Note this is a sincere question. Not intended to cause controversy. It was inspired by this post questioning another OP Grant.
Full Disclosure
I applied to the Atlas Fellowship but was rejected. However, I attended SPARC, a free in-person program that teaches rationality tools to high schoolers (and follows a similar structure to Atlas Fellowship’s summer program). I’m friends with many Atlas Fellows.
What is the Atlas Fellowship?
For those newer to the EA Community, the Atlas Fellowship is a competitive program for high schoolers. If you are awarded it, you receive,
A $50k scholarship (or $12,000 for Atlas India).
Atlas Fellows can spend this money on anything considered an “academic expense”. This includes travel expenses if justification can be provided.
A fully-funded 11-day summer program in the Bay Area in a large former fraternity on UC Berkley’s campus.
College admissions preparation for top universities. (The admissions tutors are paid $200-300/hr).
Access to the $1m Atlas Fund to learn, experiment, and build impactful projects.
For 500 finalists, they receive $1,000 and 5 free books.
Total Cost of Prizes
$50k x 100 + $12k x 20 + $1k x 500 + $1m fund = $6.74m
This does not include the instructors, venue, or travel costs.
What made me write this post?
This came to my attention after reading EA London’s monthly newsletter. It highlighted new grants that Open Philanthropy made. I learnt that OP made an additional $1.8m grant to the Atlas Fellowship in December 2022
This is on top of a $5m grant they made in March 2022 and a $5m grant that the FTX Future Fund made.
There is much discussion (even amongst Atlas Fellows) that it is not a good use of money and that high schoolers don’t need $50k scholarships; therefore, I felt raising this question is worthwhile and of interest to the wider community.
Questions I have for Open Philanthropy and the Atlas Fellowship
Why do high schoolers need $50k scholarships? If the reason is to attract talent, why is this required when programs such as SPARC and ESPR do an excellent job of attracting talented high schoolers?
Note that SPARC and ESPR have been running for close to a decade. Many alumni go to top universities worldwide (MIT, Stanford, Oxford, etc.)
I estimate each Atlas Fellow costs $80-90k, given you need to divide the total costs by the number of fellows (i.e., instructor and venue costs should be considered).
If the answer is to attract better talent, is there a significant difference in talent between Atlas Fellows and those attending SPARC and ESPR that makes this $80-90k money worthwhile? (Note that this would be over 20 lives saved through the Against Malaria Foundation).
Why was a $50k scholarship offered if a $25k scholarship would attract, say, 80-90% of the same applicants?
I suspect that a $5k unconditional grant that they can spend on whatever would attract just as many quality applications and be much cheaper.
What is the breakdown of the socioeconomic background of Atlas Fellows? What countries are all Atlas Fellows from? What about the finalists?
Atlas says they’re doing “talent search”. This connotes finding talent from under-resourced communities or poor students. Do the statistics match this?
From friends who are Atlas Fellows, they said many Atlas Fellows do not require the scholarship as their parents earn a lot and can already pay for college. This makes me question why some people are accepted as $50k to the Against Malaria Foundation, which would save over ten lives. (Even more, if you consider the cost per participant is $80-90k). What are the Atlas Fellows spending the money on if their parents have more than enough to pay for college?
What measures do they have to identify talent that otherwise would not have been identified? The Atlas Fellowship claims to do research, but many fellows seem to be identified through traditional methods (i.e., reaching out to Olympiad communities worldwide).
Atlas is planning on running a school called Atlas Academy Beta, as outlined in this document. Why does this make sense?
Rumours I would like addressed
Many rumours circulate amongst Atlas Fellows that I would like addressed and would also be of public interest,
Atlas reportedly spent $10,000 on a coffee table. Is this true? Why was the table so expensive?
There is a claim that “Atlas has no inventory stock. They overspend, buy more than they need, and if they lose an item, they buy another.” Is this true?
Clarification: added to this comment.
A contractor for Atlas reported Sydney Von Arex (one of the co-founders) saying, “I do not believe in budgets”. Is this true?
I suspect this to be true as it would match my model of Bay Area EA’s lack of planning and overspending on events. I can go into further detail on this.
[Encrypted in rot13 by the moderation team] Gurer ner ehzbhef gung Flqarl Iba Nek (bar bs Ngynf’ pb-sbhaqref) vf qngvat/qngrq Pynver Mnory’f (gur tenag vairfgvtngbe sbe Ngynf nf yvfgrq ba gur cntr sbe gur svefg Ngynf tenag yvfgrq ba Bcra Cuvynaguebcl’f jrofvgr) uhfonaq, Ohpx Fuyrtrevf. Vf guvf gehr? Jul jnf gur tenag vairfgvtngbe sbe Ngynf Sryybjfuvc gur pb-sbhaqre’f oblsevraq’f jvsr? Abgr gung nyy guerr ner xabja gb or cbylnzbebhf va Onl Nern pbzzhavgvrf.
I suspect this may be true and coincides with comments on the EA Forums about EA funding being intermingled with personal relationships.
