If you have a broadly effective altruist approach, we think 80,000 Hours is an attractive donation opportunity from a number of perspectives.
In this post, we’ll sketch some of these perspectives, and our current funding needs. This post is intended as a summary – full details on our plans and progress are provided in our annual review, and more details on our historical impact are provided here.
Although our evidence of impact isn’t as robust as some other organisations, we think the balance of evidence makes a persuasive case that 80,000 Hours is among the best funding opportunities in the community.
Note that this post is intended for people who are already familiar with our work. If you’d like an introduction, see our annual review, or ask me to send a summary (direct.ben at 80000hours dot org).
We track our impact in terms of impact-adjusted significant plan changes (IASPC). In the last three months, we’ve tracked an average of 130 per month.
- We think the marginal cost per plan change is under £250. If you include opportunity costs of staff time, that might increase to around £500.
- Each IASPC is worth at least £7,500 in donations to GiveWell-recommended charities (in expectation, on average), making our multiplier 15-fold. This is because over 2016, 13% of IASPC took the Giving What We Can (GWWC) pledge due to engaging with the guide, and we expect this to continue. GWWC estimates that each pledge is worth £58,000 in NPV, adjusted for the counterfactual, time-discounting and drop-out, and 13% of that is £7500.
- This is likely an underestimate of the donation multiplier because it doesn’t fully account for the chance of finding another new very large donor, even though we’ve advised multiple high-net-worth individuals in the past (defined as having over $100m), and influenced six large earning to give donors (defined as planning to donate over $300,000 per year by 2019)
- We roughly estimate that the donors intend to give 29% of the donations will go to GiveWell-recommended charities, 70% to meta-charity and existential risk, and 1% to animal welfare.
- See more detail of these estimates and caveats here.
For this reason, we think it’s very likely more effective to donate to 80,000 Hours than a GiveWell-recommended charity, insofar as we can maintain these levels of return, unless you require very strong evidence of effectiveness.
Compared to other meta-charities, we can also compete on our donation multiplier, though GWWC is higher. In addition, 80,000 Hours is probably a multiplier on other meta-charities and existential risk organisations.
However, the aim of 80,000 Hours is to solve talent gaps rather than funding gaps – under 30% of the plan changers seek to earn to give or take the pledge. Since we already justify our costs on the basis of additional donations, benefits besides donations come for “free”. But if you believe the effective altruism community is more bottlenecked by talent than funding, our impact on talent gaps is likely to be much greater.
£500 is very little money to significantly influence a young graduate’s career path, especially when you consider that many of the plan changes were made by top students at global top 20 universities. As one indication, average US college graduate earnings are about $68,000 (£54,000), so the market value of just one year of their labor to a top problem area is over 100-times as large as our costs.
Here are are some examples of how we’ve helped to solve pressing talent gaps over recent years, where we think it’s likely we can do something similar in 2017.
- The following effective altruist organisations either wouldn't exist or would have been significantly delayed without 80,000 Hours: Global Priorities Project, Raising for Effective Giving, dotimpact, and Animal Charity Evaluators.
- In 2016, 9 people took jobs at effective altruist organisations (Centre for Effective Altruism, Founder’s Pledge, Animal Charity Evaluators) who were significant plan changes, and say they’re significantly less likely to have the job without 80,000 Hours.
- 2015-2016, we found 46 people who reported a plan change who (i) want to work on AI safety research (ii) are concerned by existential risks (iii) have studied a relevant quantitative subject, and (iv) switched towards this path due to us. We’ve prepared materials such as the AI safety syllabus and introduced them to mentors to help them enter. One has landed a job at a top AI organisation.
- Based on a sample, 15% of the plan changers say they switched which global problem they want to work on, and now intend to work on our top problem areas (promoting effective altruism, global priorities research, AI safety research or biosecurity) and no other areas. This would be over 200 people in 2016. If you only care about the plan changers who made a switch into a top problem area, the cost is £3,300 per IASPC (£500/15%).
