilising to get their CEO back, but Anthropic employees not good at collectively mobilising to stop development of a technology they believe could kill 8 billion people?

It’s a delight to read something so thoughtful about a topic I knew nothing about, thank you :)

Cheers Tyler, and thanks for reviving your account to speak on it! Some of this stuff feels esoteric now, but I also think it sheds light on the lineage of modern EA, especially the split on the issue between the US and UK crowds (roughly).

I think this connects the dots on the story a bit, so thank you!

This is a piece worth a lot of reflection, and I really admire the steps you’re taking!

I worry also about culture and individual role modelling might not be enough to prevent the worse forms of status-seeking & greed. I don’t think that a norm of whistleblowing policies and independent boards would’ve stopped SBF—I mean, we know that Tara Mac Aulay quit Alameda way back in 2018 and warned others about it.

I think people with power in this space (funders, evaluators, incubators, implementers) need to be on the lookout for charlatans. We should narrow down which inputs a bad actor could manipulate to access our power: For a funder, this is in evaluation, due diligence, etc. For an implementer, this might be in your hiring process. Validate your assumptions and lock these things down. ‘Assume good faith’ is a great rule to have when your movement is small—‘Trust, but verify’ needs to replace it.

We also need to acknowledge—and I don’t believe this was done well after the FTX collapse—that bad actors do not have to have ill intent. I think it might be just as harmful to exaggerate the impact of your programme, or to fail to run M&E on it appropriately, whether you are deliberately concealing and defrauding donors, or just being conveniently neglectful. There’s a selection bias here with a long tail; for every outright fraudster losing billions, there can be hundreds of little exaggerations and not-being-completely-upfronts costing millions each.

When I feel the tug of these incentives myself, I’m often a little surprised that there aren’t a lot of checks and balances on these kinds of deceptions within EA. I hope that with some amount of the money that comes into the space, that we set much higher standards for ourselves and independently hold others to account.

Has anyone considered engaging with organised labour on this? They have a lot of experience in this area + often have special legal protections against retaliation for members only. Seems like a useful way into putting on the brakes from within frontier labs.

Nice, genuinely thank you for following up!

I’m reflecting on how we’ve had a lot of outsider orgs join the Forum lately (ex. Sanku) and wondering if Marginal Funding Week might get quite crowded. It was pretty hard to get above the fold last year.

Perhaps maybe (1) a pre-agreed posting schedule and/or (2) spreading it over 2 weeks could be good? Very unsure, but it crossed my mind as worth raising

Always welcome Joel!

• I’m generally pro-YODA (although I think the name is perhaps a little unserious?). I think it’s very legible to people who care about mental health, or wellbeing more generally, and I like that it’s on the same order of magnitude as a WELLBY so the conversions are easy to do in one’s head. I think DALYs are probably still going to be the only way people in EA global health will care about psychotherapy, however.
• I think if anything, I’ve approached the Cuijpers paper from the mechanisms first. In my personal experience, I’ve used things I’ve learned in therapy over and over, (but I’m not chronically depressed). It seems like basic skills (boundary setting, emotional regulation, behavioural activation) could be like learning to ride a bicycle.
  • My mental model of severe mental illness is that you have little control (without pharmacology) over the weather in your mind, but you do have some ability to regulate your response to those feelings.
  • So to me the Cuijpers paper reads as not disconfirming the belief I already held
• On treating the young: My best guess is this would make sense after a certain age, but it’s unclear what the age is. There’s decent evidence to suggest conventional psychotherapy doesn’t work on children, last I checked (memory may be completely off here).
• I read through the paper and although they try some different fits for the cubic spline, they never explain this choice as opposed to other modelling methods
• They did test regions, and found no effect. But the ‘Other’ region includes the Iran papers HLI have discussed at length before.
• There’s no moderation for whether people pursued therapy after the trials. I guess, in fairness, this is more of a question of increasing costs in a CEA, since the continued care would be counterfactual.

Yeah, anyway, I think the paper basically raises the need to investigate further.

Thank you! I’ll get back to your email.

My sense is that cyclical data like this should come out in the wash, but just require more statistical power. It would depend on the amplitude of the menstrual cycles, if it’s only a couple of PHQ-9 points per person then it should be fine. I mean, there isn’t any evidence showing men and women have differential improvements from psychotherapy, which you’d expect if this was having a large impact.

No worries—it looks like a fantastic tool and I’m really excited about it! Thanks for all the thought and effort that went into it