Edit 1: Corrected a typo in the encrypted text (‘husband’ to ‘wife’). This is a mistake that someone pointed out.
Edit 2: After seeing this post, someone DM’d me on the EA Forums with this screenshot. From it, I suspect that some Atlas money is not being used for Atlas purposes, given that Sydney encouraged someone to put “Fellow” to be reimbursed as one despite them never having applied and attending an another event (seemingly associated with LessWrong). On the announcement post for the LessWrong Lurkshop it does say at the bottom that the workshop is funded by the Atlas Fellowship. But, all this makes me question how Atlas are writing their grant requests (i.e., do they just ask for a fixed amount of money that they can spend on whatever or did they explicitly say to OP in their grant request they’re running this).
Edit 3: Someone also DM’d me this. To add to my point of too much money to teenangers, there is an 18-year-old Atlas Fellow who has both the Open Philanthropy Undergraduate Scholarship (which covers all tuition and living expenses) and the Atlas Fellowship (which is a $50k scholarship). [Two links removed by moderators; see comment.]
The disturbing part is that on the OP Undergraduate Scholarship website, it says, “and who do not qualify as domestic students at these institutions for the purposes of admission and financial aid.” But it is clear, upon investigation into their background, that this Atlas Fellow DOES qualify as a domestic student as they have grown up in the UK which means either OP doesn’t care about their own policies or this Atlas Fellow lied and it went unnoticed.
Edit 4: Updated total cost of prizes from $5.74m to $6.74m as someone pointed out I forgot to include the $1m Atlas Fund.
Edit 5: Added clarification to the claim on “no inventory stock”.
Edit 6: After this post came out, the Atlas Fellowship website was updated to remove Sydney from the cofounders. See now vs earlier (on archive.org). Why was she removed? Was this only publicly to prevent further controversy regarding grantor-grantee relationships and she will still be helping out behind-the-scenes or was she actually removed?
Not affiliated with ATLAS, and just a guess. But, with regards to the $50k:
Because being a young adult with $50k gives you way more options about how you can choose to act on the world in a way which being a broke student doesn’t. Having rich parents doesn’t really help if they aren’t in fact willing to give the $50k to the members.
When I reach back to my own experience, I spent some time I wish I could get back selling unexciting corporate software for money and giving math classes, to pay for living expenses while doing some of my early research. $50k would have afforded me with more significant freedom of action.
Note that the Thiel Fellowship, which aims to steer promising people towards being formidable forces for good in the world, gives students $100k. If you look at their notable recipients, and imagine a similar cohort producing a similar amount of impact in a less capitalistic direction, it seems at least conceivable that such a bet could be well worth it.
That’s my 2cts.
(Speculative, I don’t know how ATLAS worked beyond the public details, I don’t think I know any ATLAS fellows)
This seems like a good argument if:
you think ATLAS fellows wouldn’t counterfactually have $50k to spend on altruistic efforts
you think ATLAS fellows are in a similar reference class to Thiel Fellowship recipients
But I’m sceptical about both:
EA seems really excited about ambitious young people, and I’d expect that if someone was both really promising and had an idea for a high impact altruistic project which they couldn’t secure other funding for, it’d be relatively easy to get EA funding for it.
I expect the ATLAS team worked hard on outreach and publicity, but it’s a new initiative with a kinda-weird-to-normal-people framing. The Thiel Fellowship has been running for a while and gets a lot of publicity from Thiel and from successful recipients. I’d be surprised if ATLAS attracted a similar caliber of applicant in its first year of operations.
I guess given this, I’d be surprised if ATLAS funding made a big difference, and $50k is kind of a lot.
My experience with Atlas fellows (although there was substantial selection bias involved here) is that they’re extremely high calibre.
I also think there’s quite a lot of friction in getting LTFF funding - it takes quite a long time to come through I think is the main one. I think there are quite large benefits to being able to unilaterally decide to do some project and having the funding immediately available to do it.
“Living expenses while doing some of my early research” is one of the main purposes of the LTFF; to me Atlas feels like a roundabout way of getting that. LTFF asks you to have a specific high-impact project or educational opportunity for you to pursue, but as far as I know that wasn’t true of Atlas.
I think The Century Fellowship would make a better comparison to the Thiel Fellowship than Atlas would. It seems aimed at similar types to the Thiel Fellowship (college age people who are prepared to start projects and need to be financially independent to do so), while Atlas targets a slightly younger demographic and gives scholarships.
Atlas is posed as a talent search and development program, so I think any evaluation of Atlas should focus on how well it is searching for and developing talent that would not otherwise exist. I personally don’t know anything about how that has been turning out, or what the graduates have done/are doing with the money, so I don’t feel very qualified to evaluate it myself.
I eventually did get a grant from the LTFF, but it was once I had a more or less clear research direction. Idk, it’s possible that I could now write a good grant application for exploratory and independent thinking, but at the time I would probably have produced a very awkward ask.
Thank you, this response makes the most sense to me.