We think the marginal multiplier perspective is not however, the main reason to donate to 80,000 Hours. Rather, the majority of the value of donations comes from them giving us the chance to become much bigger in the future. This is what the Open Philanthropy Project calls “hits based giving”, and what we called the “growth approach” to evaluating non-profits.
80,000 Hours could be far bigger than it is today. If we could reach most of the students at top 20 universities who care about social impact, then we expect to cause at least 3-times as many plan changes each year. And that would mean that in twenty years, 80,000 Hours would have influenced a significant fraction of leaders in science, politics, business and other fields.
From there, we could expand into a wider range of universities and ages (e.g. into people mid-career). We could also provide advice to people who don’t already focus on social impact, with the aim of tilting them in that direction (we already do this as a side effect).
For this reason, 80,000 Hours has more room to grow than many other meta-charities. For instance, Raising for Effective Giving has already reached a large fraction of poker players, so it’s going to be hard to get 10-times larger within that audience. (CEA and perhaps some seed funding opportunities seem like the main exceptions.) This is a major additional benefit over the multiplier.
In addition, as we grow, we expand our alumni community that donates back to us. Donations now let us accelerate the growth of this community, unlocking more donations in the future.
80,000 Hours may be the fastest growing meta-charity. In the last two years, the rate of IASPC per month has increased 20-fold, reaching 150 per month (on costs that only increased 40% p.a.). We were also admitted to Y Combinator’s non-profit program that assesses non-profits using the growth approach. Any assessment of 80,000 Hours as a donation opportunity should take into account the likelihood that the organisation is 10-times larger in a couple of years.
As we’ve grown, the maximum upside of supporting 80,000 Hours has gone down (in that we have less room to grow than before), however, the likelihood of that growth being realised has gone up. Three years ago, reaching the scale we’ve reached today seemed like a long-shot; whereas now we’re confident we can get another 3-times larger.
Greater than 3-fold growth could be achieved just by focusing on our initial target audience of altruistic graduates at top universities. We could reach most of this audience through paid marketing for about £60,000 per year, so realising this upside seems likely.1
Many on the forum think that building the effective altruism community is a top priority. We think you can make a reasonable case that donating to 80,000 Hours is one of the most effective ways to grow the community.
Our career advice introduces people to the key ideas of effective altruism, and encourages them to get involved in the community. Over the past 12 months, 80000hours.org was the largest effective altruist website by traffic. Our newsletter has more subscribers than all the other organisations put together (88,000+). So we’re introducing more people to the ideas than anyone else.
We think career decisions are one of the best moments to introduce someone to effective altruism because (i) people are open to major shifts (ii) little good advice already exists (iii) we can help people to have a more fulfilling, successful career too. Contrast this with other organisations that focus on asking people to give away their money. In addition, we're the only effective altruist organisation focused on career decisions, whereas there are many organisations focused on donations and running student groups.
Over 2016, from a sample, 74% of those who made plan changes aren’t already involved with the community, and 43% said they intend to “get more involved in the effective altruism community”, which would be 608 people. There’s evidence people actually do get more involved – 80,000 Hours was the largest and highest-quality source of applicants to EAG this year, when measured in terms of acceptance rate. As shown above, we’ve also acted as a donation multiplier on meta-charities, increasing the funding for effective altruist organisations.
80,000 Hours also helps existing community members find higher impact careers, and resolve talent gaps in the community. Based on a sample, 9% of the plan changes, say they’re already heavily involved in the community. This would be 126 effective altruist careers switched per year.
Finally, 80,000 Hours contributes research and key considerations to the community, which makes it more effective over the long-term. One major example is simply the idea that career choices are important – much of the community was previously highly focused on donations. Some other examples include: the concept of talent gaps; rules of thumb for coordination; replaceability and its importance; earning to give, its pros and cons, and who should do it; the quantitative problem framework; and the growth approach.
At the same time, we’ve done this with a comparatively small budget. In 2016, 80,000 Hours only had 22% of the budget of CEA and 6% of the budget of GiveWell.2 When compared in terms of “effective altruism community growth per dollar”, 80,000 Hours has a good claim to be one of the most effective organisations.
