My current impressions on career choice for longtermists
This post summarizes the way I currently think about career choice for longtermists. I have put much less time into thinking about this than 80,000 Hours, but I think it’s valuable for there to be multiple perspectives on this topic out there.
Edited to add: see below for why I chose to focus on longtermism in this post.
While the jobs I list overlap heavily with the jobs 80,000 Hours lists, I organize them and conceptualize them differently. 80,000 Hours tends to emphasize “paths” to particular roles working on particular causes; by contrast, I emphasize “aptitudes” one can build in a wide variety of roles and causes (including non-effective-altruist organizations) and then apply to a wide variety of longtermist-relevant jobs (often with options working on more than one cause). Example aptitudes include: “helping organizations achieve their objectives via good business practices,” “evaluating claims against each other,” “communicating already-existing ideas to not-yet-sold audiences,” etc.
(Other frameworks for career choice include starting with causes (AI safety, biorisk, etc.) or heuristics (“Do work you can be great at,” “Do work that builds your career capital and gives you more options.”) I tend to feel people should consider multiple frameworks when making career choices, since any one framework can contain useful insight, but risks being too dogmatic and specific for individual cases.)
For each aptitude I list, I include ideas for how to explore the aptitude and tell whether one is on track. Something I like about an aptitude-based framework is that it is often relatively straightforward to get a sense of one’s promise for, and progress on, a given “aptitude” if one chooses to do so. This contrasts with cause-based and path-based approaches, where there’s a lot of happenstance in whether there is a job available in a given cause or on a given path, making it hard for many people to get a clear sense of their fit for their first-choice cause/path and making it hard to know what to do next. This framework won’t make it easier for people to get the jobs they want, but it might make it easier for them to start learning about what sort of work is and isn’t likely to be a fit.
I’ve tried to list aptitudes that seem to have relatively high potential for contributing directly to longtermist goals. I’m sure there are aptitudes I should have included and didn’t, including aptitudes that don’t seem particularly promising from a longtermist perspective now but could become more so in the future.
In many cases, developing a listed aptitude is no guarantee of being able to get a job directly focused on top longtermist goals. Longtermism is a fairly young lens on the world, and there are (at least today) a relatively small number of jobs fitting that description. However, I also believe that even if one never gets such a job, there are a lot of opportunities to contribute to top longtermist goals, using whatever job and aptitudes one does have. To flesh out this view, I lay out an “aptitude-agnostic” vision for contributing to longtermism.
Some longtermism-relevant aptitudes
“Organization building, running, and boosting” aptitudes
Basic profile: helping an organization by bringing “generally useful” skills to it. By “generally useful” skills, I mean skills that could help a wide variety of organizations accomplish a wide variety of different objectives. Such skills could include:
Business operations and project management (including setting objectives, metrics, etc.)
People management and management coaching (some manager jobs require specialized skills, but some just require general management-associated skills)
Executive leadership (setting and enforcing organization-wide goals, making top-level decisions about budgeting, etc.)
Fundraising and marketing
Assistant and administrative work
Corporate communications and public relations
Finance and accounting
Beth Jones (Open Philanthropy Director of Operations); Max Dalton and Joan Gass at CEA; Malo Bourgon at MIRI. (I focused on people in executive roles and gave only a small number of examples, but I could’ve listed a large percentage of the people currently working at longtermism-focused organizations, as well as people working at not-explicitly-longtermist organizations doing work that’s important by longtermist lights. In general, my examples will be illustrative and focused on relatively simple/”pure” cases of someone focusing on a single aptitude; I don’t think people should read into any “exclusions.”)
How to try developing this aptitude:
There are many different specializations here. Each can generally be developed at just about any organization that has the corresponding need.
In many cases, early-career work in one specialization can give you some exposure to others. It’s often possible to move between the different specializations and try different things. (The last three listed—communications, finance/accounting, and law—are probably the least like this.)
I’m especially positive on joining promising, small-but-growing organizations. In this sort of organization, you often get a chance to try many different things, and can get a rich exposure to many facets of helping an organization succeed. This can be an especially good way to get experience with people management and project management, which are often very generally applicable and in-demand skills across organizations. Coming into such a company in whatever role is available, and then being flexible and simply focused on helping the company succeed, can be a good learning experience that helps with both identifying and skilling up at good-fit aptitudes.
As a first pass, the answer to “How on track are you to develop a longtermism-relevant aptitude?” seems reasonably approximated by “How generically strong is your performance?” Raises, promotions, and performance reviews are all data points here. I think one of the best indicators of success would be that the people you work most closely with are enthusiastic about you and would give you a glowing reference—combined with those people (and the organization you work for) being themselves impressive.
