Hey Brian,
We’re somewhat more keen to see additional resources on AI safety compared to GCBRs, but the difference seems fairly narrow, so we’re keen to see people take unusually good opportunities to help reduce GCBRs (or to work on it if they have better personal fit). More here: https://80000hours.org/problem-profiles/global-catastrophic-biological-risks/
Hey Brian,
Just a very quick answer from me to your first question.
At 80k we rate climate change ahead of global health since it seems more pressing from a longtermist perspective (e.g. Toby Ord thinks it’s a significant existential risk factor).
So, I would say that longtermists should donate to climate change over global health from an impact perspective, if choosing equally good charities from each cause (though I think it would be even better to donate to GCBRs of AI safety).
One might think that GiveWell is better at selecting charities than FP (and they’ve certainly done more research), but I think the edge on charity selection is unlikely to be big enough to offset the difference in cause area.
Another difference is that GiveWell focuses on evidence-backed interventions, whereas FP takes more of a hits based approach, but that seems like another advantage of the FP picks to me.
Finally, I’m focusing more on direct impact above. There could be other reasons to donate to global health (e.g. for advocacy reasons—since lots of great people have entered EA via global health), though I’m pretty unsure those factors would tell in favour of global health going forward (e.g. it seems plausible to me that EA should make climate change our standard ‘mainstream’ cause rather than global health).
PS Hauke’s post is comparing GiveWell recommendations to climate change on a neartermist perspective, so doesn’t answer your question.
I don’t think there’s a really an EA reason to pick a certain college. Just pick based on the normal considerations (e.g. where you’ll most enjoy living; where you think you fit with the culture / most fun; quality of tutorials & housing & funding; general reputation; academics https://en.wikipedia.org/wiki/Norrington_Table).
I did physics and philosophy and went to Balliol and was happy about that.
The main reason was that Balliol was clearly the phys phil college at that time, with ~5 people studying it each year, out of a total of ~15 in the university. It also had David Wallace, who was a great phys phil tutor (though he’s now left). I’d guess it still the biggest phys phil, but haven’t checked. It’s also well-known for PPE.
I think this was a decent reason to choose it—I appreciated having other phys phils to work with, and they tend to be an interesting bunch (and maybe some of the most naturally EA minded people out there!). I had lots of great tutors too.
I thought at the time that Balliol also does well on other factors: central location, OK housing (though not as nice as the wealthiest ones); culture a bit more lefty; decent academically; good reputation etc.
If I hadn’t gone to Balliol I might have gone for Wadham for the social life or one of prestigious ones with beautiful grounds (e.g. Merton, New, Magdalen, St Johns). New seems to have a good combination of features.
It’s true these ones are harder to get into than the others, but in Phys Phil they used to have a good ‘pool’ system, so if you don’t get into your first choice college, they’ll assign you to another one so long as you’re above the bar in general.
Yeah I totally agree there are useful things to say, though my impression is these kinds of changes are smaller and this kind of advice is more out there already (except the last one).
I think the hope for more radical changes would be giving people more time to mull over the worldview, and maybe introducing people to a general ‘prioritizy’ mindset, that can sometimes payoff a lot (e.g. thinking about what you really want to get out of college and making sure you do).
(On the specifics, I think maths & physics probably trumps economics at A-level, if someone has the option to do both. At undergrad it’s more unclear, but you can go from maths and physics into an econ, compsci or a bio PhD but not vice versa.)
Hi there,
If you’re looking for a wider range of job listings, you might find this list of social impact job boards useful.
I’m keen to see more experiments with impact certificates. Do you have funders interested in using it?
This will mainly need to wait for a separate article or podcast, since it’s a pretty complicated topic.
However, my quick impression is that the issues Caspar mentions are mentioned in the problem framework article.
I also agree that their effect is probably to narrow the difference between AI safety and climate change, however I don’t think they flip the ordering, and our ‘all considered’ view of the difference between the two was already narrower than a naive application of the INT framework implies – for the reasons mentioned here – so I don’t think it really alters our bottom lines (in part because we were already aware of these issues). I’m sorry, though, that we’re not clearer that our ‘all considered’ views are different from ‘naive INT’.
That’s an interesting point. I was thinking that most people would say that if my goal is X, and I achieve far less of X than I easily could have, then that would qualify as a ‘mistake’ in normal language. I also wondered whether another premise should be something very roughly like ‘maximising: it’s better to achieve more rather than less of my goal (if the costs are the same)’. I could see contrasting with some kind of alternative approach could be another good option.
I like the idea of thinking about it quantitatively like this.
I also agree with the second paragraph. One way of thinking about this is that if identifiability is high enough, it can offset low spread.
The importance of EA is proportional to the multiple of the degree to which the three premises hold.
Hi Paolo, I apologise this is just a hot take, but from quickly reading the article, my impression was that most of the objections apply more to what we could call the ‘near termist’ school of EA rather than the longtermist one (which is very happy to work on difficult-to-predict or quantify interventions). You seem to basically point this out at one point in the article. When it comes to the longtermist school, my impression is that the core disagreement is ultimately about how important/tractable/neglected it is to do grassroots work to change the political & economic system compared to something like AI alignment. I’m curious if you agree.
Hi Jamie,
I think it’s best to think about the importance of EA as a matter of degree. I briefly mention this in the post:
Moreover, we can say that it’s more of a mistake not to pursue the project of effective altruism the greater the degree to which each of the premises hold. For instance, the greater the degree of spread, the more you’re giving up by not searching (and same for the other two premises).