Edit: To clarify, this doesn’t mean I agree. I still believe the amount is too much. (See Edit 3 on the post for more details).
Cheers.
This sounds like a great reason to give young adults $50k, but if someone is giving young adults $50k for this reason, I don’t see why they would add a requirement that they be spent on “academic expenses”.
Fair point!
Urgh, I feel a bit like this is digging too much into people’s social lives, but I updated after the FTX situation that being more open about people’s relationship entanglements is probably good for the community (I was aware that Sam and Caroline were dating, and thought this was a pretty relevant aspect of modeling the FTX situation, but had norms against sharing that kind of information freely). I will experiment with being a bit more open about my knowledge here, though man, I don’t want this to set a huge precedent and currently do sure feel like I would want us to think more carefully about the norms here.
I think it’s accurate that there was/is a metamour-relationship between the Open Phil grant investigator for the grant and one of the Atlas co-founders. Pretty sure they are now no longer grant investigator on that grant and have handed it off to someone else (very likely because of the COI). I do think this kind of COI matters, though I do think second-degree things like this are much less important than first-degree relationships. I feel like the COI here has been handled fine, though I can’t speak with confidence (like, I don’t think it would be appropriate to have norms that force Open Phil employees to force their partners to not get close to anyone they were evaluating).
At a privacy level, I think (though man I am holding this very tentatively and am not intending to push back strongly in either direction) it’s better for the world if people feel comfortable discussing this stuff, while also taking some steps to not expose all of someone’s private and romantic life to the public. I think in the OP it would have been better to say “I have heard rumours that the primary grant investigator for Atlas had a metamour-relationship with one of the Atlas co-founders”, like I’ve done in this comment, which I think would have almost all the relevant info across while giving people a bit more privacy. But again, no strong push in either direction, and I would be interested in discussing the norms here more.
In-particular, I think outing someone as polyamorous can have consequences that are similarly bad to outing someone as gay, and can have pretty similar social stigma attached, and I think I am currently quite strongly in favor of not outing people who are gay without their consent.
I think we need a high bar of preemptively blasting a polycule graph onto the forum, and I don’t think the concerns in this post reach that bar.
If lack of polycule auditing has been bad for the movement, then we need a team who does these audits and shares them on a need to know basis.
I think being the grant investigator for a partner’s project would be worth addressing, though not sure if this happened here.
Fwiw I think metamour is much closer to like 1.5- order gap, or can be, than e.g. friend of a friend.
Sorry, I didn’t intend to imply that being a metamour is only as significant as being a friend of a friend. Seems definitely more intense/distortive/COIy to me.
Do you know if the decision to hand off this grant was prior to the relationship or after it?
I think the question of whether the relationship was ongoing during the time of the grant evaluation is important.
I know outing people who are gay can get them as far as murdered in places where homophobia is most rampant. On the other hand, polyamory seems to be a fairly popular model in the rich Bay Area community, often not kept secretive and without such repercussions. So I’m reserved whether this comparison is fair.
Yes, telling people in the Bay Area that someone is poly seems pretty fine, but e.g. many people are not openly poly towards their family, or have relatives in other states, and having you be easily outed as poly when googling you can have large adverse consequences for their relationships beyond the Bay Area.
To give a concrete example, my (non-EA) ex was from Europe, and she had a relative who both didn’t like that she had two partners, and that I was non-white. My understanding was that the “poly” dimension was seen as substantially worse than the racial dimension. The relative’s attitude didn’t particularly affect our relationship much (we both thought it was kind of funny). But at least in Western countries, I think your bar on outing poly people who don’t want to be outed should be at least as high as your bar for outing interracial couples who don’t want to be outed, given the relative levels of antipathy people in Western countries have between the two.
(I may want to delete this comment later).
This comment impressed me a lot. It seems humble, vulnerable and wise. To be able to see the good in something that you are uncomfortable with is at least a little inspiring. I need to improve at this!
My instinct is that this conflict of interest may be a bit worse than you are painting it, although that’s just instinct. It’s unclear and may not be written in a rulebook. I think with all the scrutiny on EA stuff at the moment, orgs should be drawing up specific, clear and public conflict of interest policies—if they don’t have them already.
I don’t know which Atlas Fellows you are friends with, but out of the ones I know, this doesn’t seem to be the case. The data also supports this: I am an Atlas Fellow and Jonas polled us on who would benefit from a full-ride scholarship. Out of 50-60 respondents, 7 answered that their parents are already paying for everything. 9 answered that they already have a full-ride scholarship, and 36 answered that getting a full-ride scholarship would “help with career” or allow them to “take out fewer loans.” In addition, 23 responded that if money wasn’t a concern, they’d be able to attend a college better than the one they’re currently attending or planning to attend.