People working on this aptitude might sometimes have feelings like “I’m performing well, but I don’t feel I’m contributing to a great mission.” In early career stages, for this aptitude, I think performing well is more important than being at an organization whose mission you’re enthusiastic about, assuming the work is overall reasonably enjoyable and sustainable. Later on, when you have a relatively stable sense of your core competencies and aren’t growing rapidly, I think it’s good to give the mission more weight.
Political and bureaucratic aptitudes
Basic profile: advancing into some high-leverage role in government (or some other institution such as the World Bank), from which you can help the larger institution make decisions that are good for the long-run future of the world.
While organization-supporting aptitudes are mostly (in the long run) about helping some organization whose mission you’re aligned with accomplish its existing goals, political and bureaucratic aptitudes are more about using a position of influence (or an influential network) to raise the salience and weight of longtermist goals within an institution.
Essentially any career that ends up in an influential position in some government (including executive, judicial, and legislative positions) could qualify here (though of course some are more likely to be relevant than others).
Richard Danzig (former Secretary of the Navy, author of Technology Roulette); multiple people who are pursuing degrees in security studies at Georgetown and aiming for (or already heading into) government roles.
How to try developing this aptitude:
First, you should probably have a clear idea of what institution (or set of institutions) could be a good fit. A possible question to ask yourself: “What’s an institution where I could imagine myself being relatively happy, productive, and motivated for a long time while ‘playing by the institution’s rules?’” I’d suggest speaking with later-career people at the institution to get as detailed a sense as possible of how long it will take to reach the kind of position you’re hoping for; what your day-to-day life will be like in the meantime; and what you will need to do to succeed.
Then, you can try for essentially any job at this institution and focus on performing well by the institution’s standards. Others who have advanced successfully should be able to give a good guide to what these are. In general (though not universally), I would expect that advancing along any track the institution offers is a good start, whether or not that track is directly relevant to longtermism.
Sometimes the best way to advance will involve going somewhere other than the institution itself, temporarily (e.g., law school, public policy school, think tanks). Graduate schools present the risk that you could spend a long time there without learning much about the actual career track itself, so it may sometimes make sense to try out a junior role, see how it feels, and make sure you’re expecting a graduate degree to be worth it before going for the graduate degree.
As a first pass, the answer to “How on track are you?” seems reasonably approximated by “How quickly and impressively is your career advancing, by the standards of the institution?” People with more experience (and advancement) at the institution will often be able to help you get a clear idea of how this is going (and I generally think it’s important to have good enough relationships with some such people to get honest input from them—this is an additional indicator for whether you’re “on track”). If you’re advancing and performing well generally, the odds seem reasonably good that you’ll be able to advance in some longtermism-relevant part of the institution at some point.
I think one of the main questions for this sort of aptitude is “How sustainable does this feel?” This question is relevant for all aptitudes, but especially here—for political and bureaucratic roles, one of the main determinants of how well you advance is simply how long you stick with it and how consistently you meet the institution’s explicit and implicit expectations.
“Conceptual and empirical research on core longtermist topics” aptitudes
Basic profile: helping to reach correct substantive conclusions on action-relevant questions for effective altruists, such as:
Which causes are most promising to work on? (This could include things like making the case for longtermism)
What’s a reasonable probability distribution over things like (a) when transformative AI will be developed; (b) the size of various existential risks?
What can we learn from historical cases about the most promising routes to growing the effective altruist community?
What policy changes would be most desirable to push for in order to reduce existential risk?
How should money be allocated between potential grantees in a given cause (or generally)? (And how should it be allocated across time, i.e., “giving now vs. giving later?”)
What sorts of jobs should effective altruists be most encouraged to aim for?
I discuss this one at some length because I know it fairly well. However, I think it’s one of the hardest aptitudes to succeed at at the moment, as it tends to require very high levels of self-directedness.
Eliezer Yudkowsky, Nick Bostrom, and others who have worked to flesh out the case for prioritizing existential risk and AI safety in particular.
Most people in research roles at FHI.
Most people in research roles at Open Philanthropy (you could also think of the grantmaking roles this way).
People working on determining what 80,000 Hours’s substantive recommendations and advice should be (as opposed to how to communicate it).
Note that some people in this category do mostly conceptual/philosophical work, while some do mostly empirical work; some focus on generating new hypotheses, while others focus on comparing different options to each other. The unifying theme is of focusing on reaching substantively correct conclusions, not on better communicating conclusions others have reached.
How to try developing these aptitudes:
One starting point would be a job at an organization specifically focused on the type of question you’re interested in. So if you want to look for crucial considerations, you might try for a job at FHI; if you want to work on questions about grantmaking, you might try for a job at Open Philanthropy.
