How to get good at something useful: Part 9 Which jobs best advance your career?

When people think about how to best advance their career, they often think they should get a job at a ‘brand name’ organisation like Google, Deloitte, or McKinsey. As job fair recruiters are keen to point out, these jobs give you a general professional training, which can enable you to enter many other jobs afterwards.

Convinced by this pitch, more than half of Harvard students go into consulting, finance, law, or tech after graduating, and it’s similar at other elite universities.¹ In a book of the same name, the Dutch journalist Simon van Teutem recently called the first three the “Bermuda triangle of talent” because so many of our best and brightest vanish into them, never to be seen again. While there, they face 70-hour weeks, unfulfilling work, and higher rates of burnout.² And as we saw in part eight about skills and the future, junior roles making PowerPoints seem among the most vulnerable to automation.

I disagree with this fully negative position. We’ve worked with many people who felt they learned a lot in these options and were able to switch into something more meaningful later. And while we still have jobs at all, we’ll still have these industries. But there are often faster routes into the most impactful careers, and you’d be better served by considering them too.

In part seven, we explained how the roles that are best for career capital are those that most efficiently advance you towards impactful jobs and provide the rapid feedback that lets you learn quickly. Next, we covered which skills are most valuable to learn. Here, we’ll suggest some specific jobs and other roles that give you the conditions to rapidly learn these valuable skills, and that have worked out among the people we’ve advised over the last 15 years.

These options are most useful straight out of college, but can also be used to switch paths mid-career. Note down any that could be a good fit for you, and we’ll return to how to pick between them later.

Reading time: 20 minutes

1. Work at a smaller, growing, high-performance organisation

In college, you’re taught to answer well-defined problems with clear answers, which are possible to master relatively quickly. In the world of work, by contrast, much of the challenge is figuring out what to focus on in the first place. Projects don’t (automatically) have well-defined scopes or success criteria. Great performance might not be possible, or could take many years. You also probably don’t know how to do the basics, like run a weekly check-in meeting, read financial statements, give good presentations, or speak to a boss.³

So one of the most useful things you can do after college is to work with any high-performing, high-integrity team where you can be mentored in the skill of generally getting stuff done at work.

Many large, well-known organisations have filled this function in the past, but today cookie-cutter junior white-collar roles seem among the easiest to replace with AI, making this track less attractive than before, or at least higher risk. So what’s an alternative?

Working at an organisation that’s small but growing has always offered a number of advantages. You don’t need to wait for senior people to retire to get promoted, dramatically increasing your opportunities to advance.⁴ Your work probably involves doing whatever needs to be done, making it easier to develop a generalist ‘getting things done’ skill set, including skills like productivity, learning rapidly, entrepreneurialism, and familiarity with many aspects of organisation-building — all skills that seem likely to increase in value going forward.⁵

We frequently encounter people who want an entrepreneurial career, so they go to one of the consulting giants like BCG or McKinsey. But working at a startup is much better prep for one day founding a startup, and is more respected in the tech sector, too. Startup experience is also very useful for many social impact projects, which are often small and scrappy.

In contrast, working for a company like McKinsey best prepares you for life in large corporations where, as a cog in a machine, advancement is often driven more by political skill than the outcome of your work. At smaller organisations, it’s easier to see the effects of your actions, and this real-world feedback also supports rapid learning.

That’s not to say you should always prioritise small organisations. The main drawback to them is that many are dysfunctional, if not a total shambles. You’ll need to do your homework when selecting one. If you’re good at identifying organisations with potential, this can be an advantage. Becoming the hundredth employee at Dropbox resulted in a payday of $10 million, more than the average founder earns,⁶ and was probably easier too. This example is financial, but a similar principle applies to positive impact.

Larger organisations are less likely to be totally dysfunctional, run out of funding, or get you zero CV points. Employers know that somewhere like McKinsey has a tough selection process that brings them smart, well-rounded people. But if your CV says you worked at Random.io, that’s harder to evaluate.

Typically you’ll also get less formal training at new organisations, because resources will be more stretched. If you benefit from learning-by-doing, are more resilient to stress, and appreciate autonomy, this can work to your advantage. But we’ve also advised people who found that the stability, processes, and slower pace of larger organisations let them learn faster. So it can come down to personal preference. The important thing is that you consider roles at smaller organisations at all, rather than defaulting to the best-known options.

If you do want to work for a smaller organisation, which one should you choose? Small nonprofits are more likely to be dysfunctional than small for-profits, because nonprofits can continue to operate as long as they can tell a good story to donors, while the need to earn revenue provides real-world feedback on performance. For-profits can also grow faster and give you the chance of making a lot of money, which, like it or not, tends to open up a lot of opportunities.

