My Q1 2019 EA Hotel donation

On March 31, 2019, I donated 3200 GBP to the EA Hotel fundraiser via GoFundMe. The donation cost me $4,306.14 USD. My decision was based mainly on the information in the EA Hotel page and the documents linked from the donations list website page on EA Hotel, which include the recent Effective Altruism Forum posts.

In this post, I describe the reasons that influenced my decision to donate. I didn’t draft the post before donating, so some of the elaboration includes aspects that didn’t (at least consciously) influence my donation decision.

I limited the time I spent writing the post, and will most likely not be able to respond to comments. But please feel free to comment with your thoughts in response to my post or other comments!

NOTE: I have no affiliation with the EA Hotel. I have never visited it, nor have I closely collaborated with anybody living there. I did not show this post to anybody affiliated with the EA Hotel before posting. Nothing here should be taken as an official statement about the EA Hotel.

The sections of the post:

I like the idea of the EA Hotel
I like the skin-in-the-game of the key players
I like the execution so far
I see institutional risk reasons for lack of institutional funding: These reasons don’t apply to individual donors, so I don’t see the lack of institutional funding as a reason to dissuade me from donating
I have not been dissuaded by the reasons against donating that I have seen so far
I find the value of marginal donations high and easy to grasp
How I decided to donate and determined the donation amount

I like the idea of the EA Hotel

My interpretation of the fundamental problem the EA Hotel is trying to solve: provide low-cost and optimized transient living arrangements to people engaged in self-study or early stages of projects. The hotel’s low-cost living arrangements are further subsidized so that long-term residents don’t have to pay anything at all, and in fact, get a stipend to cover some living expenses. This means that residents can pursue projects with single-minded focus without burning through savings or having to do additional jobs just to keep themselves financially afloat.

The backdrop of the problem, as I understand:

EA communities have congregated in some of the most expensive places in the world, such as the San Francisco Bay Area, Boston/Cambridge, New York City, and London. Even outside of these, most places with significant numbers of EAs tend to be cities, and these tend to have higher costs of living.
Most EA projects have trouble raising enough money to cover costs of living in these places, even after they get funding. Moreover, most EA organizations, which are also based in these areas, do not pay enough of a premium for people to build savings that would allow them to comfortably spend months working on such projects in these expensive locations.
Tendencies within EA to donate large fractions of one’s personal wealth may have further exacerbated people’s lack of adequate savings to pursue EA projects.

These problems, specially the first one, have been widely acknowledged. Attempts to figure out a new, lower-cost city for EAs and build group housing in that city started since as far back as 2014, when the Coordination of Rationality/EA/SSC Housing Project group was created. Browsing through the archives of that Facebook group is interesting because it shows the amount of effort that has gone in over the years in identifying lower-cost living places for EAs. This is the group where EA Hotel founder Greg Colbourn first announced his intention to buy a hotel in Blackpool.

Side note: Peter McCluskey’s comment suggests that complaining about the high rent in major hubs is a signal of low status, because the most successful and influential people don’t seem to complain about it. This does not accord with my impression. More successful people tend to personally be able to afford high rents, but I’ve seen concern about high cost of living among people across the spectrum. It’s more that the set of high-status influential people doesn’t overlap much with the set of people who are most vocal about this problem and most interested in solving it.

I like the skin-in-the-game of the key players

One of the main problems with executing a project like the EA Hotel is that it is capital-intensive: a lot of money needs to be put in to implement an idea like this in a manner that is low-cost in the long run. In this case, Greg Colbourn had to spend 130,000 GBP to buy the EA Hotel, and this was after getting it at a cheap price! For comparison, Berkeley REACH pays something like $6,000 per month for rent for a cafe in Berkeley.

I find it impressive that Greg put his own money into purchasing the building and financing the first year of the hotel’s operations, at a time when getting outside support would be hard.

Also, browsing through the Coordination of Rationality/EA/SSC Housing Project group archives, it looks like both Greg and Toon Alfrink, the hotel’s manager, have participated in the group for quite a while. While that doesn’t in and of itself mean much, it does underscore their long-term commitment to the vision.