I think other jobs are promising as well for developing key tools, habits, and methods:
Academic study in fields that are relevant for the kinds of questions you want to work on. It’s hard to generalize very far here, but for conceptual questions, I think philosophy, mathematics, computer science, and theoretical physics are especially promising; for more empirical questions, economics seems most generally promising, since fluency with quantitative social science seems important. (Many other areas, such as history and political science, could be useful as well.)
Jobs that heavily feature making difficult intellectual judgment calls and bets, preferably on topics that are “macro” and/or as related as possible to the questions you’re interested in. There are some jobs like this in “buy-side” finance (trying to predict markets) and in politics (e.g. BlueLabs).
I also think there are opportunities to explore and demonstrate these aptitudes via self-study and independent work—on free time, and/or on scholarships designed for this (such as EA Long-Term Future Fund grants, Research Scholars Program, and Open Philanthropy support for individuals working on relevant topics).
I think these aptitudes currently require a lot of self-direction to do well, no matter where you’re doing them, so trying them on your own seems like a reasonable test (although given the difficulty, I’d suggest a frame of “seeing whether this is enjoyable/interesting/useful” rather than “actively pushing for a job”).
The basic formula I see for trying out these aptitudes for self-study is something like:
Examine some effective-altruism-related hypothesis or question, and get very deep into it, forming your own “inside” view (a view based on your own reasoning and logic, rather than reasoning based on what others believe).
Write up your view with strong reasoning transparency, somewhere such as LessWrong or the EA Forum.
Engage in discussion.
Some example approaches:
Closely and critically review some piece of writing and argumentation on longtermist topics. This could be a highly influential piece of writing such as Astronomical Waste, On the Overwhelming Importance of Shaping the Far Future, some chapter of Superintelligence or The Precipice, various pieces on AI timelines, etc. Or for an easier start, it might be a recent post from the EA Forum, AI Alignment Forum, LessWrong, or a particular blog you think talks about interesting topics. Explain the parts you agree with as clearly as you can, and/or explain one or more of your key disagreements.
Pick some question such as “What are the odds of existential catastrophe this century?” (very broad) or “What are the odds of nuclear winter this century?” (narrower, likely more tractable). Write up your current view and reasoning on this, and/or write up your current view and reasoning on some sub-question that comes up as you’re thinking about it.
Look into some question that’s been explicitly flagged as a “question for further investigation” (e.g. here or here). Try to identify some sub-question you can shed some light on, and write up what you find.
It could also be beneficial to start with somewhat more concrete, tractable versions of this sort of exercise, such as:
Explaining/critiquing interesting arguments made on any topic you find motivating to write about.
Making bets and/or forecasts on PredictIt, GJOpen or Metaculus and explaining your thinking.
Writing fact posts.
Closely examining and explaining and/or critiquing GiveWell’s recommendations and cost-effectiveness analyses, or GPI’s papers (all of these are publicly available and tend to explain their reasoning thoroughly).
Reviewing the academic literature on any topic of interest and trying to reach and explain a bottom-line conclusion (there are many examples of this sort of exercise in Slate Star Codex’s “more than you wanted to know” tag; researching medical questions of personal interest can be an easy way to find topics).
In general, I think it’s not necessary to obsess over being “original” or having some new insight. In my experience, when one tries to simply write up one’s current understanding in detail—even when one’s understanding is a very “vanilla” or widely accepted story—points of confusion and uncertainty often come into relief, and one often can learn a lot and/or notice underappreciated points this way. I think it’s ideal to write up underappreciated points when one has them in mind, but I also see a lot of value in straightforward, detailed explanations and critical assessments of existing arguments.
Some example milestones you could aim for while developing these aptitudes:
You’re successfully devoting time to this and creating content. (I expect this to be the hardest milestone to hit for many—it can be hard to simply sustain motivation and productivity given how self-directed this work often needs to be.)
In your own judgment, you feel you have made and explained multiple novel, valid, nontrivially important (though not necessarily earth-shattering) points about crucial longtermist topics.
You’ve gotten enough feedback (upvotes, comments, personal communication) to feel that at least several other people (whose judgment you respect, and who put serious time into thinking about these topics) agree.
You’re making meaningful connections with others interested in these topics—connections that seem likely to lead to further funding and/or job opportunities. This could be from the organizations most devoted to your topics of interest; there could also be a “dissident” dynamic in which these organizations seem uninterested and/or defensive, but others are noticing this and offering help.
