I really appreciate this post! I have a few spots of disagreement, but many more of agreement, and appreciate the huge amount of effort that went into summarizing a very complicated situation with lots of stakeholders over an extended period of time in a way that feels sincere and has many points of resonance with my own experience.
ClaireZabel
Karma: 3,403
Seconding Ben, I did a similar exercise and got similarly mixed (with stark examples in both directions) results (including in some instances you allude to in the post)
Thanks for sharing this, Tom! I think this is an important topic, and I agree with some of the downsides you mention, and think they’re worth weighing highly; many of them are the kinds of things I was thinking in this post of mine of when I listed these anti-claims:
Anti-claims
(I.e. claims I am not trying to make and actively disagree with)
No one should be doing EA-qua-EA talent pipeline work
I think we should try to keep this onramp strong. Even if all the above is pretty correct, I think the EA-first onramp will continue to appeal to lots of great people. However, my guess is that a medium-sized reallocation away from it would be good to try for a few years.
The terms EA and longtermism aren’t useful and we should stop using them
I think they are useful for the specific things they refer to and we should keep using them in situations where they are relevant and ~ the best terms to use (many such situations exist). I just think we are over-extending them to a moderate degree
It’s implausible that existential risk reduction will come apart from EA/LT goals
E.g. it might come to seem (I don’t know if it will, but it at least is imaginable) that attending to the wellbeing of digital minds is more important from an EA perspective than reducing misalignment risk, and that those things are indeed in tension with one another.
This seems like a reason people who aren’t EA and just prioritize existential risk reduction are less helpful from an EA perspective than if they also shared EA values all else equal, and like something to watch out for, but I don’t think it outweighs the arguments in favor of more existential risk-centric outreach work.
This isn’t mostly a PR thing for me. Like I mentioned in the post, I actually drafted and shared an earlier version of that post in summer 2022 (though I didn’t decide to publish it for quite a while), which I think is evidence against it being mostly a PR thing. I think the post pretty accurately captures my reasoning at the time, that I think often people doing this outreach work on the ground were actually focused on GCRs or AI risk and trying to get others to engage on that and it felt like they were ending up using terms that pointed less well at what they were interested in for path-dependent reasons. Further updates towards shorter AI timelines moved me substantially in terms of the amount I favor the term “GCR” over “longtermism”, since I think it increases the degree to which a lot of people mostly want to engage people about GCRs or AI risk in particular.
Seriously. Someone should make a movie!
Very strongly agree, based on watching the career trajectory of lots of EAs over the past 10 years. I think focusing on what broad kinds of activities you are good at and enjoy, and what skills you have or are well-positioned to obtain (within limits: e.g. “being a really clear and fast writer” is probably helpful in most cause areas, “being a great salsa dancer” maybe less so), then thinking about how to apply them in the cause area you think is most important, is generally much more productive than trying to entangle that exploration with personal cause prio exercises.
Our impression when we started to explore different options was that one can’t place a trustee on a leave of absence; it would conflict with their duties and responsibilities to the org, and so wasn’t a viable route.
Chiming in from the EV UK side of things: First, +1 to Nicole’s thanks :)
As you and Nicole noted, Nick and Will have been recused from all FTX-related decision-making. And, Nicole mentioned the independent investigation we commissioned into that.
Like the EV US board, the EV UK board is also looking into adding more board members (though I think we are slightly behind the US board), and plans to do so soon. The board has been somewhat underwater with all the things happening (speaking for myself, it’s particularly difficult because a lot of these things affect my main job at Open Phil too, so there’s more urgent action needed on multiple fronts simultaneously).
(The board was actually planning and hoping to add additional board members even before the fall of FTX, but unfortunately those initial plans had to be somewhat delayed while we’ve been trying to address the most time-sensitive and important issues, even though having more board capacity would indeed help in responding to issues that crop up; it’s a bit of a chicken-and-egg dynamic we need to push through.)
Hope this is helpful!
My favorite is probably the movie Colossus: the Forbin Project. For this, would also weakly recommend the first section of Life 3.0.
Hey Jack, this comment might help answer your question.
That’s correct. It’s common for large funders of organizations to serve on the boards of organizations they support, and I joined the EVF board partly because we foresaw synergies between the roles (including for me acting as grant investigator on EVF grants). Leadership at both organizations are aware I am in both roles.