Within for-profits, a popular option is working at a tech startup. I’d especially highlight working at an AI applications startup, since they have the benefits already discussed, and also let you learn the skill of using AI to solve real-world problems, which is likely to be one of the most useful skills of the coming decade. In particular, look for organisations deploying existing AI technology to solve important problems rather than those pushing forward the technical frontier.

Whether to work at the frontier labs such as Google DeepMind, Anthropic, OpenAI, and DeepSeek is a more difficult question. It gives you a cutting-edge understanding of AI and the opportunity to work on projects that can only be done with access to frontier technology. They’re also growing rapidly and have high pay. On the other hand, working at one of these companies might contribute to accelerating AI risks, and you might not feel aligned with your team.

While for-profits offer some advantages, if you can find a good nonprofit, you can gain great career capital there too, and many people we’ve worked with have done so. You could also consider working at think tanks, research labs, or even newly-established government agencies.

Regardless of the size or reputation of the organisation, be on the lookout for where the most talented people are heading. PayPal alumni have gone on to spawn hundreds of companies. The Church Lab set up by Harvard geneticist George Church has spun out over 40 biotech companies.⁷ If you can find the equivalent in your industry, working there is going to be one of the best things you can do for your career, especially if those people are also socially minded. That remains true even if your parents — or friends — have never heard of it. My parents haven’t heard of Anthropic, Jane Street, or the UK’s AI Security Institute, but these are rapidly growing organisations that are well respected in their fields. What matters is impressing future employers, not your Dad.

The best organisations might not even advertise their roles, so finding a job in this category requires a bit more work than simply filling in an application form. This is where making connections and the job-hunting advice we’ll cover in parts 14 and 15 become more important.

2. Take an established route into policy careers

At age 21, Elena was working at the White House on pandemic prevention. Before that, she helped to direct tens of millions of dollars towards funding for a strategic stockpile of personal protective equipment (PPE) to protect key government workers during a future pandemic. How did she end up in such an impactful position so fast?

Although Elena’s case is extreme, it’s not uncommon for young people to end up in very influential positions in government. While the field contains a tremendous diversity of roles, people tend to enter them through a couple of standard routes. These routes can sometimes allow you to advance remarkably quickly while also building connections and professional skills that will be useful in a career outside of policy. The main categories depend on your country, but usually they are:

• Fellowships and leadership schemes. One- or two-year programmes that aim to accelerate people into leadership positions in government. Some of these are aimed at experienced professionals, others at recent graduates.

• Working for a politician. People often advance from working as a researcher, staffer, or campaign assistant into key government positions, or even elected office. It’s demanding and prestigious, and gives you connections. (Though bear in mind it will associate you with a single political party.)

• Think-tank internships or research roles. These are jobs that will help you learn about key policy issues. They’re also prestigious, and open up positions in government, party politics, and the social sector more broadly.

• Roles at key government departments or agencies. This means applying to entry-level jobs as a government employee, and then trying to work your way up into more influential positions, or if you’re already experienced, finding a senior role using your current skills. Ideally you’d work on something relevant to a problem you want to help solve, but that might not be possible in your first job.

At this point you may be thinking, “This all sounds very well, but how do I actually get any of these jobs?” There’s a standard playbook, which Elena’s story illustrates.

She decided she wanted to work on AI or biosecurity to prevent future pandemics, but was pursuing a technical degree and had little prior experience. So she started talking to everyone in the field she could find, attending conferences and getting in touch with friends of friends: anyone she thought might be able to give her advice. She also started to read all about the areas of policy she was interested in.

When her summer holidays came along, she decided to drive to Washington, DC, sleep on a friend’s couch, and get coffee with every connection she’d made. Not many young people are pumped to work on pandemic preparedness, so she impressed people with her enthusiasm. One of her meetings led to an internship in the executive branch. Once she had that position, she worked super hard, opening up more opportunities and eventually ending up in the White House.

You can attempt a similar strategy. Make a couple of connections to people in policy through conferences or friends of friends. Then visit the city where they work — most likely the seat of political power, whether Washington, DC, Beijing, Brussels, or London — and try to use your connections to meet as many people working in policy as possible. Have some specific areas of policy you’re excited about and tell people you’re looking for a way in. Our job board also has lists of fellowships, think-tank internships, and relevant government jobs. It doesn’t matter too much which kind of job you get. Once you have a toehold, you can work hard and advance from there, switching between different types of policy jobs over time.

