High-impact roles hire differently. Here’s what you need to know.

How to thrive under pressure (or at least fail gracefully)

We post over three thousand jobs each year.¹ Many come from small-to-medium nonprofits and companies in AI safety, biosecurity, and other neglected areas.

My last post covered ways to make those organizations notice you — from getting referrals to building a more compelling online profile.

But getting noticed is only step one. Next, you’ll need to outcompete everyone else who got noticed.

Frankly, I only want you to get hired if you’re actually the best candidate. But equally, if you are the best candidate, I want you to get hired!

The orgs we work with tend to be quirky: they have unusual standards, from the traits they look for to the way they test you. I’ve seen those quirks throw off promising applicants, spoiling what could have been a great match.

I can’t make the orgs less quirky.² But I can help them avoid false negatives by telling you what to look out for, so you can put your best foot forward.

If you want to get a high-impact role, and you make it past the resume screen, here’s what you need to know next:

A good application opens the gates to the underworld, but your heart must be weighed before you reach the Field of Reeds. (Anubis represents your interviewer.)

Prepare for an in-depth process

The typical “high-impact” hiring process³ is longer than most in the private sector. It has the usual interviews, but also a series of work tests (or even a multi-day work trial). These can take many hours to complete, though you’ll usually be compensated.

Also, high-impact orgs tend to run lean. They often won’t have dedicated hiring teams, and this may be the first time they’ve ever hired for your position. Don’t be surprised if they miss their projected timeline, or if each step takes a while.

Learn about the org ahead of time

When I was at Coefficient Giving in 2025, my team hired an editor. Several applicants professed excitement about our work in areas we hadn’t funded since 2017 (or ever). That was a red flag for us, as it would have been for most high-impact orgs.

A giant corporation can afford low-context employees; you don’t need to know Google’s full business model to write their code. The jobs we post are different. Before your first interview or work test, take the time to understand the org’s work and worldview, and to form your own views about it.⁴

Here’s why:

1. Most of these orgs are small. They need people who can wear multiple hats. This is easier when each employee understands the mission. I’ve seen lawyers run conferences, accountants meet with applicants at job fairs, and writers catch math mistakes in spreadsheets — all things that require context.

2. Mission alignment matters. Nonprofits lag the private sector on pay, status, and broad career capital. They find and keep great employees by hiring people who really care about their mission. If you have the right skills but don’t seem to get the big picture, they’ll worry that you might not stick around. You’re also more likely to get quick promotions and decision-making power at a small org, which could go badly if you aren’t on board with their strategy.

You don’t have to agree with everything you hear; high-impact orgs value honest feedback and encourage internal debate. But you do have to understand their work, and value their ultimate mission (even if you’d tweak the tactical approach).

Be honest and admit uncertainty

Here are questions I’ve asked, or been asked, in job interviews:

• “What do you think we’re doing wrong?”

• “In your last job, did you actually have a positive counterfactual impact, relative to what the average employee would have done? What makes you say that?”

• “What’s something you believe that most of your friends don’t? Whatever you say, I will argue with you.”

Many high-impact orgs are borderline maniacal about honesty and calibration. When you talk to them, acknowledge your mistakes and uncertainties — and don’t be sycophantic.⁵ Don’t hesitate to disagree, say “I don’t know,” or admit to feeling confused. You still want to present your best self, but you should be less ‘salesy’ than you would for the average interview.

You shouldn’t be dancing on strings, and you should always tell the truth.

What to know about work tests

Nearly all of these orgs use work tests. The tests themselves tend to be simple, but plenty of candidates still stumble over them. A few ways to avoid that:

Follow the instructions. I’ve seen a lot of smart people completely miss something I asked for: they breach the word count, forget to share the document, de-anonymize themselves, or fail to flag uncertainties.

Don’t rush to get started. Slow down, read carefully, and outline what you need to do.

Share your reasoning. Work tests are meant to capture how you think. Orgs don’t only want to know how well you write or code; they want to understand what you’d be like as a colleague. How do you set priorities or choose between options? Can you notice when you’re missing information, or when your favorite idea has downsides?

Pay close attention to any request to share your reasoning, and consider adding notes on your process even for mundane tasks. The essay “Reasoning Transparency” shares how Coefficient Giving, the largest funder in the space, thinks about this; it’s been hugely influential for many orgs, and is worth reading before your next work test.⁶

Use your time intentionally. Work tests are also meant to capture how you… work. They use time limits, and often give you more than you can handle, to imitate the real world (where you never have time to do everything you want).