My very rough impression/guess is that for people who are an excellent good fit for this aptitude, a year of full-time independent effort should be enough to mostly reach these milestones, and that 2-3 years of 20%-time independent effort (e.g., one day per week) should also suffice. (For this kind of role, I think there’s a lot of important “background processing” of ideas, so I’d expect a 20%-time year to be more than 1⁄5 as productive as a full-time year.) I would generally consider this “clock” to start as soon as someone is carving out time and forming an intent to try this work (I wouldn’t wait until they are successfully spending time on it, since this is one of the most challenging things about the work, as noted above).
Contrast with research “paths.” Rather than aiming to work on a particular topic such as AI governance or cause prioritization, I’m suggesting starting with whatever topics you have the energy and interest to write about, and I think that someone who succeeds by the above criteria has a good shot at building a career around research on some topic in the general vicinity. Because of this, it should be possible to try/explore this aptitude without needing a particular job offer in a particular area (although again, I think the success rate will generally be low).
Basic profile: helping to communicate key, substantively well-grounded messages and ideas to particular audiences. The audiences could be very general (e.g., writing for mass-market media) or more specialized (e.g., writing for policymakers on particular issues). Example messages could be the importance of global catastrophic risks, the challenge of AI alignment, the danger of covert state bioweapons programs, the general framework of effective altruism, and many more.
Kelsey Piper and other journalists at Future Perfect.
Authors of mass-market books such as The Alignment Problem.
People who do social media and/or podcasting, e.g. Julia Galef and Rob Wiblin.
People working at think tanks, whose main goal is to put key ideas in terms that will be more compelling to particular policymaking audiences.
How to try developing these aptitudes:
First, you should have some idea of what sort of target audience you’d like to communicate with. A possible question to ask yourself: “What’s a type of person that I understand and communicate with, better than most EAs / longtermists do?”
Then, you can try to get any job that involves communicating with this audience and getting feedback on a regular basis—whether or not the communication is about EA/longtermist topics. The main aptitude being built is general ability to communicate with the audience (although understanding of EA/longtermist topics will be important at some point as well). So if you’re interested in communicating with fairly general/widespread audiences, most jobs in journalism, and many in public relations and corporate communications, would qualify.
I also think there’s a lot of opportunity to build this sort of aptitude through independent work, such as blogging, tweeting, podcasting, etc. I expect that some of the people with the greatest potential as communicators are those who find it relatively easy to create large amounts of content and connect with their target audience naturally. (Though for anyone doing independent public work, I’d advise taking some measures to avoid publishing something unintentionally offensive, as this could affect your career prospects for a long time even if the offense is the result of a misunderstanding.)
As a first pass, the answer to “How on track are you to develop a longtermism-relevant ‘communicator’ aptitude?” seems reasonably approximated by “How generically successful are you by the standards of the (communications-focused) career track you’re on?” The more successful, the better position you’ll likely be in at some point to find ways to communicate important longtermist ideas to your target audience.
Building a following via independent content creation would also be a clear sign of promise.
In both cases, it seems realistic to get a pretty good read on how you’re doing within 2-3 years.
Basic profile: founding, building, and (at least for some time) running an organization that works on some longtermist goal. Some people found organizations primarily as a way to have independence for their research or other work; here I am instead picturing someone who is explicitly aiming to invest in hiring, management, culture- and vision-setting, etc. with the aim of building an organization that can continue to function well if they leave.
(Not all organizations are founded by someone who is explicitly focused this way; sometimes an organization is founded by one person, but a lot of the “entrepreneur” work ends up done by people who come in later and take the top executive role.)
Some pretty clean examples (with the organization that was founded in parentheses, regardless of whether the person is still there) would be Ben Todd (80,000 Hours); Jason Matheny (CSET); Elie Hassenfeld and myself (GiveWell). Many other longtermist organizations had a fair amount of early turnover at the top (leaving it somewhat unclear who did the bulk of the “entrepreneur” work) and/or are academic centers rather than traditional organizations.
How to try developing this aptitude:
Entrepreneurship tends to require juggling more duties than one can really learn how to do “the right way.” It crucially relies on the ability and willingness to handle many things “just well enough” (usually with very little training or guidance) and focus one’s energy on the few things that are worth doing “reasonably well.”
With this in mind, I generally think the person best-suited to found an organization is the person who feels such strong conviction that the organization ought to exist (and can succeed) that they can hardly imagine working on anything else. This is the kind of person who tends to have a really clear idea of what they’re trying to do and how to make the tradeoffs gestured at above, and who is willing and able to put in a lot of work without much reliable guidance.