We also see 80,000 Hours as helping to reduce existential risk. Right now, we think artificial intelligence is the most pressing form of existential risk (though we also encourage people to work on biosecurity and nuclear security). In Chapter 15 of Superintelligence, Bostrom proposes three key ways to mitigate existential risks from artificial intelligence (AI): (i) community capacity building (ii) strategy research (i.e. global priorities research) and (iii) technical safety research. Addressing these three priorities in reverse order:
- To contribute to technical safety research, we created a pipeline of about 50 potential technical AI researchers who are concerned by existential risks (as mentioned above).
- We encourage people to work on strategy research (which we call global priorities research) to work out how best to mitigate existential risk. The Global Priorities Project at Oxford likely wouldn’t exist without 80,000 Hours.
- We build the existential risk community. In part, we do this by building the effective altruism community (as covered above). But we also grow the AI community directly. For instance, the Assistant Director of the Future of Humanity Institute, Niel Bowerman, who does the day-to-day management of the organisation, was influenced by 80,000 Hours. There are also plan changers at many of the top AI companies and research institutes.
- As covered, we’re also a multiplier on donations to meta-charities and existential risk organisations.
If you would like to learn more about how 80,000 Hours helps to mitigate existential risks, email me (direct.ben at 80000hours dot org).
In previous years, 80,000 Hours hasn’t had a large room for more funding. However, given our recent success, we now think we should aim to scale up more aggressively. Several recent hires have also given us more capacity to invest in marketing and hiring.
For this reason, we’d like to grow our budget faster than the rate at which our existing donors can increase their funding, so we’ll need to welcome new donors or we won’t make our targets.
As further evidence, a survey of meta-charity donors carried out by Open Phil and 80,000 Hours found that they expect to give about £4.5m this year, and not all will go to meta-charities. Given that CEA is aiming to raise £2.5m-£5m alone, the capacity of meta-charity donors is going to be used up this year. This means we need new meta-charity donors, or good meta opportunities will go unfunded.
This all means that donations to 80,000 Hours this year will have an especially important impact on how fast we can grow.
However, we also think many people put too much weight on room for more funding as a factor. Our marginal donors are effective altruists, so if you donate to us, at worst you free up another effective altruist to donate elsewhere. Unless you have better information or unusual values compared to the rest of the community, this isn’t a bad outcome. In fact, by donating quickly you’ll save us and other donors time, which lets the community be more effective as a whole. We wrote more about donation coordination here.
We’d like to raise at least £1.7m. With this funding, over 2017, the aim is to triple the rate of IASPC. Here are some of the priorities we’d pursue to do that:
- Dramatically improve the career reviews and problem profiles, so we have in-depth profiles of all the best options, including advice on how to enter each area. For instance, we’d like to have in-depth profiles covering the main policy options, working at effective non-profits, and tech entrepreneurship.
- Create mentor networks. The easiest, scalable way to get more high-value plan changes seems like it’s to create ways to spot high-potential people and get them involved with our community. This year, we used a form on our website to find people interested in AI safety research careers, found a group of 50 with the relevant background, and found someone to mentor them. We’d like to do something similar for other key areas mentioned above. We could also do more to connect people with the effective altruism community.
- Make the online guide more engaging. We think there’s a lot that could be done here. For instance, we could create high-quality video content instead of articles. If this went well, we could put the guide on a MOOC platform like Coursera, which we expect to double the rate of sign-ups, and may improve the completion rate.
- Scale up outreach with the aim of saturating most of our target universities. We’d start by doing intensive outreach to one university in spring, then apply what we learn to 5-10 universities in our annual September outreach campaign. There’s scope to double or triple the number of students who use our advice, which could hugely grow the rate of plan changes. We could use digital advertising to gain new subscribers from target universities for under £4 (more detail).
This is roughly how the target breaks down:
- Cover our existing commitments to 6 full-time staff and freelancers over 2017 (£585k).
- Maintain at least 12 months’ reserves to give us financial security (to have this Dec ‘17, we need funding to cover 2018, which is £670k).
- Increase salaries about 30% to match comparable organisations and attract better new staff (£350k over two years).