Personally, my parents plan to give me $0 after I move out. We don’t earn that much. Even with Atlas money, I’d go into ~$150k debt if I go to a top US or UK school. I’m a senior and got accepted into Cambridge, but may not attend due to finances. I have multiple friends who are Atlas Fellows (mostly international students) that would be on the edge about attending a top US school if/when accepted, due to cost reservations. For many of us, the Atlas money frees us to take gap years, to pursue projects for a few months without worrying about rent, and from taking corporate jobs after graduation to pay back college debt. I’m sure I’m biased, as a recipient of the money. I’m sure that some Atlas Fellows, like you described, would not be heavily impacted by the $50k. But I hope to add another perspective.
I too was an Atlas Fellow and I agree with this comment. I personally am fortunate enough where I could afford college at full cost, even without Atlas. Despite this, the money from Atlas is allowing me to pursue opportunities that I definitely couldn’t have before. Earlier this year I was discussing taking a gap year to go work abroad for a company that I believe can greatly improve the world. I wasn’t sure if I would be paid, I knew my parents likely wouldn’t have given me money to travel and do this work for a year, whereas with Atlas, I had pretty high confidence I could use my money to make this possible. The scholarship is still helpful for students who could already pay for college.
This seems like a poorly worded question. Many people would be incentivised to say yes if there is even a 1% probability that answering this question may lead to existing fellows getting full-ride scholarships
Hmm I guess so, people who can’t afford college on their own are more likely to answer. Though don’t think that’s sufficient incentive to lie.
Out of fellows I’ve talked to about this, I can count 12 distinct people that can’t easily afford college and 5 that can. Some quotes:
Though I guess a few examples + the sample of fellows I’m close with isn’t sufficient evidence. How would you design a better question to gauge the percentage of fellows that can afford college comfortably from their parents’ income? If you want more data, I can perhaps poll the server.
Note: Written on mobile
Most of the questions here are on whether Atlas Fellowship is wasting resources. I think we should give projects some space to breathe before scrutinising rumours on their coffee tables. It hasn’t been much time since this project has been founded. Cofounders of Atlas Fellowship have a strong track record in creating a lot of value and their time is quite expensive. The team is only 4 people afaik. Probably their spending on physical goods is very low within their overall budget, so spending their time on cutting costs of physical goods isn’t worth their time. Sometimes spending your time on cutting costs is wasting resources because your time is more valuable.
You might ask “why they don’t hire another operations specialist to cut costs”. Unfortunately, hiring someone has many hidden costs explained in this blog post, so it’s usually more difficult than it sounds: https://blog.givewell.org/2013/08/29/we-cant-simply-buy-capacity/
Both the cofounders and the grantmakers of this project have a strong track record in raising tons of money. So if they think offering high schoolers 50k scholarships for a fellowship might create $500m+ impact in expectation, that seems to be a hypothesis worth testing out. Maybe the SPARC and ESPCR programs you mention are really better. But let’s just let the project to breathe a few years so we can really have data on this.
This project will also have verifiable, measurable results in a short amount time. So I think there are already strong accountability mechanisms on whether the project really delivered its promises. If they don’t deliver on their promises, we will find out that maybe their time was better spent cutting these costs. If they get strong results, we will discover that spending thrir time on improving other aspects of the project was better use of their time. So I think we will be in a much better place to discuss Atlas in a few years.
If you think the project is doing harm, that’s fair to discuss. I think it’s also fair to ask the grantmakers the rationale behind large grant decisions such as this one. But unless there is a suspicion of the project doing harm, I think it’s better to let project founders do their thing for some time before micromanaging all the small decisions they make. Funders/investors of startups don’t usually do that sort of thing and allow founders to execute their vision independently. It will be clear whether they were really effective when they will have to demonstrate results.
I recognise that not all of your questions are about wasting money, my response is not relevant to those questions.
I think there’s value in discussing ops decisions. They still represent how an org runs.
The most expensive coffee table I find in my local IKEA is $150, the cheapest $16. I wouldn’t quibble over the outright choice to buy the $150 one and agree at some point spending time comparing items to save cash is not a good use of time. I agree the difference saved between these tables is less than the value of some people’s time, sure.
But $10,000? That’s just unreasonable. That amount of money could save two lives and would be appalling spending if this is true. There has to be a cheaper alternative that can be found in little time. Say 20 minutes. Which would mean there is no one at Atlas whose time is below $30,000 an hour? Hard to believe.
I think I’m more sympathetic to some types of discussion than others.
If someone really established that Atlas spent 100x extra money on a coffee table than it was needed, then I think it would be fair for them to make sarcastic jokes about Atlas coffee tables in public or over the internet.
But I’m much more resistant to a culture where project founders are expected to respond to every similar rumour with a paragraph like “actually, it wasn’t a coffee table, it was a much larger piece of furniture for a bar, and it wasn’t actually $10000, it was $4000. Here is the screenshot of our digital receipt.”
If we establish a culture where every rumour about every similar product purchase is expected to receive an explanation from the relevant organisations, many projects would move much more slowly. If this wasn’t just a rumour, I think it’s more justified to criticise but I still feel resistant to the idea of founders being expected to explain all their weird-looking spending on physical goods that make up less than 0.1% of their budget.