Also, though you didn’t ask: I don’t receive any compensation for my work as an EVF board member.
Hey, I wanted to clarify that Open Phil gave most of the funding for the purchase of Wytham Abbey (a small part of the costs were also committed by Owen and his wife, as a signal of “skin in the game”). I run the Longtermist EA Community Growth program at Open Phil (we recently launched a parallel program for EA community growth for global health and wellbeing, which I don’t run) and I was the grant investigator for this grant, so I probably have the most context on it from the side of the donor. I’m also on the board of the Effective Ventures Foundation (EVF).
Why did we make the grant? There are two things I’d like to discuss about this, the process we used/context we were in, and our take on the case for the grant. I’ll start with the former.
Process and context: At the time we committed the funding (November 2021, though the purchase wasn’t completed until April 2022), there was a lot more apparent funding available than there is today, both from Open Phil and from the Future Fund. Existential risk reduction and related efforts seemed to us to have a funding overhang, and we were actively looking for more ways to spend money to support more good work, especially by encouraging more people to dedicate their careers to addressing the associated risks. Also, given the large amounts of funding available at the time, and relatively lower number of grantmakers, we wanted to experiment with decentralizing some influence over funding to other people with experience in the space who understood our long-run priorities but had visions for how to use the funding that were somewhat different from our grantmakers’.
So, we were experimenting with more defaulting to saying “yes” to shovel-ready grant requests that seemed aimed at our core priorities, especially when they didn’t require enormous amounts of funding. (As others have noted and as I’ll explain below, we modeled the amount of funding at stake here as being in the low millions of dollars, rather than the higher sticker price, since it was to purchase a durable asset that can be resold.)
What did we think about this grant? In the abstract, we thought the idea of buying real estate to use for events was a reasonable one.
When evaluating a grant, my team tends to focus on the question “to what extent (per dollar) will this grant result in more promising people focusing their careers on doing as much good as possible in our longtermist focus areas?” (Other Open Philanthropy programs use different criteria.) I and other people on my team have invested a fair amount in collecting data and developing metrics for evaluating the value-for-money of grants aimed at this goal, though also it’s still a work in progress in many ways.
When we surveyed ~200 people involved or interested in longtermist work, we found that many had been strongly influenced by in-person events they’d attended, particularly ones where people relatively new to a topic came into contact with relatively experienced people working full-time in that area. (Some other data and our general experiences in the space are largely supportive of this too). Overall, we feel fairly confident in-person events like workshops, learning retreats, and conferences can be very impactful for people considering big career changes, and have historically often been good value-for-money via helping people interested in our longtermist focus areas make professional connections and deepen their understanding about these topics.
We support projects that cumulatively run dozens of events per year (disproportionately in the Bay Area and UK), and we saw several reasons a venue purchase could be valuable for both increasing the number and quality of impactful events, and potentially saving some money:
The number of events was growing fast.
Suitable venues are often of limited availability and book up far in advance, making more spontaneous events challenging.
Rented venues sometimes have issues that aren’t discovered until an event has begun (and we’ve seen events go substantially worse, or have to execute difficult mid-event moves, because of issues arising with the venues after the event was already underway).
Ops capacity to organize before and after events is a resource we value highly. Organizing and preparing for events takes more time and may go worse on average when you need to orient to the layout of different venues, move in and out all the equipment, learn about new areas (and travel to/from them) and available vendors that the event allows, etc.
Renting venues varies a lot in price, but venues for events near top EA hubs can be very expensive. Often, the cost of the venues was in the $30-100k range (though there were also smaller, shorter events aimed at students with venue costs in the $10k-ish range, and large conferences can be much more expensive).
Other locations are often cheaper, but have a more difficult time attracting busy professionals, relative to ones that take place near where they live, and demand more time from ops staff, senior staff, and (sometimes) instructors helping with events (and, we generally highly value all those folks’ time, even if they are willing to travel).
The fact that the funding was going to be used to purchase a durable asset means that for the purposes of our cost-effectiveness analysis, we modeled the financial stakes as being somewhere in the low millions (rather than the meaningfully higher sticker price of the property). Our grantees also stood to potentially save money through the investment, since they wouldn’t end up paying rental costs to run their events. We still wouldn’t have been surprised if the investment turned out net negative from a financial perspective, but losing most of the investment seemed unlikely. Given this model of the costs, we thought it was consistent with our overall experiment in lowering barriers to longtermist funding to “default to yes” for a shovel-ready grant.