Skin-in-the-game of key players is very important for me when making a sizable donation to an organization that has not yet achieved the level of financial stability that would make its survival automatic.

I like the execution so far

This breaks down into a lot of pieces:

I like both the high demand and the competence exhibited in meeting the demand effectively
I like that the residents are very positive about the hotel
I find the resident profiles and their projects reasonably impressive
I like the openness about the hotel operations and the detailed information available on the website
I like the cost-effectiveness of the hotel and think it solves the problem (of providing low-cost living) that it set out to solve

I like both the high demand and the competence exhibited in meeting the demand effectively

The public announcement about the specific EA Hotel, after buying it, was made by Greg Colbourn on June 18, 2018. Occupancy started climbing through July and August. On August 9, the website had been launched and bookings could be made through the website.

By September, the hotel appeared to be running in full steam, and even got press coverage in The Economist and The Times. For the last six months, the hotel appears to have been operating at reasonably high occupancy. The occupancy chart across time is available with the header “Number of guests over time” on the EA Hotel wiki.

Starting a new project and immediately getting traction is impressive. It’s even more impressive to have been able to cope well with the demand and keep the hotel residents satisfied.

I like that the residents are very positive about the hotel

Hotel residents appear quite positive about their experience at the hotel; see for instance Matt Goldenberg (halffull) and the comments on the other posts discussing the hotel.

I find the resident profiles and their projects reasonably impressive

The description of the residents so far and the projects and work completed are both decent, and particularly good for something that’s just gotten off the ground.

I like the openness about the hotel operations and the detailed information available on the website

The EA Hotel website offers a lot of valuable information for potential hotel residents, but also for others interested in understanding the progress of the hotel. For instance, it includes a description of current guests and their work, a historical occupancy chart, and detailed information about food, cleaning and laundry, showers, wifi, and more.

The EA Hotel team has also been quite forthcoming in sharing information in their fundraiser posts, including information they didn’t initially expect to share.

I like the cost-effectiveness of the hotel and think it solves the problem (of providing low-cost living) that it set out to solve

Let’s dig more into this. The official estimate for the EA Hotel is a cost of 5700 GBP per resident per year, including the stipend paid out to them. That translates to about $7500 USD per person per year.

How low-cost is this compared to other options? It’s definitely cheaper than the average deal in the San Francisco Bay Area or New York City. My anecdotal evidence suggests it would be very very hard to get a place in the San Francisco Bay Area where even the rent + utilities is lower than $7500/year, and once you add food, it will clearly exceed that range. Landing a good deal like this would itself take weeks if not months of searching, time that can burn savings or runway.

I know of two categories of options that would be cheaper than the EA Hotel:

Living with others who are taking care of the main living expenses: This could mean living with parents or with a spouse, a sibling, or friends, who are shouldering the majority of the household living expenses. Depending on the living arrangement, the marginal cost of the resident may not be too high. For instance, you might be living with your parents in a room in their house, but they wouldn’t have rented it out otherwise, so the effective rent cost is zero. Food, utilities, etc. are all shared, and the marginal cost of your food and utilities would likely be in the $200 to $400 range. Similar arrangements may exist with a spouse or sibling.
Living in a much lower-cost part of the world: Example places include Eastern Europe, Southeast Asia, and South Asia.

The disadvantages of these options:

Some people may not have others willing to take care of main living expenses, or may not be comfortable seeking their help: In particular, a person who quits a well-paying job to work on an EA project may not feel comfortable or welcomed reducing living costs by going and living with their parents.
Living in places with different culture and language may not be conducive to focusing on an EA project: Going to Thailand or Romania to cut costs may mean too much adjustment time and effort and not enough resources left to focus on the project. For some people, this option might actually work well, based on their background and comfort levels in different places. But for others, this option may not work.
In both cases, one does not have the advantage of being plugged into a network of other people also working on and motivated to make progress on similar projects.

Based on this analysis, I think the EA Hotel offers a good option that is not clearly dominated by any other. Moreover, the general idea of having such a hotel, if successful, could be replicated in other locations. That would lead to a wider range of price points to meet the needs of people with different trade-offs.