I sure think the coffee table is a distraction, but in any case, I left a comment with the story of the rumored coffee table here (TLDR: It’s a really cool art-piece/coffee table that cost around $2200, Lightcone bought it from Atlas a few months ago at list price): https://forum.effectivealtruism.org/posts/fMrtoKBFK7p6oRHpu/atlas-fellowship-why-do-100-high-schoolers-need-usd50k-each?commentId=tPDmJWrimG9KpBCfi
I agree with most of this sentiment (and I don’t like sharing of unsubstantiated rumours). Once the rumour is out there though I don’t see the harm in responding to every rumour with a paragraph. There aren’t that many rumours flying around on the forum. It would take 5-10 minutes to put this particular rumour to bed. I don’t think this would slow down projects meaningfully to respond. Yes it might not make sense to explain tiny percentages of spending, but if that is where the scrutiny happens to fall why not just clear it up quicksmart?
Fortunately @Habryka was here to clear it up, but the head of the org could have also taken a few minutes to quash the rumour.
I don’t know the exact source of this quote, but I can offer a related quote from myself during a bunch of recent renovation work I’ve been coordinating.
When we first started working on renovating the hotel property Lightcone is currently working on, the contractors we had hired would generally make extremely budget-conscious decisions, calling us about a large number of individual decisions, checking in with us far too frequently, and when given a budget, acted with great risk-aversion on trying to desperately avoid going over the budget even a tiny bit. This ended up costing our core team a bunch of time, time that we valued at substantially above the savings that would usually result from this kind of checking in.
We tried for a while to figure out what the best way to communicate what we want is, and at least with some new contractors, the instruction that we have given a few times is “for this project you are working with an unlimited budget. We are primarily concerned with speed and quality, not how much money we spend. If you are considering making a purchase or a financial decision above $15k that we haven’t previously discussed, please briefly check in with me, though it will probably be fine, I generally trust you to make reasonable tradeoffs here.”
I was definitely quite concerned that this would cause us to drastically overspend, and it has caused some issues in one or two places, but I think overall the time savings of our core staff time, as well as the quality improvements and the speed improvements of the overall projects have been worth it (though we have gone somewhat over our intended overall budget for our renovation plans, and we’ve been doing more detailed budget-setting and accounting since the FTX collapse, so I also don’t think this kind of policy is obviously great).
I don’t have a ton of privileged information about how Atlas makes decisions here, but I at least could see someone walking away from Lightcone with the impression that we don’t believe in budgets, and I would be happy to defend our current policy.
Full disclosure: I am an Atlas Fellow, and after attending the first summer session, was hired by Atlas as a junior counselor to teach at the last 2 programs in the summer. I have no ongoing employment contract with Atlas, and am writing from my own personal beliefs. Most of these thoughts on the purpose of the scholarship were developed before I was hired, as I asked quite a few instructors/fellows “Do you think this is an effective use of $10m?”, and developed some of my own takeaways (these opinions are not directly those of any Atlas Instructor, as I don’t want to incorrectly echo someone else’s take).
I agree with most replies on this post, but I figured I’d give some thoughts as a fellow, aswell as answer a few questions people haven’t replied to yet.
Besides this being false (based on moderator comments), I think this comment + your point on students who can already pay for college, leads me to believe that you may misunderstand the point of the scholarship. The appeal/benefit of the scholarship to myself, as well as many other fellows I’ve spoken to is that it explicitly does not need to be spent on college tuition.
Nearly all large scholarships I’m aware of (both in and outside of EA) require you to spend the money on college tuition. Pursuing research, funding small projects, hiring tutors/experts for 1-1 learning, traveling to conferences, etc… are all things that are quite hard to do as a teenager in high school. Having $50,000 to go do these things and learn in the most optimal way for every student, is the benefit I see with the Atlas scholarship. This is true for students who could/couldn’t already afford college.
In the case that you were correct about this student who received the OP scholarship, I still would disagree with you that it is bad that they got both OP/Atlas funding, based on my reason above. I can think of a plethora of ways that student could greatly increase their trajectory outside of college, which again, the Atlas scholarship allows.
I’ve wondered quite a bit if Atlas was the best use of $10 million, and I support that you are posing this question.
I think the efficacy of Atlas can only be feasibly noted 5-10 years from now (possibly sooner, not super sure on time frame here), but I think in theory, Atlas’s effectiveness beats AMF.
In the instance of AI alignment, taking ~60 students (assuming 60 or so fellows meet this criteria) who are planning on majoring in Comp-sci/math/something adjacent, and exposing them to arguments around AGI risk, seems worth it. Afaik, there are <500 full time alignment researchers (if not many fewer), and this exposure may lead to a significant increase in future researchers in this field. This is notably hard to pin a value on, and it is hard to estimate how many fellows will actually turn into AI researchers. Given that the money is available before attending the program, fellows simply could not give a shit about the curriculum, but from what I personally observed, this wasn’t the case. Anecdotally, most of the people I’ve kept up with from the program have kept up their interest into AI risk (though this is subject to selection bias on who I’ve kept up with).
