(All writing on this account is personal and not endorsed by my employer.)
I work on the Navigating Transformative AI team at Coefficient Giving, as Luke Muehlhauser's Chief of Staff. Previously I worked at GiveWell in recruiting, ops, and Chief of Staff roles; and I started my career as a high school math teacher. I learned about EA in 2014 when I stumbled on Scott Alexander's blog.
If you'd like to see more of my personal writing, my Substack is linked above.
hey Ana, nice to meet you! :)
here are six quick thoughts, and I want to apologize in advance that these are extremely rushed and unedited -- if anything here seems super off, lmk and I'll either clarify or retract. Also these points are not connected to each other; they're all separate thoughts.
Agree!
Yeah it's reasonable to bring up the possibility, though in my case I don't think it's a compelling narrative. I actually feel pretty embarrassed about my switching time, and I stand by what I said in the main post: Given repeated signals that it would be possible for me to make a switch, I should have been less scared and stuck!!!
Of the people I've spoken to about their own career switches:
What they all share is the same experience of stuckness, and that's what I'm trying to write about w/ this post.
Yes, I think this is right. I'm not trying to address all problems people might face vis-a-vis switching careers to AI safety. And more specifically, I'm not trying to address the particular type of problem where someone is trying hard to switch into an AIS role but is having trouble doing so. Instead I’m trying to address a different problem, where some people aren’t trying hard because they’re stuck with emotional blockers.
Hopefully it'll be clear to people not experiencing the problems I'm addressing that they should ignore my advice!
yo, I will respond more later, at which point I’ll delete this comment and upload a new one. I don’t have time to write something careful at the moment, which is why I should not enter Forum discussions in general. 🤦♂️
thanks for your reply!
Tony, thank you for writing this post, and for your work at Fortify! I really enjoyed reading, and expect I'll come back for more re-reads later.
I don't think the framing offered in this comment is useful because it relies on the assumption that GiveWell's expected grantmaking is fixed at [whatever amount has been currently announced by GiveWell], and that expected grantmaking would not further increase with additional funds. I'm not saying that people should definitely donate to GiveWell (more on that below), just that this particular argument for not doing so doesn't hold up.
It's probably worth examining the historical parallel here:
There's certainly a point at which GiveWell would reach full funding saturation based on available opportunities and staff capacity, but I don't think there's any basis to assume that the saturation point happens to be $1B, especially since GiveWell has asserted that it is funding-constrained and generally has a strong track record of being truthful about its sincere beliefs.
If I were currently deciding where to donate in global health, I expect I wouldn't try to guess at GiveWell's absorptive capacity and would instead use these frames to think through the decision:
When I've done this kind of thinking in the past, I've ultimately settled on giving to CE/AIM for my GHD giving - that's mainly because I tend to favor catalytic investments and hits-based work in my personal giving, and because I'm cripplingly contrarian and easily influenced by arguments that big institutional actors are missing things. I could definitely see myself giving in the future to e.g. the Mid-stage Global Health Fund or another similar funder!
But fwiw, I think that the question of "Where can my dollars have the most impact in global health" is not at all clear, including because many of the questions about ecosystem effects, funder diversity, etc. are not settled at all and because we don't know yet how GiveWell will behave at large funding levels. I'd be very interested in seeing more public discussion of those issues, and I think it's not a good idea to make decisions based on probably-wrong heuristics about GiveWell's likely absorptive capacity.
(I work at GiveWell and my thinking is inevitably informed by that experience, but the specific claims I'm making here rest on public information and personal opinions, not insider information - I'll keep it that way if this thread continues. I also don't donate to GiveWell right now, and haven't since I started working there: I currently give mostly to animal welfare and GCR opportunities, and I've donated ~20% of my likely 2026 giving to CE/AIM)
Cool discussion, thanks for sharing it!!
GiveWell has internal documentation about our compensation philosophy. I'm not going to quote it directly, but the gist is: "We don't want to overpay because that would be financially irresponsible and we don't want to underpay because that would inhibit our ability to execute our mission." Like Coefficient Giving (see Sam Anschell's comment), we benchmark our compensation to organizations that we think (1) do similar work and (2) try to hire similar talent. Our policy as written doesn't consider the spillover effects of our approach to compensation (e.g. its effects on the optimality of talent allocation between impact-focused organizations).
It's also worth noting that the unique setup of our retirement benefits (we increase salaries in lieu of offering e.g. a % match on 401k contributions) makes total compensation a better cross-org comparator than salary. But, I'm just making a pedantic point here, not trying to sneakily imply that GiveWell's total compensation isn't higher than many impact-focused organizations - it is!
@Ben_West🔸 this is not really a substantive comment, but just wanted to say that this is the only time I've seen someone mention Tolstoy's The Kingdom of God is Within You (imo one of the all-time great theological tracts) online or in real life—that was pretty exciting. Thanks for writing this; I really enjoyed the read!
One comment on a trivial thing:
You said that Satan's temptation of Jesus is not typically interpreted as meaning that governments are intrinsically Satanic. I think that's right. But, I don't think it's correct to say that the Bible clearly states that civil authority is intrinsically godly.
Your quote from the Cambridge Bible argues this, but their reasoning proves too much: If God is sovereign, you could use the same argument to say that everything (including sin) is divinely mandated because it's all created by God. Some Christian theological factions do actually hold that view–for example, by saying that God created sin in order to demonstrate His power and glory– but most do not.
I think the mainstream Christian view, sourced from a variety of Scriptures (but especially from Jesus' comments on Roman taxes in Mark 12) is something like "civil authorities are somewhat morally neutral; you should default to obeying their commands unless they're clearly in opposition to God's commands."
My best guess is that Tolstoy and many other Christian anarchists would actually agree with the precise logic of that statement, but where they'd differ from mainstream Christians is in the their interpretation of what's "clearly in opposition to God's commands". They'd argue that the set of civil commands that are contrary to God's commands is just very large—maybe including all civil commands—because such commands enabled the governments in their contexts to do ungodly things.
But, it's complicated; I think the debate in Christian theological circles about how to understand the morality of institutional behavior is very much alive!
Looking forward to attending!