So my general approach to entrepreneurship would be: if there’s no organization you have a burning desire to create (or at least, a strong vision for), it’s probably not time to be an entrepreneur. Instead it could make more sense to try for a job in which you’re learning more about parts of the world you’re interested in, becoming more aware of how organizations work, etc. - this could later lead to identifying some “gap in the market” that you’re excited to fill.
I do think that if you have any idea for an organization that you think could succeed, and that you’d be extremely excited to try to create, giving this a shot could be a great learning experience and way of building a general “entrepreneur” aptitude. (Edited to add: and potentially having impact in other ways, such as via philanthropy—see discussion below.) This is true even if the organization you have in mind does not do longtermist-focused work (for example, if it’s a conventional tech startup). Though it’s worth keeping in mind that it could take a long time (several years, sometimes >10 years) to get a successful organization to the point where one can responsibly step away and move onto something else.
In the first couple of years, I think you’re doing reasonably well if your organization is in a reasonable financial position, hasn’t had any clear disasters, and has done pretty well at attracting talent. Beyond that, I think it tends to be a big judgment call how an organization is doing.
“Community building” aptitudes
Basic profile: bringing together people with common interests and goals, so that they form a stronger commitment to these interests and goals and have more opportunities and connections to pursue them. This could be via direct networking (getting to know a lot of people and introducing them to each other); meetups and events; explicit recruiting; etc. Referring new people to resources and helping them learn more is also an important component.
People organizing local, university, etc. EA groups, organizing EAGx’s, etc., as well as many of the people at the Centre for Effective Altruism.
How to try developing this aptitude:
There is likely some community you’re a part of, or set of people you know, that you can immediately start working with in this way: networking and making introductions; organizing meetups and other events, etc. This can initially be done on free time; if you start to build a thriving mini-community, I’d suggest looking for funding to transition into doing the work full-time, and looking into whether you can expand the population you’re working with.
I find it a bit harder to articulate “on track” conditions for this aptitude than for most of the others in this piece, but a couple of possibilities:
To use a meetup-type model as a concrete example: I’d roughly think that you’re doing well if, within 1-3 years of calendar time (whether full-time or part-time), you’ve had a major hand in organizing a set of people that interacts regularly; has a good number of people who are highly engaged and showing up regularly; and has some number of people who you think are likely to devote a lot of their career to longtermist impact, and are likely to succeed at this. (Specific numbers are hard to give since communities vary significantly, but double-digit regular attendees and a handful of highly promising people are roughly what I have in mind.)
Other versions of community building might look less like “organizing a community of regularly-meeting effective altruists” and more like “creating events that cause new connections to happen.” Here I think you can look at who is “endorsing” your events via their attendance (especially repeat attendance) and/or via recommendations to others, to get a sense of how you’re doing.
A slightly more generalized statement of what it looks like to be “on track”: you’re providing a space and/or service (whether this is a networking service, social space, discussion space, etc.) that a good number of people value and recommend to others; you have a strong sense of your target audience and the value that people are getting out of the space; and there are a number of especially promising people that are making heavy use of what you’re providing.
Software engineering aptitude
Basic profile: I think software engineering can be useful for longtermist goals in multiple ways:
AI labs (mostly in industry, but sometimes in academia) have demand for software engineers, and it may be growing.
In some cases, engineers work directly on AI alignment research; Anthropic, DeepMind, MIRI, and OpenAIF all have roles like this.
In other cases, they may work on building large and capable AI systems (e.g., AlphaStar and GPT3) that can then be analyzed and characterized. If the lab they’re working at is committed to reducing AI risk above other goals, and therefore is cautious about publicizing and deploying these systems while investing heavily in analyzing them and using them to help with alignment research, this can (at least arguably) be good for longtermist goals.
Software engineering can also be useful at any organization doing heavy analysis, which could include organizations working in politics (e.g.) and potentially on biosecurity and pandemic preparedness (I don’t currently know any examples of the latter, but think it’s reasonably likely there will be some down the line).
Software engineering also tends to pay well, especially for people who join successful startups early on. (It is also probably a useful background for someone who wants to start a tech company.) So it could be good for would-be philanthropists.
Catherine Olsson and Tom Brown have both done software engineering at OpenAI, Google Brain, and Anthropic.
How to try developing this aptitude:
Software engineering is a relatively well-established career path. You can start with something like App Academy or Lambda School. (For roles at e.g. DeepMind and OpenAI specifically, one probably needs to be in the top few percent of people in these programs.) Just about any software engineering job is probably a good way to build this aptitude; the more talented one’s peers, the better.
See the “On track” section of Organization building, running and boosting.
Information security aptitudes
(In this case, there isn’t as much difference between an “aptitude” and a “path.” The same applies to the next section, as well.)