There are numerous other parts of the curriculum outside AI risk, and I think exposing young people who are about to enter their career to these ideas is valuable. I just use AI risk as an example here. Atlas’s thesis is not simply to birth AI researchers, so I don’t want to falsely convey this.
Even outside of fellows, the application process included analyzing an essay on AI risk, and the $50k scholarship incentivized applicants to truly comprehend the argument presented, and build a thoughtful counterargument. I think there is some value purely in this incentive, and given that you believe a high amount of applicants were recruited from “Olympiad communities worldwide”, I imagine a high percentage of these students had never deeply considered AI risk before.
I’m curious if others think this is valuable, but my main point is the high value of the scholarship leads to a high incentive to carefully digest and think about these readings.
From the projects I have seen Atlas fellows go on to pursue, I believe this past cohort of Atlas/Atlas India fellows will collectively save >2000 lives (assuming all $10m could have gone to AMF). I don’t have hard evidence for this claim, and it is frankly anecdotal.
Sydney left Atlas as a co-founder in mid-december. This was announced to all fellows then, so I imagine there was just some lag in removing her from the website.
The document addresses the reason fully.
Additionally the document notes that it shouldn’t be shared publicly. I think you ask some important and good questions in this post, but I think this combined with you unnecessarily outing an employee, not engaging with most comments that provide answers to your questions, and your spreading of unconfirmed rumors are all harmful norms to promote on a public forum. I think posts like this raise similar questions while not being harmful and fostering beneficial discussion, and I wish this post was written similarly.
To be fair, the request not to share the document publicly is not very prominent: it’s a non-bolded sentence halfway down the second page. It would be easy to miss if you hadn’t read the document carefully. (When I’ve written private documents in the past, I generally have it in red in the header or something similarly unmissable.)
I do think the moderators should probably remove the link. If we have decent evidence that the OP knew the document was private, I think that speaks very poorly of them and I’d support more draconian repercussions.
In general, there is no reason to expect the Atlas’ founders to spend money needlessly. Nobody is suspecting that they are spending it on themselves (excepting the alleged expensive table), and just like enterprises I expect them to be at least trying to use their resources in the most efficient way possible.
You raise imho valid arguments. To address some of your points:
I guess the Atlas Foundation is going off a model where impact is heavy tailed, in which it makes sense to spend what seems disproportionate resources on attracting the most talented. In such a model, attracting a fellow from the 99th “potential impact” percentile rather than 10 fellows from the 95th percentile would still be worth spending some marginal 45k for, even though it sounds excessive.
“From friends who are Atlas Fellows, they said many Atlas Fellows do not require the scholarship as their parents earn a lot and can already pay for college.” If true, this is evidence in favor of offering them such a ludicrous amount of money. They do not really need the money, so the marginal value is reduced and you need to offer more money to entice such potential students (or think of other benefits). And an unfortunate fact of life seems to be that a person’s financial earnings are highly correlated with those of their parents. Taking earnings as a proxy for potential impact means that a program like the Atlas Fellowship should also consider privileged students as people worth attracting.
And maybe that’s just me, but some of the phrasing comes off as somewhat combative (on the other hand I am aware that many people here think we should state our opinions more directly). As an example, the question in the title: “why do high schoolers need $50k each?” is not really truthful and sounds rhetorical, because nobody has claimed that the applicants need that money, just like high frequency traders do not need high compensation but still firms pay that amount to hire them.
I would usually not go around tone-policing, but I think it would be beneficial in controversial times like this to remember that as **a community we wanted to move away **from evaluating charitable initiatives based on how they sound and instead evaluate them on their results. In that vein, I do not think that it is helpful to quote rumoured single sentences by founders without any context (“not believing in budgets”) and without actually engaging with that sentence. The founders do not owe us accountability of private sentences that they might have uttered at some point.
Hits based giving means that Open Phil should not police the furniture of their grantees, and I am also unsure whether the way they manage inventory is indeed of public interest, as they are not soliciting donations from the public at the moment.
The steelman for these programmes is that they’re essentially talent arbitrage. A high-school students earning potential is very little at 18, but just 5 years later it could be a 6 figure sum. If someone’s earning potential is analogous to the amount of impact they could make, and you’reable to change the area they focus on, then you could imagine 50k USD is quite cheap (esp for influencing ones entire career).
The 50k USD here would be to attract the most talented students.
Some criticism of this theory of change:
the programme selects for privileged students who are unlikely to be value aligned with effective altruism
It’s really hard to influence someone’s career
the value of someone’s career is low
the values instilled by the programme are not aligned with doing good (or are actively harmful)
talented Americans are not neglected, and ATLAS is likely to get outcompeted by other status/wealthy corporations
Note: this comment doesn’t necessarily attempt to reflect my opinions. It’s also low effort/and written on mobile.
I don’t really know what this means. I’ve seen the Atlas venue stock rooms, which IMO seem reasonably stocked to me if you were expecting to use the venue on an ongoing basis.