We discussed Owen’s reasoning behind selecting this kind of venue, and we thought it broadly made sense. The main goal was to have a venue relatively near Oxford with sufficient capacity for the types of events we thought could be highly impactful.
Despite the points above, we were relatively uncertain about this specific opportunity, and there was some internal debate over the amount and structure of our funding; we also expressed our uncertainty to Owen.
We were mainly uncertain about:
Owen being the person driving this forward, despite him not wanting to be the main point person actually helping oversee the property and run relevant events over the next few years
Whether refurbishing and maintaining the property would end up being much more time-consuming and expensive than Owen forecasted
Whether the process was moving too quickly and whether more should be done to investigate the state and expected value of the property, and whether a better opportunity might arise in the coming years
Whether the number and kinds of events (and the value of those events) had been sufficiently well-scoped, and whether this venue would prove a good fit for a sufficient fraction of them.
Internally at OP (including with our communications team), we discussed how this would reflect on us somewhat, but mostly from the perspective of us as a funder, rather than the wider ecosystem as a whole. I think that it was a mistake on my part that the funding conversations didn’t focus more on that broader question, and I regret it, though I think we should have a high bar to rejecting otherwise solid-seeming grants on optics grounds because that kind of approach can be hard to limit and then can end up being surprisingly costly to impact. (My recollection is that we thought of it mainly as a longtermist/existential risk reduction field resource and didn’t think enough about how what was then called CEA providing fiscal sponsorship might make it sound more like an EA project.)
We also told Owen that we would want to check in about venue usage and think about the value it was generating in terms of the community growth metrics we generally use, and discuss selling the property if it didn’t seem like things were going well. (Proceeds from a sale would be used as general funding within EVF, and that funding would replace some of our and other funders’ future grants to EVF.)
Where do I stand on this grant now? With the huge decline in available funds since November 2021, I don’t know whether we’d make this grant again today. I still think it could turn out to be importantly effort-saving and event-increasing-and-improving relative to regularly renting one-off event venues, and it could also lead to a higher number of impactful events being run. But it’s currently too soon to say whether the usage will justify the investment. If we were considering a similar grant now, we’d want to get more into the details of modeling the effective financial cost (incorporating the resale possibility), hammering out plans and predictions for future operations and usage, etc. but I think we might consider it to be good enough value-for-money that we’d want to go forward with some possible projects in this vein.
Why isn’t there a published grant page right now? (This isn’t my domain but) we typically aim to publish grants within three months of when we make our initial payment, but we’re currently working through a backlog of older grants. Wytham is one of many grants in that category.
- I’m less approving of the EA community now than before the FTX collapse by (Dec 16, 2022, 7:40 PM; 225 points)
- Should EVF consider appointing new board members? by (Feb 5, 2023, 6:01 AM; 188 points)
- Keep EA high-trust by (Dec 22, 2022, 2:58 PM; 156 points)
- Why did CEA buy Wytham Abbey? by (Dec 6, 2022, 2:46 PM; 142 points)
- Reflections on Wytham Abbey by (Jan 10, 2023, 6:30 PM; 82 points)
- Overall Karma vs Agreement Karma by (Dec 28, 2022, 10:04 PM; 66 points)
- 's comment on CEA Disambiguation by (Dec 19, 2022, 8:05 PM; 52 points)
- 's comment on Announcement on the future of Wytham Abbey by (Mar 26, 2024, 7:46 PM; 47 points)
- 's comment on Should EVF consider appointing new board members? by (Feb 6, 2023, 2:29 AM; 27 points)
- 's comment on Isn’t this a conflict of interest and self-dealing on the part of CEA/EVF leadership? by (Dec 14, 2022, 11:30 PM; 22 points)
- 's comment on Wall Street Journal Reporter Looking to Speak to EAs by (Dec 15, 2022, 5:52 PM; 20 points)
- The case for transparent spending by (Dec 15, 2022, 5:42 PM; 19 points)
- 's comment on Reflections on Wytham Abbey by (Jan 11, 2023, 3:13 PM; 13 points)
- 's comment on Isn’t this a conflict of interest and self-dealing on the part of CEA/EVF leadership? by (Dec 14, 2022, 11:30 PM; 8 points)
- Isn’t this a conflict of interest and self-dealing on the part of CEA/EVF leadership? by (Dec 14, 2022, 10:56 PM; 1 point)
This is also great, thank you!
