TL;DR
- Don’t spend too long thinking about the pros and cons of applying to an opportunity (e.g., a job, grant, degree program, or internship). Assuming the initial application wouldn’t take you long, if it seems worth thinking hard about, you should probably just apply instead.
Key Points
- It’s very easy to end up spending a similar or greater amount of time figuring out whether to apply than it’d take just to actually apply, given that it often takes just 30-120 mins to fill out the initial application form. You should probably just apply instead!
- It’s very hard to predict what you’ll get offers for, so it’s usually better to just apply and find out what actually happens.
- When you apply, or when you do each later stage of the process, you’re not deciding whether you’d want to accept the opportunity if offered - you’re just deciding whether to spend the time on whatever the first/next stage of the application process is.
- And given that most applicants to most opportunities get rejected anyway and that it’s often hard to work out whether you’d want to accept the opportunity, you should probably not bother thinking much about that till you actually have an offer (or perhaps a many-hour work-trial task)..
- You’ll also often get some benefits even from the application process alone (e.g., learning about your personal fit for the role, improving your career capital).
- I suggest a heuristic of applying for ~20 things a year if you are looking for work and ~5 things a year even if you think you probably don’t want to / can’t take a new opportunity.
Epistemic status and authorship: This was drafted quickly by Michael and then edited and expanded by Akash and Michael. It mostly reflects Michael’s thinking, and parts of that thinking might be off.
Apply to things!
If someone wants to test their fit for a given line of work or build their career capital, my key recommendation is usually to apply for lots of "actual" jobs, grants, degree programs, internships, or other such things. (See the final section for collections of opportunities to apply to.)
I think this is more useful than focusing on reading up on an area, doing independent projects, taking little courses, etc. It can make sense to also do the latter things, but that’s probably less good than just making loads of applications. So you should probably only do the other things on top of making many applications and if you happen to have extra time, rather than instead of making many applications.
Why am I such a fan of applying to things?
Applying to things often has good expected value, because:
- You might actually get one of these opportunities, which could allow for a lot of testing of fit, building of career capital, and direct impact.
- You’ll get at least some evidence about your fit for that opportunity and for similar opportunities/paths
- The evidence comes from seeing what the work tests or interviews involve, seeing how you feel you did in the work tests / interviews and how much you enjoyed them, and seeing how far through the selection processes you get.
- To be clear, getting a single rejection provides only a small amount of data, since rejection is essentially the default outcome. But lots of rejections for a given type of role eventually provides a decent bit of evidence, as does getting far in one work test or getting an offer.
- You may build career capital from the selection process, even if you don't get or accept an offer. This career capital comes from:
- building your skills and knowledge during the work tests
- the hirers recommending you for other roles or connecting you with other people, even if they don’t make you an offer for their role at this time or even if you don’t accept it
- the hirers reaching out to you for other opportunities later (perhaps ones that are more entry-level, or after you’ve had time to build your skills more)
And the EV doesn’t have to be very high to make applying worthwhile, given that:
- Early stages of selection processes tend to not take up much time
- For most jobs, by the time you get to more time-consuming stages, the EV is probably higher, since now your odds of actually getting the job are better, the work test may be more useful for testing fit, etc.
How many things should I apply to, and how much time should I spend thinking?
My rough suggestion, essentially just based on what I did/do rather than lots of reasoning, is to:
- Apply for something like 20 opportunities per year when actively seeking work
- Apply for something like 5 opportunities per year even when planning to not leave your current role
- Since those 5 things might turn out, on further reflection, to be worth changing your plans for, and/or you might learn a lot from applying or be able to defer an offer
- Have heuristics that help you quickly determine whether an opportunity probably meets the “top 20” or “top 5” criteria. Some examples include taking into account how likely you are to want the role, how good the role would be if so, how likely you are to get it, how time consuming the application process is in expectation, and maybe other factors.
- Then simply apply without any further thought whenever the answer seems to probably be “yes”, rather than trying to carefully consider each of those factors in any detail.
- You don’t need to end up confident about whether an opportunity is worth applying for - you just need to make it fairly likely that you apply to most of the things you should apply for, that you won’t spend even longer on applications than you should (e.g., applying for 50 things is probably too much), and that you won’t spend too long deciding whether to make each application.
- Think harder (only if and when when you get either to long work tests or to an actual offer.
"20" and "5" feel pretty made up, but I feel pretty confident that the overall procedure is a good one.
What, specifically, am I suggesting not spending much time on?