I see institutional risk reasons for lack of institutional funding

According to a comment on a question asking why the EA Hotel is having trouble fundraising:

EA Grants is taking a long time to review the hotel’s application. There have been rumours that this is because of one or several of the following: (a.) staffing issues at CEA which have only recently been resolved; (b.) desire to perform an audit/review of the hotel. Also, to see some operational changes at the hotel before extending runway; (c.) a strategic decision to delay funding the hotel as a countermeasure against cultural dilution, or PR risk. (c.) seems unlikely.

I can’t speak to the accuracy of the claims here, but they are suggestive.

The EA Hotel is a sufficiently distinctive project, and also an incubator of many other projects, that institutions funding it may perceive risk to their reputation from even one of the funded projects by the EA Hotel going wrong. I don’t expect this to prevent the EA Hotel from getting institutionally funded in the long run. But I do expect the process to take longer. I think a stronger, longer track record will be needed before it gets funded by EA Funds, EA Grants, or other larger sources of funding in the EA community.

This means that individual donors, who have less of these institutional risk concerns, have a bigger role to play in the beginning.

I have not been dissuaded by the reasons against donating that I have seen so far

This applies first and foremost to the reasons I had access to prior to making my donation decision. But I’ve been keeping an eye on the comments on the Grue Slinky’s $100 prize for best argument post, the question post by Milan Griffes, and the case for the EA hotel post and even now, I don’t see anything that changes my stance meaningfully.

My synthesis of the arguments so far is that there is a strong prior in favor of supporting individuals and projects directly rather than supporting the EA Hotel. In this view, the evidence emerging from the EA Hotel so far does not meaningfully alter that prior. Different variants of this story:

Challenging the idea of the hotel as a hub: The very idea of having people congregate at the EA Hotel to work doesn’t make sense; it doesn’t add value relative to them going to other places and working.
Challenging the financing model of the hotel: The EA Hotel is a great idea, and should continue to exist as an option for individuals working on projects, but individuals coming to the EA Hotel should get their own funding and pay at cost. This will reduce the EA Hotel’s operational costs to near-zero, because it is recovering costs from all residents (this is Argument 2 in Brendon Wong’s comment).
Challenging the specific accomplishments so far: The specifics that we have of the EA Hotel residents and their outputs is not yet strong enough evidence that the EA Hotel can generate significant value. However, if in the future, much more high-value stuff comes out of the EA Hotel, the EA Hotel would become a worthwile donation target.

I see all the critiques as having some value. In fact, it may well be the case that the sustainable model for the hotel in the long term would be to put a much lower ceiling on the number of long-term residents for whom fees are waived. Nonetheless, I think the evidence so far is positive enough, and the people involved in running the hotel have enough skin in the game, that I expect with high probability that they will make the necessary adjustments based on the feedback they get. In the meantime, I think there is a lot of value in this experiment, and I want there to be enough room for the EA Hotel team to plan the transition and tweaking of their models without causing too much disruption.

I find the value of marginal donations high and easy to grasp

The EA Hotel has a relatively easy-to-grasp operational math: it costs about 240 GBP per day to run, with the cost varying a bit based on current occupancy, and in particular on the mix of residents (how many are paying cost and how many are staying for free). The fundraiser uses this math to ask for 130,000 GBP to cover 18 months of operations. Greg Colbourn gives a figure of 265 GBP in this comment but clarifies that actual costs in January and Febrary where 25% less after factoring in rent paid costs; I’m taking 240 GBP as somewhere in between the theoretical figure and the observed figure for January and February.

In the same comment, Colbourn says:

Our costs are ~£8k/month, so even buying a month or two runway would make a big difference in terms of giving us some breathing space to work on getting more money coming in. It’s also approximately continuously divisible in that every ~£265 will keep us going another day.

With this model of spending, I can concretely think of my donation of 3200 GBP as having added about 13 days to the hotel’s runway.