I don’t know what it means to “have no inventory stock”, like, there clearly are storage rooms. And isn’t restocking the primary point of having an inventory stock in the first place? So of course if you lose an item you buy another one, what’s the alternative?
I meant that there is no database of what items have been ordered (and with quantities) meaning that, if someone wanted to, they could steal something (eg, a laptop) and it would likely go unnoticed.
(Speaking here as the Communications Officer for Open Philanthropy.)
In response to point #4: During the grantmaking process, the investigator for the initial Atlas grant followed our relationship disclosure policy. Our internal decisionmakers reviewed their disclosure and approved them to continue working on the grant.
I was pretty skeptical of Atlas when I first heard about it, and I’m still concerned about overspending, bad optics, attracting grifters etc. That notwithstanding, this post doesn’t even attempt to engage with what a programme with Atlas’s goals should look like, and instead just reads as a long list of speculative gesturing.
In part, this makes sense as part of identifying talent. It’s not obvious what the elasticity of grant money to attracted talent is, but it would be surprising if the elasticity is zero. As you note yourself, however, the money is also there for “academic expenses”. Presumably identifying top talent is only an instrumental goal, and it is just as important to help said talent utilise their skills once you find them?
For many teenagers, they are likely financially constrained and this sort of funding can let them attend academic conferences or pay for textbooks etc. While it is true that they could in principle apply to EAIF or LTFF for these sorts of expenses, the website suggests that Atlas evaluates “the use of scholarship funds … on a case-by-case basis”, so this wouldn’t be carte blanche for Atlas fellows to spend, and instead merely represents the diversification of funders within the community, which I see as a good thing.
In any case, my best guess is that the discretionary nature of this spending means that topline number of $50k is an upper bound which does not reflect the median Atlas fellow.
Of course, this leads to the next consideration:
Would it?
I personally would be in favour of a smaller amount of money (somewhere in the range of $10k to $30k is my best guess), but I don’t have good data on this. My impression is that many similar grant schemes offer comparable amounts of money (see: 776 Fellowship, O’Shaughnessy Fellowship, Thiel Fellowship, Century Fellowship) and those which offer less money often do so because they are aiming to provide funding for a much more limited time interval.
Given you seem to know a lot about optimal models of funding, I’d love to see the data you have which makes you believe this!
In a similar vein to my previous comment, I’d be curious where you’re getting your data from, and would love if you could publicise the survey you must undoubtedly have done to come to these conclusions, since you are not yourself at Atlas fellow!
I’d also point out that if the average Atlas fellow does not require the money, then the money needed to attract them is higher, so this point and the previous point seem in tension.
The document you link seems to address this, when it says that “we suspect that for many Atlas Fellows, the things holding them back from whatever ambitious and genuinely impactful projects they’re most interested in is a lack of deep, specialized technical knowledge—and the strict requirement to spend 8 hours in a classroom every day”.
I’m not sure what the goals of the organisers of Atlas are, but judging by the funding structure and pop-up programmes like this, my guess is that they’re not aiming for just talent search but also helping nurture the talent they find, which explains why they may diverge with programmes you’ve been involved with like SPARC.
As a final note, I find it a bit disconcerting that this post ends by speculating about the relationship status and preferences of people in the community, and I generally think that organising events or making grants does not represent consent in having people gossip about your personal life.
Strong downvoted for hostile tone, e.g. “In a similar vein to my previous comment, I’d be curious where you’re getting your data from, and would love if you could publicise the survey you must undoubtedly have done to come to these conclusions, since you are not yourself at Atlas fellow!”
The sarcastic tone is very patronising here. I suggest rewording this.
My priors come from,
non-trivial offering approx. $500 scholarship.
Stripe Atlas offering almost nothing but still attracting many exceptional teenangers.
Emergent Ventures offering grants on the order of $1,000s but still attracting exceptional people.
I find it curious you fail to engage with any of the other responses, and consider your time better spent on looking for dirt on people’s Facebooks and Linkedins, especially since you claim:
That notwithstanding, I was not attempting to be patronising. I simply expected people (especially SPARC alum) who make strong claims on the forum to have a strong evidence base backing those claims up—for example, you question the validity of internal Atlas polling on the basis of respondents lying and bad survey methodology.
I’m assuming based off of your reply that your beliefs are influenced more by your priors than by having well-evidenced research or data on this, which is fine, but in that case, I think it would have been useful if you had flagged that level of epistemic confidence or rephrased your assertive claims in the original post!
On the object level, none of these seem to be great priors. There’s no evidence for what non-trivial’s scholarships have attracted, it’s not clear that any of the projects you listed in the twitter thread about Stripe Atlas are the sort of thing Atlas would be trying to attract and in the case of Emergent Ventures, the “s” is doing a lot of work for you, since most grants are significantly higher than $1,000 and closer to the order of $10,000.
There is a difference between failing to engage and choosing not to engage.
Your tone is very patronising here. I suggest rewording this.