Basic profile: working to keep information safe from unauthorized access (or modification). This could include:
Research on theoretical and cutting-edge issues in information security—what attacks could theoretically be carried out, how they could theoretically be defended against, etc.
Working at a company, helping it to (a) define its information security goals and needs; (b) define the kinds of solutions that could be practical; (c) rolling out and supporting solutions so that information remains secure in practice.
This post by Claire Zabel and Luke Muehlhauser states, “Information security (infosec) expertise may be crucial for addressing catastrophic risks related to AI and biosecurity … More generally, security expertise may be useful for those attempting to reduce [global catastrophic risks], because such work sometimes involves engaging with information that could do harm if misused … It’s more likely than not that within 10 years, there will be dozens of GCR-focused roles in information security, and some organizations are already looking for candidates that fit their needs (and would hire them now, if they found them) … If people who try this don’t get a direct work job but gain the relevant skills, they could still end up in a highly lucrative career in which their skillset would be in high demand.”
I broadly agree with these points.
Unfortunately, there aren’t many effective altruists with advanced information security careers as of now, as far as I know.
How to try developing this aptitude:
Working on information security for any company—or working in any field of information security research—could be a good way to build this aptitude. I would guess that the best jobs would be ones at major tech companies for whom security is crucial: Amazon, Apple, Microsoft, Facebook, and (especially) Google.
See the “On track” section of Organization building, running and boosting.
Basic profile: following an academic career track likely means picking a field relatively early, earning a Ph.D., continuing to take academic positions, attempting to compile an impressive publication record, and ultimately likely aiming for a role as a tenured professor (although there are some other jobs that recruit from academia). Academia is a pretty self-contained career track, so this is a case where there isn’t a lot of difference between an “aptitude” and a “path” as defined in the introduction of this post.
Being an academic could be useful for longtermist goals in a few ways:
You might do research that relates substantively to key longtermist questions, which would cause this aptitude to overlap with “conceptual and empirical research on core longtermist topics” aptitudes.
You might have opportunities to raise the profile of important longtermist ideas within your field. You could think of this as being a sort of specialized “communicator” role. (Global Priorities Institute often aims for some combination of this point and the one above.)
You might have opportunities to advise policymakers and the public, as an expert.
You might have opportunities to help introduce your students to important longtermist ideas, including by teaching courses on effective altruism and longtermism (example). (As a side note, I think there could also be a lot of potential impact in being a K-12 teacher who looks for opportunities to introduce students to important ideas in effective altruism and longtermism.)
Additionally, some academic fields open doors for potentially high-impact non-academic roles. AI is perhaps the best example: studying AI and having impressive early-career accomplishments (even prior to earning one’s PhD) can be a good way to end up with a “scientist” role at a private AI lab. Economics also can lead to strong non-academic opportunities, including in policymaking.
Many academic fields could potentially lead to these sorts of opportunities. Some that seem particularly likely to be relevant for longtermists include:
Biology, epidemiology, public health, and other fields relevant to biorisk
Economics and philosophy, the two priority fields at Global Priorities Institute
Hilary Greaves at Global Priorities Institute; Stuart Russell at Center for Human-Compatible AI; Kevin Esvelt.
How to try developing this aptitude:
The academic career path is very well-defined. People entering it tend to have fairly robust opportunities to get advice from people in their field about how to advance, and how to know whether they’re advancing.
In general, I would encourage people to place high weight on succeeding by traditional standards—both when picking a field and when picking topics and projects within it—rather than trying to optimize too heavily for producing work directly relevant to longtermist goals early in their careers.
My answer here is essentially the same as for political and bureaucratic aptitudes.
There are almost certainly aptitudes that have a lot of potential to contribute directly to longtermist goals, that I simply haven’t thought to list here.
Sometimes people are able to do roles that others can’t because they have two (or more) of the sorts of aptitudes listed above. For example, perhaps someone is a reasonably strong software engineer and a reasonably strong project/people manager, which allows them to contribute more as a software engineering manager than they could as either a software engineer or a nontechnical manager. In the effective altruism community, “conceptual and empirical research” often goes hand in hand with “communicator” (as with Nick Bostrom writing Superintelligence).
I think it’s good to be open to building hybrid aptitudes, but also good to keep in mind that specialization is powerful. I think the ideal way to pursue a hybrid aptitude is to start with one aptitude, and then notice an opportunity to develop another aptitude that complements it and improves your career options. I wouldn’t generally recommend pursuing multiple aptitudes at once early in one’s career.
