Our Progress in 2016 and Plans for 2017

by Holden Karnofsky 0 Comments
Share

Throughout the post, “we” refers to GiveWell and Good Ventures, who work as partners on the Open Philanthropy Project.

This post compares our progress with the goals we set forth a year ago, and lays out our plans for the coming year. In brief:

Progress in 2016

Last year's post laid out plans for 2016. This section follows the order of that section, and quotes from it to allow comparisons between our plans and our progress.

Grantmaking

In the intro to last year's post, we wrote:

This year, we have a general goal of focusing on making grants in the most outstanding causes we’ve found. This is a departure from past years’ goals, which revolved around building knowledge and staff capacity. We expect to prioritize building knowledge and staff capacity again in the future, but we think this is a good year to focus on increasing our grantmaking. We currently are bottlenecked in terms of management capacity, and we believe that focusing on grantmaking will likely lead to a lot of learning that will inform future hiring and capacity building.

2016 progress was consistent with this: we did little hiring, and increased grantmaking greatly. We recommended over $100 million worth of grants. (A comparable figure for 2015 would have been under $20 million.)

Potential risks from advanced AI

Last year, we wrote:

Potential risks from advanced artificial intelligence will be a major priority for 2016 ... Daniel Dewey will work on it full-time, and Nick Beckstead and I [Holden] will both put significant time into it as well. Some other staff will be contributing smaller amounts of time as appropriate ... We won’t necessarily recommend a large amount of grantmaking in this cause, though it’s possible that we will. Our main goal is to lay the groundwork for more grantmaking in the future.

We recommended approximately $11 million in grants, including $5.56 million to start UC Berkeley's Center for Human-Compatible AI, $2.4 million to Montreal Institute for Learning Algorithms (writeup forthcoming), approximately $2 million to the Future of Humanity Institute, and $500,000 to Machine Intelligence Research Institute. Our grantmaking has focused on (a) creating more spaces for work on technical safety research in academia; (b) "shovel-ready" grants to existing organizations. We also put substantial effort into writing up our views on the topic, making connections with people in relevant fields, and developing a general grant process that we'll be using for more grants on technical safety work. We feel much better informed about this area, with a clearer sense of our future potential activities, than we did a year ago. We still consider it a highly promising philanthropic area, and we plan to continue allocating partial time from several staff (especially Nick Beckstead, Helen Toner and myself) as well as all of Daniel Dewey's time. Working this way in 2016 did not require cutting or suspending any other top-priority work, and we don't expect it to in 2017 either. We see this as a very long-time-horizon cause, and we don't think we can point to tangible impact of our work yet.

Biosecurity and pandemic preparedness

Last year, we wrote:

We hope to recommend at least $10 million worth of grants in the area of biosecurity in the next year. A full-time Program Officer will be starting work in this area soon.

With Jaime Yassif as our new Program Officer who started in mid-2016, we've recommended a bit over $21 million in grants in this area. $16 million of that was for the Center for Health Security; other major grants included $1.8 million for the Blue Ribbon Study Panel on Biodefense and $2.7 million for the Emerging Leaders in Biosecurity Initiative. We focused initially on "shovel-ready" grants to organizations that relevant staff had pre-existing knowledge of. We then spent some time setting priorities more systematically. In the near term our top three priorities will be: (1) risk assessment; (2) managing the risks posed by advances in biotechnology, its global spread and decreasing barriers to access; and (3) building capacity for independent policy research and advocacy. In the coming year we plan to refine our strategic priorities and set longer term goals for the program. Most of our grantmaking in this cause was done after mid-year and has aimed to build capacity for key institutions; we don't think we can point to tangible impact of this work yet.

Criminal Justice Reform

Last year, we wrote:

We hope to recommend at least $10 million worth of grants in each of criminal justice reform and farm animal welfare. We have cause-specific staffers leading our work in both areas.

We recommended about $20 million worth of Criminal Justice Reform grants in 2016.[2] Priorities include sentencing reform (including $3 million for the Alliance for Safety and Justice), organizing (including Faith in Texas, Mijente and The Ordinary People Society), prosecutorial accountability and reform (such as The Accountable Justice Project), and piloting and promoting alternatives to incarceration (such as restorative justice). We expect to publish significantly more content on this topic in the coming months. We've seen several cases where our grants may have had an impact, and will likely write more about some of these in the future (while some are best not to share publicly). So far, we believe the most significant case is that of a recent bill signed in Illinois that our grantee, the Alliance for Safety and Justice, estimates will result in approximately a 3000-person reduction in the prison population (we plan to discuss this in more detail in a future post). It is our understanding that ASJ developed the strategy for the bill, and worked very closely with the bill's champion Jehan Gordon-Booth, to cover all bases necessary in the state to get it passed. It passed with strong bipartisan majority and support from key law enforcement associations. Our grant to ASJ constituted 54% of its income in 2016; we believe this grant substantially increased the scope of ASJ's work, and that the work ASJ did in Illinois is among the work that would have been unlikely to occur without our support.

