colophonemes

The Long-Term Future Fund is for ensuring the continued survival and flourishing of future generations. Climate change is one future risk, but there are many others — including ensuring that the risks posed by smarter-than-human artificial intelligence are understood and mitigated, but also e.g. mitigating pandemic risk from genetically engineered pathogens.

As with all the Funds, choosing organisations to grant to will involve asking questions about scale, crowdedness, and tractability (how big is the problem, how many people are already working on the problem, and how hard is the problem to solve). For more on this see [1],[2].

Climate change is clearly an important future risk, but it's a very hard problem to solve (requiring massive international cooperation), and there are already lots of people working in this space (while I wouldn't necessarily agree that all future risks of climate change are 'very well understood' because of nonlinearities in modelling complex systems, part of the reason it's well-documented is because there has been considerable research effort expended over the last 20-30 years). By contrast, there are many future risks that have less attention, but may have similarly severe consequences for future generations. It's important we aren't blindsided by one of these risks just because we have a better understanding of one particular area.

To be clear, none of the above precludes that the Fund could donate to orgs working on climate change research/mitigation/prevention, just that there might be reasons to look at other risks as well, even if they're less well understood.

For more on research into effective giving opportunities in climate change, see https://www.givingwhatwecan.org/cause/climate-change/

[1] https://www.youtube.com/watch?v=67oL0ANDh5Y [2] https://80000hours.org/2014/01/which-cause-is-most-effective...

Yeah, I agree that it's much easier to get on board with the fundamental proposition of EA if you're of a consequentialist disposition, because the 'most' in 'do the most good' implies a maximising view. I don't think that it's inherently antithetical to other value systems, but agree that because there are more consequentialists, it's more culturally consequentialist.

The reason we publish fund manager grant history/writeups is so that you can have a sense of what their values are, and can make some calls about whether that accords with your views. Without presuming to speak for him or pre-empt any decisions, I strongly suspect that Lewis is unlikely to grant to anything on the more speculative/controversial side of animal welfare (in general I think it's more likely to focus on corporate cage-free programs and meat replacement tech). We think there are a lot of good reasons not to use the Funds[1], and if you're worried that you're going to end up funding something that's harmful, you shouldn't donate to that Fund.

[1] https://app.effectivealtruism.org/funds/faq?tag=why-use-ea-f...

Good point — a nice feature but not a point of differentiation from Givewell's donation page... I think the others still stand – in general you gain additional flexibility with the Funds without compromising effectiveness.

I'd add that these things will differ between cause areas, so it's hard to give a neat, roll-up answer. DALY's and NPV income estimates are fairly good in the global health space, but other cause areas may need to make more speculative comparisons because they deal with things that are harder to measure.

In all cases, when a Fund makes a grant, the fund manager will provide a write-up outlining their rationale. For now, each fund manager has a list of previous grants they've made, so you can get a sense of the organisations they are likely to grant to (though by no means does that imply that they necessarily /will/ grant to these specific orgs in future).

One of the key advantages of the Funds model is not locking you in to a particular charity up front, which means that as more information about funding gaps, intervention effectiveness etc comes to light, fund managers can respond accordingly.

While there are people who identify as effective altruist who hold views that are controversial, I think it's important to not paint with such a broad brush. As your linked article notes, there is wide disagreement on a range of thorny philosophical issues across a range of cause areas, but the scare quotes/selective quoting makes it seem like these views are unquestioningly accepted by a plurality of people in the community.

Effective altruism is a broad church. The unifying themes are trying to make the world a better place, using reason and evidence as tools for making decision, expanding our circle of compassion, and having epistemic humility (i.e. knowing that we could be wrong about things and being open to changing our minds in proportion to the strength of new evidence). The conclusions people draw can hinge on deep and ultimately irreducible value judgements — a strength of the community is that we can (in general) have these disagreements respectfully and work together productively where there is common ground.

Our baseline for effectiveness (specifically with regard to the Global Health and Development Fund, which is managed by GiveWell founder Elie Hassenfeld) is Givewell's top charities. So, it's entirely possible that Elie will choose to grant to these, in which case your donation will be exactly as effective as donating to Givewell, but you'll get the advantage of using the EA Funds platform — e.g. single tax receipt (donating to one org instead of 3-4), easy recurring donations, the ability to change your donation preferences later, allocating to causes outside of global health, donation tracking etc. etc.

However, we do think there's a good chance of getting higher returns. Givewell recommendations skew towards what makes sense for an average, individual donor with a particular risk profile. By having access to a larger pool of funds, you could seed new high-expected-value charities that wouldn't necessarily make it onto Givewell's recommendation list, but are nonetheless potentially high impact (examples of recent things in this space — New Incentives, Charity Science Health etc).

Thanks for the comment — we welcome feedback, even if it's harsh, so no worries!

That said, I think some of this hinges on a naïve interpretation of what effective altruism is about. The idea of 'earning to give' is counterintuitive ('do more good by working in finance - whut?'), and so has been one that the media has run with. Accordingly, I think it's considerably overrepresented in many people's minds compared to how most people in the EA community actually think about things[1]. We're always looking for talented people to do direct work, and we see one of our missions as finding and attracting people to work on important issues — whether that be in animal welfare, global development, politics, research etc. etc. Indeed, nearly everyone working on this project left significantly higher-paying jobs to come work at CEA because we think it offers the best chance for us to have a positive impact.

Agree that donating to large non-profits does have the problem of the money displacing unrestricted funds (in effect, subsidising other projects within the org). That's part of the reason most of the non-profits we end up supporting are fairly small and tightly focused on a specific, well-validated intervention. We're putting previous grant history up is so that donors can make an informed decision about whether or not they think a fund manager will represent their values, but it's not necessarily an indication of where grants will be made in future.

[1] E.g. see https://80000hours.org/2015/07/80000-hours-thinks-that-only-...

For sure, we want to represent a broad range of views within areas that are potentially high impact. We chose these as our initial funds partly because they seem like cause areas that reasonable people can disagree on (both on questions of values, and on empirical questions about relative risks and how to solve problems).

It's just that it's an estimate that hinges on a few judgement calls (e.g. whether or not to cost services that AMF receive pro bono) — we're in agreement with Givewell's estimate. For comparison, the WHO think that LLINs work out to be around ~$30/DALY and AMF run a pretty tight ship.

Given the tricky nature of cost-effectiveness estimates I don't think the numbers are hard enough that it'd be wise to talk about the difference implying anything as exact as an intervention being '20% more efficient'. It's more to get a reasonable ballpark for comparisons between other interventions.

(Also, seems there was a typo in the OP, should be DALY not QALY).

Thanks! Ideally this makes it significantly easier for effective non-profits to raise money. We're not locking in the set of organizations we fund. Being flexible about which cause you support at any given time, given the available facts is a core part of effective altruism.

We're especially excited about funding new nonprofits that seem like they're going to be able to have a big impact, because at the moment they can spend north of 30-40% of their time fundraising (which, especially when they're new, would be time better spent on validating their programs and scaling up).