Aptitude-agnostic vision: general longtermism strengthening
I think any of the above aptitudes could lead to opportunities to work directly on longtermist goals—at an AI lab, EA organization, political institution, etc. And I think there are probably many other aptitudes that could as well.
However, some people will find themselves best-suited for an aptitude that doesn’t lead to such opportunities. And some people will develop one of the above aptitudes, but still not end up with such opportunities.
I think such people still have big opportunities to contribute to longtermist goals, well beyond (though including) “earning to give,” by doing things to strengthen longtermism generally. Things that have occurred to me in this category include:
Spreading longtermist ideas within personal networks. I don’t think people should promote longtermism aggressively or in ways that risk annoying their friends. But I think people who are connected and respected will have natural opportunities to get people excited about important ideas who would ordinarily not be open to them. (And success in any career is likely to lead to personal networks full of other successful people.) People who are good at this might also become a sort of expert in how to communicate about longtermism with a certain kind of person.
Showing up in (and/or creating and hosting) longtermist and effective altruist spaces such as local meetups, EA Global, dinners and parties, talks, etc. You can raise the quality of these events both by your presence and by your feedback (noticing what’s suboptimal about them and being vocal about this). I don’t think people should necessarily attend these events only for personal benefit—there’s a lot of good to be done by making them better, such that new people attending them immediately have good experiences and encounter people they respect and can learn from. And hosting/creating events along these lines can often be a good idea.
Being a role model. You can aim to simultaneously be a deeply informed, committed and vocal longtermist, and a person whom non-longtermists think highly of and are glad to know. I think role models are important and impactful, so this could make a real difference within whatever communities you’re in.
Being a vocal “customer” of the effective altruist and longtermist communities. I value it when someone says, “I feel unwelcome in the community because ___” or “I have trouble engaging with longtermists because __” or “I really value events like __ and wish there were more.” People with different perspectives can notice different things along these lines, and help make the longtermist community better at retaining future people like them.
Raising children. I feel a bit odd mentioning this one, and my intent is certainly not to tell anyone they “should” have children. But I believe that raising children takes a ton of work and probably makes the long-run future better in expectation, so it would also feel odd not to mention it. As of now, I’d guess that longtermists with children also significantly increase demand for the longtermist community to become more parent- and child-friendly; this seems like a good (and not minor) thing for longtermism and longtermists.
Donating. I’ve listed this one relatively late because I think “earning to give” has probably been overemphasized compared to the above. But I think there is a lot of potential impact here.
There isn’t currently an “obvious” and arbitrarily scalable place for longtermists to donate, analogous to GiveWell’s top charities. But if one doesn’t have particular donations they’re excited to make, I think it makes sense to simply save/invest—ideally following best investing practices for longtermist values (e.g., taking the optimal amount of risk for money intended to benefit others over long time horizons, and using charitable vehicles to reduce taxes on money that’s intended for charitable purposes—I hope there will be writeups on this sort of thing available in the future). There are debates about whether this is better than giving today, but I think it is at least competitive.
Donor lotteries also seem like a solid option.
Either of these means you don’t have to stress on an annual basis about optimizing your donations; I think that’s a good thing, because I think that time and energy can be better spent on aiming for success as a professional and person, contributing to all of the above.
Being prepared to do direct longtermist work if the need/opportunity arises. The future is hard to predict, and many people who see no track toward direct longtermist work now may end up with a big opportunity in the future.
A late-career job switch can be difficult: it could involve a major reduction in pay and other aspects of status, recognition, appreciation, and comfort. I think anyone who has set themselves up to be truly open to a late-career job switch has (just by that fact) accomplished something impressive and important. I’d guess that your odds of being able to do this are higher if you have significant “reserves” in terms of physical and mental (and financial) health.
I’d guess that anyone who is succeeding at what they do and developing aptitudes that few can match, while being truly prepared to switch jobs if the right opportunity comes up, has—in some sense—quite high expected longtermist impact (over the long run) via direct work alone. I think this expected impact will often be higher than the expected impact of someone who is in a seemingly top-priority longtermist career now, but isn’t necessarily performing excellently, sustainably or flexibly.
I would think anyone who’s broadly succeeding at many of the above things—regardless of what their job is—is having a large expected longtermist impact. I think being successful and satisfied in whatever job one has probably helps on all of these fronts.
How to choose an aptitude
I imagine some people will want a take on which of these aptitudes is “highest impact.”
My main opinion on this is that variance within aptitudes probably mostly swamps variance between them. Anyone who is an outstanding, one-of-a-kind talent at any of the aptitudes I listed is likely having enormous expected impact; anyone who is successful and high-performing is likely having very high expected impact; anyone who is barely hanging onto their job is likely having less impact than the first two categories, even if they’re in a theoretically high-impact role.