If you don’t land a role right away, then try to find a way to do the work on a small scale. People will take you more seriously if you’ve written something on the topic, helped to run a conference, or otherwise made a visible contribution. For instance, Jeffrey Ding realized there was huge interest in what was happening with AI in China, but very little knowledge of it. He started curating and translating Chinese-language articles and sending them out in a newsletter. It quickly grew, reaching 17,000 subscribers. Today he’s an assistant professor of political science specialising in the role of technology in great power conflict. He’s testified to the US Senate and his research has been featured in The Washington Post and the Financial Times.

Otherwise, you may need to first do something to build relevant career capital for 1–2 years. This could be anything that lets you learn about the policy area you want to work on (e.g. a climate nonprofit to work on climate policy), but the most common option is graduate school, as we’ll cover next.

3. Go to graduate school in carefully considered subjects

Bilal had done well in maths as an undergraduate and was given an offer to do a PhD in cosmology. But like many students, he still didn’t have much idea what he wanted to do with his life. In his final year of college, he avoided thinking about the question. Not sure what else to do, he accepted the offer.

After accepting, he immediately had doubts. Even in a best-case scenario, he thought his PhD would play a very small part in improving our best theories of the expansion of the early universe. But having just lived through COVID-19 brought home the urgency of some of the other problems facing the world. He also realised he didn’t really want to be an academic cosmologist, and the PhD didn’t seem likely to open up any roles he couldn’t enter more directly. One week in, he decided to leave.

He felt anxious about the decision. Everyone else seemed to know what they were doing, but for the first time he didn’t have a clear goal. After reading our guide, he started to explore different problems he could work on. He took a job that would let him improve his programming skills and began studying machine learning on the side. After a series of fellowships, he landed a job at DeepMind working with Neel Nanda on AI interpretability, where he can now hopefully contribute to a highly urgent problem, and in less time than a PhD would have taken.

Many lack Bilal’s courage. Most PhD students in the US take longer than seven years to finish, and many never finish at all.⁸ Worse, PhD programmes can be demoralising. A survey in Nature found 41% of graduate students had “moderate to severe” anxiety, and 39% had moderate to severe depression, compared to 6% of the general public. According to standard DALY weightings, in terms of years of health lost, this makes graduate school a public health hazard comparable to all STDs combined.⁹

This isn’t to say doing a PhD is always the wrong choice. Several years later, Bilal applied to PhDs again, but this time in machine learning rather than cosmology. A PhD is still necessary in many careers — not just academia.¹⁰ Used well, it can be like a multi-year research retreat, surrounded by some of the world’s smartest people, where you can explore an important topic in-depth. If research turns out to not be for you, the flexibility you have over your time means you can explore other options — many startups have been started by PhD students, including 80,000 Hours itself.

Master’s programmes can be even more attractive. A one- or two-year master’s in a quantitative subject, like economics, can open up technical roles that would be hard to enter otherwise.

If you do decide to do graduate study, which subject should you choose so you don’t end up regretting it? Well, first, don’t rush into it. In part 11, we’ll make the case for exploring in the year or two before starting graduate school, if possible. But when it comes time to choose, here are three key criteria for comparing programmes:

1. How useful is it to your longer-term plans? Arts students often drift into master’s degrees in history, the classics, or literature, even though they don’t really plan to work in the field. Would-be entrepreneurs are also often tempted to do MBAs. But MBAs are mainly useful for advancing in large corporations. Think about what you might actually do long term and whether graduate school is an efficient way to pursue it.

Also think about how easy it would be to learn the skills outside of grad school. It’s easier to read about history in your spare time than geometric algebra, so it’s especially helpful to focus on quantitative skills at school.

Keep in mind that some graduate programmes give you much better odds of landing a research position than others. For instance, 70% of statistics PhDs can get research positions, whereas only about 40% of psychology PhDs do. It’s even worse in philosophy, where even those admitted to top-10 programmes have under a 25% chance of getting a permanent academic position.

2. How good is your fit for the subject? If you’re good at it, you’ll enjoy it more, be more likely to complete it quickly, and be more likely to work in the area in the future. It’s usually better to do a subject you love that’s a bit less generally useful than to struggle through an option that seems best in abstract.