Farm Animal Welfare

As quoted above, last year we wrote that we hoped to recommend at least $10 million worth of grants in this area. We recommended about $15 million worth of grants in 2016 (not including an additional $2.5 million we recommended in 2015). This included about $3 million each to Mercy for Animals (grants here, here and here) and The Humane League (here, here, and here). Our main focus was on corporate campaigns for better treatment of both layer and broiler chickens, both in the U.S. and abroad. Our grantees have now secured cage-free pledges from all of the top 25 US grocers, foodservice companies, and food manufacturers, and 16 of the top 20 fast food chains, which will benefit >250M hens annually once implemented. They have also secured global cage-free pledges from the world’s two largest foodservice companies, the world’s largest discount grocery chain, and several of the world’s largest hotel chains, and regional cage-free pledges from Latin America’s largest fast food operators and France and Britain’s largest grocers. And they have secured commitments from the largest US foodservice operators and several fast food chains to reduce the suffering of >200M broiler chickens annually. We also supported work on fish welfare, work on promoting animal product alternatives, and nascent work to build stronger farm animal welfare organizations in China and India. We view these as long-term efforts and don’t expect to see tangible progress from these grants soon.

Other U.S. policy

Last year, we wrote:

Alexander Berger’s top priority continues to be supporting Chloe and Lewis’s work (on criminal justice reform and farm animal welfare, respectively). We’re less set on what his other priorities will be, but with his remaining time, he is likely to work on some combination of:
  • exploring funding opportunities around a cluster of science and technology policy issues, such as policy around emerging technologies (potentially including advanced artificial intelligence) and other topics related to science policy and infrastructure.
  • investigating a set of funding opportunities in macroeconomic stabilization policy.
  • maintaining or expanding upon our grants in immigration policy and land use reform.
I’d guess that his work (aside from supporting Chloe and Lewis) will result in about $10 million worth of grant recommendations in 2016, but with high uncertainty.

Alexander spent less time on the above priorities than anticipated, instead focusing on supporting Chloe and Lewis, helping Open Philanthropy move toward spinning out as an independent organization, and working with the science team on grants around the use of gene drives in malaria control (discussed more below). We recommended about $2 million in grants to think tanks to work on macroeconomic stabilization policy (grant writeups here, here, here, here and here), as well as a renewal for Fed Up (writeup forthcoming). We also recommended $1-2 million for immigration policy and land use reform. We mostly think it is too early to see especially clear impact from these grants. We plan to lower the priority of new grantmaking (as opposed to renewals for existing grantees) in macroeconomic stabilization policy, immigration policy and land use reform this year. We've made a relatively small number of grants in these causes, and this is related to the fact that Alexander's attention has been divided between multiple priorities. (We've also updated in favor of the "expert" model of philanthropy described in this blog post.) For now, we anticipate mostly keeping our work in these areas in “maintenance mode,” while Alexander works on some other priorities (more below).

Scientific research

Last year, we wrote:

We still have a substantial amount of work to do before we’ll be ready to set focus areas within scientific research. We don’t plan to set a particular goal for 2016, because (a) we are allocating senior generalist time to potential risks from advanced artificial intelligence; and (b) we think time spent on scientific research will be much more productive after our scientific advisors come on board.

We had three Scientific Advisors - Chris Somerville, Heather Youngs and Daniel Martin-Alarcon - start in mid-2016. We've been investigating several areas of interest but have not completed any investigations yet. We have also:

Effective altruism

Last year, we wrote:

There is a strong possibility that we will make grants aimed at helping grow the effective altruist community in 2016. Nick Beckstead, who has strong connections and context in this community, would lead this work.

We recommended about $1 million to Founders' Pledge, and $1.4 million for SPARC and Center for Applied Rationality. We also made recommendations for Machine Intelligence Research Institute and Future of Humanity Institute (writeup forthcoming), though we've classified these as falling under potential risks from advanced AI rather than under effective altruism. We are hoping to recommend grants for 80,000 Hours and Centre for Effective Altruism soon. At that point we believe we will have covered most of the major opportunities in this area, but we will continue to keep an eye out. We expect to have substantive updates on these organizations' progress and the impact of our funds within about a year, but we don't at the moment.

Other grantmaking

Last year, we wrote:

As discussed above, we have significantly better capacity - and context - for evaluating potential grants than we’ve had in the past. As such, we are making a moderate effort to identify potential grants that don’t fit neatly into an existing focus area but are highly “shovel-ready” (in the sense that we can identify a specific potential grantee) and appear strong overall based on the knowledge we’ve picked up in the course of our many cause investigations and discussions. We don’t expect to make a large number of these grants, but we expect to make some.

We recommended about $5.5 million in grants on outstanding opportunities to mitigate risks from climate change: $2.5 million for research on solar geoengineering (writeup forthcoming) and $3 million to encourage reduction in hydrofluorocarbon emissions. We also made about $1 million worth of other "one-off grants" (see here, here and here). We wrote about our recommendation to Good Ventures to grant $50 million to GiveWell's top charities. We aren't including this figure when we say (above) that "We recommended over $100 million worth of grants" in 2016.