- Hire two additional entry-level staff members to work on writing career reviews, giving workshops, or design (or one senior staff member) (£175k over two years)
- Expand our marketing budget (£165k).
- Subtract from the target the £250k of cash we already have on hand.
(See more detail on our plans and our budgets in our annual review)
Even if we made this target, it wouldn’t exhaust our room for more funding. If we raised more, we could increase our reserves, which would make it easier to attract staff, or we could pursue these expansion opportunities more aggressively (e.g. hire more, larger marketing budget).
We see the following as some of the main arguments against donating to 80,000 Hours, and we suggest nearby alternatives.
- Our impact is spread out over several problem areas. If you’re confident that one particular area is much more effective than the others, then it may be better to support that area directly. For instance, if you mainly just care about animal welfare, consider supporting ACE, which is a top pick by Open Philanthropy’s program manager in that area. If you just care about global poverty, then GWWC might provide a higher multiplier, at least in terms of donations (though GWWC is now merged with CEA). If you’re focused on AI risks then consider MIRI or FHI, which were picked by Daniel Daniel and Nick Beckstead at Open Phil, as well as Larks on the forum (though we think 80,000 Hours is still a contender).
- If you don’t think the problems that the effective altruism community focuses on are among the most pressing, then we’re not having a large impact.
- If you’re unwilling to consider weak evidence of impact, or unable to spend enough time to investigate us to gain confidence in our effectiveness over first order charities, then it may be better to support something first order with stronger evidence behind it, such as GiveWell’s recommendations within global health. Similarly, if there's another meta-charity you understand much better, it may make sense just to donate to it.
- If you think some aspect of our career advice is badly wrong, such that it’ll cause some users to have far less impact. (If so, please tell us which part!)
- If your top cause is building the effective altruism movement, it’s plausible that creating better movement strategy, marketing and “infrastructure” (e.g. good local groups; a central community website) is the top priority right now. CEA is more focused on these activities than 80,000 Hours, so we’re keen to see CEA funded at least to the point where it can cover its basic activities in these areas.
- You think most of our impact comes from the effective altruism community rather than our programs, and that we haven’t done much to help the community. This could be true if you think the majority of the value of effective altruism community comes from word-of-mouth growth, and isn’t helped by organisations like 80,000 Hours.
- You think the average value of the plan changes is declining over time, in a way that’s not captured by the impact-adjustment of the metrics. The best argument we’ve heard is that most of the value comes from the plan changes rated “10” recruited via one-on-one outreach i.e. maybe what we score “10” is actually worth “50”. The 10s are growing much more slowly than the 1s, so if you have this view, our marginal impact is worse. We, however, think the 1:10 ratio is reasonable, and that growth in the 10s will pick up with some delay – 10s take longer than 1s.
80,000 Hours is an attractive donation opportunity purely on the basis of its marginal multiplier, both in terms of donations and talent. But in addition, we have major potential to grow, and as a side effect, we’re one of the most-effective groups growing the effective altruism community. We’re also attractive if you’re mainly concerned by existential risks, and we have more room for funding than we’ve had in recent years.
One of these perspectives by itself might not be convincing. The nature of our work makes it hard to quickly convey a compelling and robust case for impact. However, we think if you consider the balance of evidence, there’s a good case to be made.
If you’d like to help us grow in 2017, leave your email in this form.
If you’d like to learn more about our progress over 2016, see our annual review.
1. There’s about 200,000 undergraduates at the world’s top 20 universities. 25% choose careers each year, so that’s 50,000 per year. About 30% care about their social impact, which is 15,000. We estimate we can get newsletter subscribers at target universities for under £4, so the cost to get 15,000 is £60,000 per year. The cost to merely reach them would be much less.
2. In 2016, our budget was about £250,000.
CEA’s budget was £1,332,298. Archived link, retrieved 18 Dec 2016.
GiveWell’s budget was approximately $5m. Archived link, retrieved 18 Dec 2016.
We spent $3.0m in the 12 months from December 1, 2014 to November 30, 2015. We currently project expenses of $4.9m for December 1, 2015 to November 30, 2016.