I agree that if there were only, say, 2x differences in the impact of actions, EA could still be very worthwhile. But it wouldn’t be as important as in a world where there are 100x differences. I talk about this a little more in the podcast.
I think ideally I’d reframe the whole argument to be about how important EA is rather than whether it’s important or not, but the phrasing gets tricky.
Hi David, just a very quick reply: I agree that if the first two premises were true, but the third were false, then EA would still be important in a sense, it’s just that everyone would already be doing EA, so we wouldn’t need a new movement to do it, and people wouldn’t increase their impact by learning about EA. I’m unsure about how best to handle this in the argument.
Hi Greg,
I agree that when introducing EA to someone for the first, it’s often better to lead with a “thick” version, and then bring in thin later.
(I should have maybe better clarified that my aim wasn’t to provide a new popular introduction, but rather to better clarify what “thin” EA actually is. I hope this will inform future popular intros to EA, but that involves a lot of extra steps.)
I also agree that many objections are about EA in practice rather than the ‘thin’ core ideas, and that it can be annoying to retreat back to thin EA, and that it’s often better to start by responding to the objections to thick. Still, I think it would be ideal if more people understood the thin/thick distinction (I could imagine more objections starting with “I agree we should try to find the highest-impact actions, but I disagree with the current priorities of the community because...), so I think it’s worth making some efforts in that direction.
Thanks for the other thoughts!
Given a set of values, I see there as being multiple layers of heuristics, which are all useful to consider and make comparisons based on:
Yardsticks (e.g. x-risk, qualys)
Causes (e.g. AI alignment)
Interventions (e.g. research into the deployment problem)
Specific jobs /orgs (e.g. working at FHI)
Comparisons at all levels are all ultimately about finding proxies for expected value relative to your values.
The cause level abstraction seems to be especially useful for career planning (and grantmaking) since it helps you get career capital that builds up in a useful area. Intervention selection usually seems too brittle. Yardsticks are too broad. This post is pretty old but tries to give some more detail: https://80000hours.org/2013/12/why-pick-a-cause/
INT = good per dollar by definition (when used in the quantitative rather than heuristic way), so in that sense it’s exhaustive, though in practice, people often miss some factors that are not as naturally captured by the framework:
People often assess ‘importance’ just based on one yardstick (e.g. QALYs) when there are effects on other relevant metrics (e.g. x-risk reduction, economic growth) (and the choice of which yardsticks to focus on in the first place is where a lot of the action is).
Value of information—can be included in either I or N, but often missed.
Coordination considerations e.g. portfolio approach, comparative advantage, trade with people with other values.
Cross-cutting epistemic considerations such as regression to the mean & epistemic humility & how to deal with unmeasured factors – partially covered in my recent podcast. People often only report their ‘unadjusted’ estimates and don’t account for these.
Movement building effects e.g. promoting one cause might bring people into others.
Funging e.g. if you solve one problem, it might free up resources to work on another.
I find it useful to try to make a ‘direct’ estimate using INT, and then to have a separate ‘all considered’ estimate that aims to take account of all the above.
The application of INT also gets more complicated depending on whether you’re interested in the resources spent in a certain year or over all time. Likewise, there are issues like complementarities between different forms of resources (e.g. funding vs. labour) that can mean the analysis is different for different resources.
You could also have another category of timing considerations, such as these. Toby’s soon, sharp, sudden framework helps to capture some of timing factors as well, or you could think of them as guides to what’s most important.
As an alternative, I think it’s also useful to think of cost-effectiveness analysis of specific interventions as a separate framework that provides a different perspective.
INT is also only about how pressing causes are in general. In practice if you’re making a real decision, you also need to consider your personal fit, career capital, the quality of the specific opportunity etc. as well as other moral considerations besides good done.
Agree. We should also probably expect it to happen: the income distribution is very heavy tailed, and it becomes easier to donate the more money you have, so we should probably expect the largest couple of donors to account for most of the money.
Otoh, the total US non-profit sector is something like 300 billion per year, and I think billionaire philanthropy is under $30bn, so that would suggest 10% from billionaires as a base rate. (Though a lot of this is to fund local services, churches etc. where we might expect a broader base.)
That’s the bit I’m most unsure about.
I think the longtermist EA amount is around $30m per year—and I have reasonable data on that.
I then guessed that there’s another $50m of near termist donations (based on typical ratios of near termist to longtermist donors). Note this needs to include all effective animal advocacy and Founder’s Pledge. However, it might be that most of this category overlaps with the GiveWell donations, so I might have been overoptimistic. Still I would guess that it’s ~$20m, making for $50m+ in total within other.
I’ve updated it to $50m rather than $80m.
I think total EA funding is something like (per year):
$250m Open Phil / Good Ventures
$80m from GiveWell (excluding Open Phil)
$50m other
So that’s 66% Open Phil. Note that Open Phil seems to be 90%+ Dustin and Cari.
Sorry I don’t have sources for these figures—they’re my personal rough estimates. (Though Open Phil’s grants are almost all published.)
One other thing to note is that funding is only one component of EA – we also have the value of the labour of community members and our ideas. Even if there was no funding at all, we could still accomplish a bunch (by going to work in existing institutions like govt., academia, non-profits, or working as volunteers).
Quick note on the figures for 80k: about 66% of our funding comes from Open Phil in recent years. You can’t divide the size of the grant by our annual budget, because we’re also building up reserves with the scale of the org. You instead need to divide the grant by our total income.
I think it’s a good question, but it’s pretty complex, so it would take me a while to elaborate, sorry!