I’m sceptical of programs like this for the following reasons:
Giving $50k is an anti-filter for hardcore-ness and willingness to make it work in EA/Impact (I agree with the comment that a smaller amount would likely do the same job)
EA already has a somewhat high bounce rate, I suspect at younger ages it’s even higher, I understand the logic of needing only one hit or homerun to make back all the EV but actually is a $50k grant making a difference on this front anyway, it seems likely there are better ways to incentives people, Charity Entrepreneurship seems to of succeed in finding talent and starting charities despite minimal funding and financial incentives, maybe something like a worldwide talent search focused on underutilised regions and geographies make more sense (The India aspect seems good)
The coffee table thing seems indicative of poor decision-making and optics, never has my thought process been in any way impacted or influenced by a piece of furniture (Perhaps you can argue for the overall environment design mattering, but generally I think if this is an issue then there are broader issues to think about and tackle in EA) there’s definitely something purer about a minimalistic and “Cheaper” environment. (Also I’m sure a $200 coffee table can also look 99% as good)
Disclaimer: I’m very sorry if my quick thoughts and naive perspective derailed anywhere and hurt anyone’s sentiments. Not intended to cause controversy.
I think it’d be great to inculcate something that Open Phil does with their undergraduate scholarships for internationals aspiring to study in US and UK universities, that is include the factor of “not residents/citizens of US and UK” in some way. I mean to say the present Atlas International could be redefined to cover most of the world including Central and West Africa, South America, India, China, etc, with say $20k scholarships. And instead of other subdivisions like Atlas India and Atlas China rather launch an Atlas+ for the US, Canada, UK, and remaining regions, with say $10k scholarships. This could help tackle the privilege thing you mentioned.
I think the privilege factor rightly comes in not only in terms of the rich students getting the scholarships but also in terms of who is having the opportunity of having an actually ‘good education’, which again falls in the aspect that certain schools and/or institutions which only certain students can afford to attend, get better writing and articulation skills. I’m saying this because though Atlas says it has a strict no-BS policy for answers to the questions of their applications, I firmly believe the answers need to be written, approached, formed, and framed in a certain way to even get the minimum attention/consideration while admissions review by the staff. I’m not saying faking or lying at all, just that you need to mold and present your answer and thoughts in a way that has to certainly incline with the thought processes of Atlas people that you gauge from their site and available information. And this is pretty similar to the Common App US colleges process. I’m an Atlas finalist as well as a rationality camp alum and didn’t fake anything anywhere but this writing style thing/skill one has to implement which generally is better for students of elite schools.
With all these being said, I don’t think Atlas should be discontinued though, like EA as a whole has flaws and improvements to be made in the community, Atlas too can surely grow and improve upon. And if $10M+ could be spent on PR for a book, certainly giving even $50k to 100 high school students each for their experimentation sounds good enough as it also might lead to some good results/nice outcomes.
How big a barrier is this in practice? Do all the fellows spend all the money and are “finding” ways to make sure they do? Or are they rarely hitting the $50k limit and actually spending this money positively?
Atlas Fellow here. Amongst casual conversations, I get the vibe that most of us have barely dipped into our scholarships, given that it’s been less than a year. Many fellows I know (including myself) haven’t touched the money yet. Some have spent a few thousand. Some have used it to pay for their first semester of college. There’s no reason to “find” ways to spend money as the money doesn’t expire—if I don’t need money now, I can defer withdrawal to a year down the line, or whenever I need it in my future career.
Epistemic status: Very Uncertain
It’s an interesting question whether orgs should respond to criticism posts like this.I think the benefits of responding possibly outweigh the harms.
Maybe higher experienced and knowlegable EAs feel comfortable to let these kind of posts slide, as they trust the orgs and people involved. Less experienced Forum peruses (people like me) who have less knowledge and EA background might find these kinds of posts disturbing and feel like a response is warranted to clear the air. For example even Habryka’s reply about the coffee table was fantastic and basically quashed that criticism. I feel it would have been even better if it had come from the org’s leaders.
It’s not only this post about Atlas. OpenPhil and others haven’t responded in depth to serious criticisms recently, nor have they clearly stated why not.
I Have listed potential benefits and harms of responding to weigh below. It’s hard to weight benefits vs harms,
Potential Benefits of responding
- Quashing rumours early and quickly
- Answering reasonable questions to set minds at ease. For example in ths case Explaining the EV calculations of why 50k was given wouldn’t be very hard I don’t think. It must be written somewhere, I would imagine they could probably almost copy paste the reasons.
- Give confidence both within and outside the EA community that orgs are open to criticism and engaging with it.
Potental Harms of responding
- Dragging out drama and making it even more higher profile by continuing unhelpful or even toxic conversations.
- Giving weak criticisms increased airtime by responding rather than ignoring them
- Making whoever responds a target for further grilling and stress.
- Monetary/Time costs spent responding (I think these are often overstated though and can be easy to hide behind)
I’m sure there are many more benefits and harms as well. Although my instinct is that early, in-depth replies is the best way to respond I’m very uncertain.