I also believe that successfully building an aptitude—to the point where one is “professionally in demand”—generally requires sticking with it and putting a lot of time in for a long time. Because of this, I think people are more likely to succeed when they enjoy their work and thrive in their work environment, and should put a good deal of weight on this when considering what sorts of aptitudes they want to build. (I think this is particularly true early in one’s career.)
With these points in mind, I suggest a couple rules of thumb that I think are worth placing some weight on:
“Minimize N, where N is the number of people who are more in-demand for this aptitude than you are.” A more informal way of putting this is “Do what you’ll succeed at.”
“Take your intuitions and feelings seriously.” A lot of people will instinctively know what sorts of aptitudes they want to try next; I think going with these instincts is usually a good idea and usually shouldn’t be overridden by impact estimates. (This doesn’t mean I think the instincts are usually “correct.” I think most good careers involve a lot of experimentation, learning that some sort of job isn’t what one pictured, and changing course. I think people learn more effectively when they follow their curiosity and excitement; this doesn’t mean that their curiosity and excitement are pointing directly at the optimal ultimate destination.)
I do believe there are some distinctions to be made, in terms of impact being higher for a given level of success at one aptitude vs. another. But any guesses I made on this front would be pretty wild guesses, quite sensitive to my current views on cause prioritization as well as the current state of the world (which could change quickly). And I think there’s potential for enormous expected longtermist impact within any of the listed aptitudes—or just via aptitude-agnostic longtermism strengthening.
Some closing thoughts on advice
Throughout this piece, I’ve shared a number of impressions about how to build an aptitude, how to tell whether you’re on track, and some general thoughts on what rules of thumb might help to be successful and have impact.
I’ve done this because I think it’s helped me try to get across a general framework/attitude for career choice that I think is worth some weight, and can help complement other frameworks that longtermists use.
But I’m generally nervous about giving career advice to anyone, even people I know well, because career choice is such a personal matter and it’s so easy for an advice-giver to be oblivious to important things about someone’s personality, situation, etc. I’m even more nervous about putting advice up on the internet where many people in many situations that I know very little about might read it.
So I want to close this piece by generally discouraging people from “taking advice,” in the sense of making a radically different decision than they would otherwise because of their interpretation of what some particular person would think they should do. Hopefully this piece is useful for inspiration, for prompting discussion, and by raising points that one can consider on the merits and apply their own personal judgment to. Hopefully it won’t be taken as any sort of instruction or preference about a specific choice or set of choices.
I’ll also link to this page which contains a fair amount of “anti-advice advice,” including quotes from me here (“A career is such a personal thing”), here (“When you’re great at your job, no one’s advice is that useful”), and here (“Don’t listen too much to anyone’s advice”).
Some of the content in this section overlaps with that of 80,000 Hours’s content on working at effective altruist organizations, particularly with respect to how one might prepare oneself for a role at such organizations. However, my section excludes research-based and other “idiosyncratic” roles at such organizations; it is about jobs based on “generally useful” skills that could also be used at many non-effective-altruist organizations (some of them giving an opportunity to have longtermist impact despite not being explicitly effective-altruist). In other words, this section takes a frame of “building aptitudes that can be useful to help many organizations, including non-effective-altruist ones doing important work” rather than “going to a non-effective-altruist organization in order to build skills for an effective-altruist organization.”
I’d expect most investigations of this form to “balloon,” starting with a seemingly straightforward question (“What are the odds of nuclear winter this century?”) that turns out to rely on many difficult sub-questions (“What are the odds there will be a nuclear war at all? How much particulate matter does a typical nuke kick into the air? Are there bigger nukes that might be deployed, and how much bigger?”) It can be very difficult to stay focused on a broad question, handling sub-questions pragmatically and giving a reasonable amount of depth to each. But allowing oneself to switch to answering a narrower and narrower subquestion could make the work more tractable.
This takes into account the fact that this kind of work can be very hard to put a lot of hours into. I’d expect even people who are a great fit for it to frequently struggle with maintaining focus and to frequently put in less time than they intended; nonetheless, I’d expect such people to achieve roughly the kind of progress I outline on the calendar time frames discussed.
This section is similar to 80,000 Hours’s discussion of “nonprofit entrepreneur,”, with the main difference being my emphasis that entrepreneurship experience with a non-effective-altruist organization (including a for-profit) can be useful.
For example, “online organizing”—asking people to take relatively small actions on compelling, immediate topics, resulting in their becoming more engaged and reachable on broader topics.
Also see my comments here (under “Not focusing on becoming really good at something”), which were made anonymously but which I’m now attributing.