I’m suggesting not usually spending much time on the following things unless either (a) you’ve gotten up to a many-hour work test or a job offer or (b) you’re doing these things mostly like “case studies” to learn about career paths rather than this individual role (see the following section):
- Reading a lot about a given opportunity
- Going to Q&As about the role
- Talking to people about the role (e.g., people who work at the relevant org)
- Thinking about your odds of getting the role, thinking about whether they’d want to accept the role (e.g., would they have to move and are they willing to? Would they enjoy the job?)
Spending lots of time on that before applying to an opportunity would essentially be spending lots of time:
- armchair reasoning about something you can more cheaply just gather actual data on
- trying to inform a decision you probably won't end up getting/having to make anyway (whether to accept a role you currently have a low-ish chance of being offered)
What is worth spending time thinking about?
To be clear, there are some activities that look like “spending lots of time deciding whether to make each application” but that I do think are often worthwhile. These include:
- Spending time working out what types of opportunities to focus one’s search on
- E.g., should you in general be looking for roles at EA orgs, corporations, gov, academia, or elsewhere? In research or ops? In AI governance or animal welfare?
- Thinking and learning about specific roles can be useful for this, as something like concrete case studies.
- Spending time working out what hirers are looking for in cases where hirers do have things in mind that they aren’t telling people. Or spending time ingratiating yourself in cases where - unfortunately! - that will work on hirers.
- Thankfully, both of those things will rarely be relevant for EA orgs, since they generally try to (a) make it as clear as possible what they’re after and (b) just see whether people can actually do well in work tests and things like that, rather than having their selection process confounded by “Who put the most effort into / had the most success with guessing our secret selection criteria and quirks?”
- But it’s relevant for some other orgs, unfortunately.
- Spending time thinking about which skills and aptitudes you want to develop (see e.g. here and here).
Miscellaneous additional points
- One way to frame much of this advice is “Be more empirical and less theoretical - go and gather data rather than focusing on armchair reasoning”.
- E.g., 80,000 Hours write “Many people try to figure out their career from the armchair, but it’s often more useful to go and test things in the real world.”
- My guess is that hirers would generally be actively happy to get more applicants who are at least (say) 2% likely to be great fits and at least 10% likely to take the role if offered, rather than wanting such people to screen themselves out.
- This is because such applications can be high EV from the hirer’s perspective, given how valuable it is when a person is a great fit and joins the org, and because evaluating each individual applicant at each individual stage tends to be quite quick for hirers (e.g., ~10 mins per 2 hour work test evaluated).
- Don’t screen yourself out just because the org says they need you to start in a given month but you want a new job by 2 months earlier or could only start 2 months later, or something like that!
- If you end up getting an offer, there’s a decent chance you’ll realize you actually can rearrange things to start earlier or later. There’s also a decent chance the org will turn out to be willing to rearrange things or wait for you so you can start earlier or later.
- Again, this comes from how high value getting a good job and getting a good hire are for an applicant and an org, respectively.
- Note that the “decent chance” need only be high enough to justify the small time cost of you applying and the org evaluating you.
- You might even be able to defer the offer for much longer.
- You might also end up politely declining an offer while also expressing appreciation and enthusiasm and asking if the org can reach out to you again after whatever time you might be available again.
- Don’t screen yourself out just because you don’t have every single qualification. If you would be a good fit for some parts of the role, but not others, you might still be the best candidate for that precise role, or the org might end up wanting to shape the role somewhat to suit you. Let the application process help you figure out if you’d be a good fit.
- Consider a wide action space. The space of possible career options is large—often much larger than people initially think. Don’t limit yourself to the first two or three job ideas that come to mind. The most impactful thing you can do is often not intuitive.
- When in doubt, apply* provides advice that I mostly agree with and that overlaps with this post’s advice
- People who find this post interesting might also be interested in the following notes and recommendations I collected previously:
You have hereby convinced me, O Wise One - but to what shall I apply?
Hard to say, as I don’t know you at all!
But some things you might want to check out are:
My thanks to Janique Behman and Sean Engelhart for feedback on earlier drafts of this post.
Huge +1 to this post! A few reflections:
From my perspective as an applicant, fwiw, I would urge employers to reduce the scope of questions in the initial application materials, more so than the time commitment. EA orgs have a tendency to ask insanely big questions of their early-stage job applicants, like "How would you reason about the moral value of humans vs. animals?" or "What are the three most important ways our research could be improved?" Obviously these are important questions, but to my mind they have the perverse effect that the more an applicant has previously thought about EA ideas, the more daunting it seems to answer a question like that in 45 minutes. Case in point, I'm probably not going to get around to applying for some positions at this post's main author's organization, because I'm not sure how best to spend $10M to improve the long-term future and I have other stuff to do this week.
Open Phil scores great on this metric by the way - in my recent experience, the initial screening was mostly an elaborate word problem and a prompt to explain your reasoning. I'd happily do as many of those as anyone wants me to.