I can’t speak to sub-Saharan Africa, but in India the government have put together a very comprehensive survey of mental wellbeing, which they ran in 2016 and are running again now. They randomly sampled households across the country, and asked people about their feelings and symptoms, which they then categorised into different mental conditions, rather than asking directly. This found results that largely concurred with previous GBD estimates for India. But I can’t say for sure whether there might be other biases, and there are almost certainly issues with the GBD’s Africa estimates due to a lack of ground truth data in many regions and methodological issues.

Here’s a paper on the kinds of surveys and interviews used to form the GBD prevalence estimates, and here’s one arguing that prevalence is sharply underestimated.

Not particularly interested right now, but maybe in the future :)

I will take a note—we are trying to do more comms in 2026 and this could be a great thing to post about. If we do, I’ll reach out :)

5% of Americans identify as being on the far left

However, I would strongly wager that the majority of this sample does not believe in the three ideological points you outlined around authoritarianism, terrorist attacks, and Stalin & Mao (I think it is also quite unlikely that the people viewing the Tik Tok in question would believe these things either). Those latter beliefs are extremely fringe.

My hobby horse around these parts has been that EA should be less scared about reaching out to the left (where I’m politically rooted), and thinking about what commonalities we have. This is something I have already seen in the animal welfare movement, where EAs are unafraid to work with existing vegan activism, and have done a good job of selling philanthropic funding to them, despite having large differences in opinion on the margins.

As you note, it’s not unreasonable that EA looks very far left from some perspectives. GiveDirectly is about direct empowerment, and I would argue that a lot of global development work, especially economic development, can be anti-imperial and generally concord with Marxist ideas of the internationale. Some better outreach and PR management in these communities would go a long way in the same way that it has for the political centre-left, who seem to get lots more attention from EA.

I strongly agree, and would add that this is a big concern of mine in direct intervention delivery also. Kaya Guides is fortunate enough to have a team that understands digital marketing, and we decided early on to recruit participants for our intervention using Meta ads. When I joined, I implemented direct conversion tracking from our intervention. Combined, we have reduced recruitment costs to around US$1, which is substantially cheaper than forming partnerships in the early years of an organisation, and is much more flexible.

I have been advocating for it where I can within AIM, for interventions that support it (and sometimes go a step further, and try to advocate for selecting interventions that are well suited for digital delivery).

Let me know if there are ways I can help advocate for better growth marketing within EA interventions, I am very passionate about this!

A few points:

1. There is still a lot of progress to be made in low-income country psychotherapy, which I think many EAs find counterintuitive. StrongMinds and Friendship Bench could both be about 5× cheaper, and have found ways to get substantially cheaper every year for the past half decade or so. At Kaya Guides, we’re exploring further improvements and should share more soon.
2. Plausibly, you could double cost-effectiveness again if it were possible to replace human counsellors with AI in a way that maintained retention (the jury is still out here).
3. The Happier Lives Institute has been looking at these kinds of interventions; their Promising Charities Pure Earth and Taimaka both appear to improve long-run mental health sustainably, by treating lead poisoning and malnutrition.

I am surprised at this, only because I remember the Gulf states were quite keen on bringing production into their countries and I would’ve thought they’d have declared Halal sooner!

(Yep, I’m not having a go at the mission here, more at the nuances of measurement)

Small drive-by question for you: In your opinion, if C. Elegans is conscious and has some moral significance, and suppose we could hypothetically train artificial neural networks to simulate a C. Elegans, would the resulting simulation have moral significance?

If so, what other consequences flow from this—do image recognition networks running on my phone have moral significance? Do LLMs? Are we already torturing billions of digital minds?

If not, what special sauce does C. Elegans have that an artificial neural network does not? (If you’re not sure, where do you think it might lie?)

(Asking out of genuine curiosity—haven’t had a lot of time to interface with this stuff)