Not the intended audience, but as a US person who lives in the Bay Area, I enjoyed reading this really detailed list of what’s often unusual or confusing to people from a specific different cultural context
I generally directionally agree with Eli Nathan and Habryka’s responses. I also weak-downvoted this post (though felt borderline about that), for two reasons.
(1) I would have preferred a post that tried harder to even-handedly discuss and weigh up upsides and downsides, whereas this mostly highlighted upsides of expansion, and (2) I think it’s generally easier to publicly call for increased inclusivity than to publicly defend greater selectivity (the former will generally structurally have more advocates and defenders). In that context I feel worse about (1) and wish Scott had handled that asymmetry better.
But I wouldn’t have downvoted if this had been written by someone new to the community, I hold Scott to a higher standard and I’m pretty uncertain about the right policy with respect to voting differently in response to the same content on that basis.
Quite. I was in that Stanford EA group, I thought Kelsey was obviously very promising and I think the rest of us did too, including when she was taking a leave of absence.
No worries, appreciate ppl checking :)
As noted in the post, I got Scott’s permission before posting this.
I strongly disagree with Greg. I think CFAR messed up very badly, but I think the way they messed up is totally consistent with also being able to add value in some situations.
We have data I find convincing suggesting a substantial fraction of top EAs got value from CFAR. ~ 5 years have passed since I went to a CFAR workshop, and I still value what I learned and think it’s been useful for my work. I would encourage other people who are curious to go (again, with the caveat that I don’t know much about the new program), if they feel like they’re in a place of relative strength and can take a discerning eye to what they’re taught.
If I, with (mostly) admirable candour, describe a series of grossly incompetent mistakes during my work as a doctor, the appropriate response may still be to disqualify me from future medical practice (there are sidelines re. incentives, but they don’t help)
I think doctor is a really disanalogous example to use; doctors are in one of the relatively few professions where screwups regularly lead to death; we want to some somewhat risk-averse, with respect to doctors (and e.g. pilots or school bus drivers), at least if the screwups are the very dangerous kind (as opposed to like, being terrible at filing one’s paperwork), and aren’t based on a reasonable CBA (e.g. enrolling patients in a clinical trial with a drug that looked promising but turned out to be dangerous). For lots of other professions, this example looks way less compelling; e.g. I doubt people would think that a startup founder or movie director or author who had a bunch of failures but also some big wins should be banned from their profession or ostracized in their community. I think in-person overnight events about psychology are in a pretty in-between risk category.
You said you wouldn’t tell anyone about your friend’s secret, but this seems like a situation where they wouldn’t mind, and it would be pretty awkward to say nothing…etc.
This isn’t your main point, and I agree there’s a lot of motivated cognition people can fall prey to. But I think this gets a bit tricky, because people often ask for vague commitments, that are different from what they actually want and intend. For example, I think sometimes when people say “don’t share this” they actually mean something more like “don’t share this with people that know me personally” or “keep it in our small circle of trusted friends and advisors” or “you can tell your spouse and therapist, but no one else” (and often, this is borne out when I try to clarify). Sometimes, I think they are just trying to convey “this info is sensitive, tread with care”. Or, they might mean something more intense, like “don’t share this, and aim not to reveal any information that updates others substantially towards thinking it’s true”.
Clarification can often be useful here (and I wish there were more verbal shorthands for different levels of intensity of commitment) but sometimes it doesn’t happen and I don’t think, in its absence, all agreements should be taken to be maximally strict (though I think it’s extremely important to have tools for conveying when a requested agreement is very strict, and being the kind of person that can honor that). And I think some EAs get intense and overly scrupulous about obeying unimportant agreements, which can be pretty unwieldy and divorced from what anyone intended.
I think “do you keep actually-thoughtful promises you think people expected you to interpret as real commitments” and “do you take all superficially-promise-like-things as serious promises” are fairly different qualities (though somewhat correlated), and kinda often conflated in a way that I think is unhealthy and even performative.
Drift isn’t the issue I was pointing at it my comment