The aspect of this that’s hard to quantify: I’m not actually just buying 13 days of the hotel’s operations. My goal is to help the hotel extend its runway so that it can be better equipped to do some combination of raising more money and making the operational changes needed to survive with whatever money it has. How good 13 days is in terms of making the hotel more likely to fundraise enough money to survive in the long term, I don’t know. Nonetheless, it is still a relatively clear calculation with a potential for high impact.

The other part that is hard to predict in advance: how much my donation might trigger other donations. It is possible that some large donations can spur other large donations by sending a signal. It’s even possible that this post might convince more people to donate to the EA Hotel, though that is not the primary purpose of my writing this post.

One area of concern I have is that the fundraiser has not gone well. I am not concerned in terms of what it says about the value of funding the hotel, but I do have concern in terms of what it says about the hotel’s ability to raise funds in the longer term. However, I chose to donate despite this because:

The EA Hotel team seems to be superfocused on actually running the hotel, and doing a good job of it. It’s natural if this doesn’t leave a lot of time or energy for fundraising.
As mentioned earlier, both Greg Colbourn and Toon Alfrink seem to have enough skin in the game that I’m somewhat confident that with more resources they will be able to focus their attention on the problem of funding.
Judging from the past history of similar EA projects, such as Berkeley REACH, there may well be an inflection point coming soon when the donation volume to the EA Hotel goes up. With Berkeley REACH, the situation seemed quite dire till June 2018, but then with both Eliezer Yudkowsky and Ben Hoffman plugging REACH, a number of people made pledges. Later, REACH also got institutional funding from BERI and EA Funds. Today, REACH has $3,528 in committed monthly funds.

How I decided to donate and determined the donation amount

Unlike the other sections, this section is a more personal one that discusses how I traded off my own finances against what I saw as a funding gap that needed filling. For background, see my donation history page and my 2018 donations post.

A few months ago, back when the first EA Hotel fundraiser post was published, I had looked at the EA Hotel fundraiser and dug around a bit on their website. I remember doing so while doing the initial round of adding documents related to the EA Hotel on the donations list website (the date of this commit was January 9, about two weeks after the fundraiser post).

Since then, I didn’t think actively about the EA Hotel till I saw more posts of theirs on the Effective Altruism Forum in the past week. On March 30, I added a bunch more documents related to the EA Hotel to the donations list website. As I read through the documents, I started thinking through the arguments described in this post. I came to the tentative conclusion that there was an important funding gap that urgently needed to be filled.

I ran this thinking by my esteemed friend and collaborator-on-many-projects Issa Rice, who did not immediately point out major flaws in my thinking. I slept over the thought on the night of March 30. On the morning of March 31, I reviewed the information, and read a new post The case for the EA Hotel. This solidified my decision to donate.

When I had initially been pondering how much to donate, I had been thinking of an amount in the range of 1500 to 4000 USD. I ultimately decided on a value slightly above the high end of this range. My reasons for donating as much as I did were:

In the past few months, I’ve ended up spending less on some other projects than I had anticipated, and therefore built up savings at a greater rate than expected. This has made it more feasible for me to donate a larger amount.
In light of all the overhead, both for me and for the recipient, I wanted a donation amount that made a big enough difference to justify the overhead. One week of operations did not intuitively feel like enough, but two weeks did.

The main reasons I didn’t donate more:

The primary constraint was how much I felt I could comfortably afford to donate without compromising my long-term saving targets. Although I could probably donate a little more, being confident of that would have required more careful financial calculation. The amount I ultimately ended up donating was at the high end of what I felt I could donate without having to do much calculation.
I am also interested in seeing the dynamic of how further donations play out. I think this amount is enough to serve as social proof that outsiders are willing to give money to the EA Hotel. Depending on how fundraising dynamics proceed, I may make one more donation.

The main reasons I didn’t defer the donation:

I felt I had enough information (thanks to all the extensive discussion on the EA Forum about the EA Hotel in the past few days) and was unlikely to get anything in the next few days that would change my stance significantly.
I felt a sense of urgency given that the existing runway was expected to last till May. Donating now, I felt, could extend the runway enough to let the fundraiser pick up pace.
Also, since I’m quite busy during the week, I felt that the weekend is a good time to make the donation and do the auxiliary work such as documenting it in this post.