Focus on the most important output, and don’t be afraid to use estimates if precision would cost a lot of time. Most tests allow comments or footnotes to explain your decisions: “I rounded these figures to save time. I made the following assumptions…”

Give yourself a few minutes at the end to review your work and add more comments. For a multi-hour test, ten minutes of review outweighs ten marginal minutes of work.

If you don’t finish everything, add a note to show your thinking: “Here’s why I prioritized X. I think I spent too long on Y. My next step would be Z.”

Don’t (only) use AI. In the LLM era, I’ve seen tests with barely-edited AI responses. That’s a bad idea. The org can use AI themselves; they want to see what value you would add, and how your thinking would contribute to the work. Even if a test doesn’t forbid AI, your contributions should be clear.

Trust your feelings. Work tests are meant to be realistic. If you enjoy a test, that’s a sign to seek out similar roles. If you hate a test, and it’s not just anxiety, that’s a sign to do something else. If you don’t like the first four hours, the next four thousand won’t be much fun either.

Before I joined 80,000 Hours, I took a five-hour editing test from another org. After the first hour, I had a headache and a creeping sense of dread. I was starting to see AI governance the way you’d see raw fish after a bad case of food poisoning. So I closed the tab, sent an apology to the hiring manager, and thought about what I’d learned.⁷

Don’t assume rejection means failure

Unlike your work history, interviews and tests are within your control; they represent your current abilities. That makes a late-stage rejection especially frustrating. It can feel like a signal that you aren’t cut out for this work.

But rejection doesn’t mean you didn’t meet the bar — it usually means someone else was an even better fit. For every person I’ve hired, I rejected multiple qualified candidates. The further you get, the more likely it is that you left a good impression.

Good impressions pay off! Most high-impact orgs share candidates with each other. After Coefficient Giving rejected me in 2018, they referred me to the Centre for Effective Altruism (CEA), which wound up hiring me. My current manager was hired after the Global Priorities Institute (RIP) rejected her and referred her to 80,000 Hours. When I hired a part-time editor at CEA, I kept the top few candidates on a list and sent them any writing role that crossed my desk, with an offer to refer them if they applied. Later on, I hired one of them for a different role.

This could be you — and it’s more likely if you are easy to advocate for, which is basically the same as being easy to hire. I’m more likely to recommend you if I recall what made you special, can share examples of your work, or felt like you were especially sharp and engaged when we spoke.⁸

Don’t think of the application grind as merely a series of ‘job or no job’ decisions. It’s also the process of building a reputation.

What to read next

• Check out our revamped career guide — all our best ideas in one place! Since you made it this far, you might like part 14, “How to actually get a job”.

• If you’re having trouble landing interviews or work tests in the first place, try my last post: “How to become a better applicant in one week.”

• I’ve been cleared to share two actual 80,000 Hours work tests, in case you want to see what they look like! (Neither is in active use.)

  • Operations: Designing a time off process

  • Writer-researcher: Improving an article

• Alison Green’s guide to interviewing covers the basics very well, including things other guides often miss (like building context or practicing out loud).

1. That number might be a lot higher soon. If major AI companies go public, their shareholders may direct billions of dollars toward our priority problems, unlocking a ton of new jobs.

2. Nor would I necessarily want to: the quirks are often there for good reasons! Many features of “high-impact hiring” are meant to fix problems with the standard application process.

3. I don’t love the phrase “high-impact” (a bit presumptuous, no?). But it’s the cleanest way I’ve found to describe the jobs and orgs we recommend. Think of “high-impact orgs” as “orgs 80K likes in spaces including AI safety, biosecurity, effective altruism, global health, animal welfare, and others that relate to our priority problems”.

4. Your understanding doesn’t have to be perfect. Asking good questions shows that you care about the material and are comfortable admitting uncertainty; both are highly valuable.

5. I know the limits of my own intelligence. If someone agrees with everything I say, I get suspicious.

6. You can skip the secondary recommendations.

7. I learned that I hate editing academic writing.

8. This is another reason to be honest about your skills, knowledge, and interests. Even if you slightly reduce your odds of getting hired right away, I’m more likely to recommend you elsewhere if I feel like I can trust you. (Plus, if you land in a role where you aren’t a good fit, it’s worse for your impact and your long-term career.)