Progress on other fronts

Last year, we wrote:

  • We hope that by the end of 2016, the Open Philanthropy Project will operate as an independent organization from GiveWell, with its own employees and financials, though some individuals will continue to do work for both organizations.
  • We continue to commission case studies on philanthropy’s past successes as part of our history of philanthropy work. We also continue to collect relevant cases informally. By the end of 2016, we hope to publish a summary of what we’ve learned about the history of philanthropy, which will be a significant update from our 2012 post on the subject.
  • We hope to continue improving cross-cutting aspects of how we work. Our interests include improving our process for producing public writeups, refining our approach to cost-effectiveness estimates and comparing the good accomplished by different grants, refining our thinking about how much giving we should recommend to GiveWell's top charities, and finding ways to improve our general forecasting abilities (since our grantmaking decisions rely on views about what developments are likely in the future). We expect to do some work on all of these fronts, but it will not be a major priority with a concrete goal this year.
  • We still don't have the capacity we'd need to identify focus areas within the category of global health and development - beyond our existing connection to GiveWell's work and our support of the Center for Global Development. We don’t plan to build this capacity in 2016 and expect to revisit this in a future year.

The first of these - spinning off into an independent organization - turned out to be more complex and involved than we had anticipated. We put significant time into it in 2016 and are planning to complete the spinoff by mid-2017. We put less time than anticipated into the other points listed above. We did not publish an updated writeup on the history of philanthropy, and may or may not do so in 2017. We made some progress on the third point, including a revised approach to information sharing, starting to explore best practices for forecasting, improvements to our grantmaking process and new hires who focus on it, and an internal "Program Officer guide" that we expect to eventually create a public version of.

Takeaways from 2016

Speaking broadly, the main significance of 2016 for us is that we greatly increased our grantmaking, to the point where it is now comparable to that of major foundations. We expect to be able to maintain this pace and quality of giving recommendations (i.e. $100+ million per year, though with substantial year-to-year variation, from grants that we consider reasonably well-investigated and good expected value for money). If we had a goal of increasing the annual level of grantmaking further, we think we could fairly easily do so, and this has affected how we think about giving now vs. later. We now have a much more tangible sense of what the day-to-day key questions and operational challenges are for giving in this range. Many new questions and challenges have come up. At this level of giving, we now see more importance for careful public communications, the need for a more streamlined and systematic grantmaking process, and the need for more thinking about the question of how much grantmaking to aim for in each of our focus areas and how to decide when a grant is worth making. In particular, we believe that we will gain a lot from sharpening our thinking on several points related to worldview diversification - and the relative appeal of different worldviews - before increasing our annual giving level much more.

Plans for 2017

For reasons outlined above, we aren't planning much staff expansion in 2017. We plan to continue grantmaking roughly as we have been, while trying to improve the underlying processes and intellectual framework.

Continued grantmaking

Overall, I will not be surprised if 2017 grantmaking is lower than 2016 grantmaking, as there were some significant one-off grants in 2016 (particularly in the area of gene drives). But I'd guess that we will recommend at least $100 million worth of grants, and I'll consider it quite surprising if we don't recommend at least $80 million worth.

Cause prioritization framework and "last dollar" analysis

We recently wrote:

Our framework for “giving now vs. later” has gotten a lot more complex compared to last year’s, and it’s very far at this point from being as systematic as I would like. Currently, we apply the framework very loosely and informally ... In 2017, I hope to put significantly more time into the issues that were preliminarily addressed in this post, such as the value of the “last dollar” ...

A major goal of 2017 will be to reach and publish better-developed views on:

This work is likely to include addressing several of the fundamental questions we listed recently: how to compare cost-effectiveness across very different causes, how to make decisions under different kinds of uncertainty, how best to realize the benefits of worldview diversification, and which types of beings warrant our moral concern. It may also include some degree of further investigation on the moral value of the far future, and the question of how likely it is that we can have non-negligible impact on humanity's long-term prospects. We hope this work will put us in better position to make systematic day-to-day decisions about our grantmaking, and to choose the best possible cause prioritization and funding allocation as we scale up our giving further.

Capacity building

We hope to improve our staff capacity and processes on a few fronts:

Other

Footnotes
  1. [1]

    Throughout this post, "recommended" means that a recommendation was made by an Open Phil grant investigator and approved by Cari and Holden. However, the recommendation is generally conditional on finalizing a grant agreement and completing due diligence, and in many cases the next step after this is a recommendation to a donor-advised fund that has final say in what grant (if any) is made. Grants are not posted to our database until completed, so in many cases the date of a grant's being "recommended" in the sense used here is different from the date that a grant appears in the database.

  2. [2]

    Last year's review stated, "We hope to recommend at least $10 million worth of grants in each of criminal justice reform and farm animal welfare. We have cause-specific staffers leading our work in both areas. (In each case, we’d include grants already made in these areas - since late 2015, when both staffers joined - as part of the $10 million.)" However, we decided that this accounting method is needlessly confusing, and we're now just listing numbers for calendar-year 2016.