3. How good are the back-up options? If the research path doesn’t work out, or you decide to change track, will this programme give you other options? Watch out especially for subjects that don’t really help in anything outside of academia, such as a literature PhD. In contrast, studying more foundational, quantitative subjects has historically kept more options open. If you’ve studied maths, for instance, you can transfer into economics, physics, biology, machine learning, and so on, but the reverse is not true. Economists can transfer into the rest of the social sciences and public policy, when the reverse is harder. Having a quantitative degree opens up jobs that require quantitative skills, including many in technology, policy, and the private sector. Even if you end up pursuing a career that doesn’t call for quantitative skills, a quantitative degree is often seen as a signal of smarts, one reason they’re associated with higher income.

The subjects we most recommended studying are effectively the same as the fields of expertise we highlighted in part eight. You may not be immediately interested in these subjects, but it’s worth giving them a chance. We’ve advised plenty of people who couldn’t imagine studying anything except what they were currently studying, only to try out some other subjects and find they enjoyed them more than they expected.

Many subjects besides those we most highlight can make sense for the right person. History PhDs aren’t among our top recommendations due to their poor prospects in academia and lack of backup options. Between 2016 and 2021, UK universities awarded close to 4,000 History PhDs, yet the entire history staff across all UK universities totalled fewer than 3,000 people in 2021. But there are many topics we’d love to see more historians tackle, so if your fit is good enough, it still can make sense.

For instance, we’ve argued existential risk is an idea that should be taken much more seriously by society, but there’s been little study of its intellectual history. At the same time, many dismiss it by noting the long history of different cultures believing Armageddon is nigh. Historian Thomas Moynihan decided to investigate whether that was really true. He found that almost all cultures have actually assumed the universe would last forever. Belief in Armageddon is common, but it’s seen as part of a cycle, or a transition to an infinite heaven or hell. The idea that the world could end and have no successor, and that this could be caused by human hands, is a modern invention. It has roots in the Enlightenment, and only became common from the twentieth century after it became technologically possible.

Whatever subject you choose, consider the reputation of the department and the mentorship you’ll receive from your supervisor. One summer I worked in the Cavendish Laboratory museum in Cambridge. It was founded by J. J. Thompson, who I learned is perhaps the most successful graduate supervisor of all time. Seven of his students went on to win Nobel Prizes, a batting average of about one in 20.

Also consider whether the day-to-day environment will suit you, and whether it’ll help you make connections in key geographic hubs, such as studying in Georgetown for DC policy connections, or at Berkeley for connections in AI. Try to speak to current students about all of these factors. Whether you’ll get funding to avoid debt is also a big consideration. If you’ve thought of some graduate programmes that might work for you, add them to your list of ideas for next steps.

4. Get a verifiable valuable skill

Suppose you can’t find any roles in the categories above. Sometimes it’s possible to spend 3–12 months doing a crash course that will teach you a concrete skill, and then use that to get a job — and potentially advance a lot faster, too.

The first thing that’s crucial is that the skill is verifiable. You need to be able to prove your competency to employers and it needs to be something employers value.

The second thing is that you can find a good environment to learn it rapidly. This could be a formal course, like a data science bootcamp, language school, or an online course, like Bluedot’s course in AI governance. But you could also do something like produce a bunch of design work and put it online, or study a technical skill intensively with a friend, as Daniel, who we met in part seven, did by replicating machine learning papers. Finally, you could also do a side project, like volunteering to help run a conference or entering a forecasting competition to improve your judgement. Pause for a moment to see if you can think of anything like this that might help you enter the options you’d like to end up in.

5. Do anything where you might excel (even if it’s a bit random)

We met a magician, Romil Depala, who had a chance of landing a national TV show in India, and was deciding between that and . . . consulting. The magician path seemed more exciting to us, because the skills and connections he could make seemed more unusual and valuable for spreading ideas about the world’s most pressing problems than another consultant.

He didn’t sell a show to a channel in the end, but he did release a documentary featuring several stars in Bollywood. He also performed for high-profile clients, including the India cricket team and Shah Rukh Khan. Later, he was able to work with several magicians to create some of the most-viewed content channels on Meta and Snap — accumulating billions of impressions and getting featured in the The Economist.

Working in this world showed him the major reach of viral content — billions of people watching channels mainly featuring pimple popping and ASMR — but he’s now able to put these skills to good use by helping impactful projects capture similar online attention. This year, he’s already helped several AI risk nonprofits get tens of thousands of extra views for their videos and helped with several political campaigns, and he’s just getting started.

While it’s easy to focus excessively on the ‘hard’ aspects of career capital, like having a well-known employer and formal credentials, the best career capital doesn’t come from either of these. It comes from doing impressive work. Doing impressive work stretches you, builds your reputation, and allows you to make connections with other high achievers. In fact, extreme enough success in almost any field can lead to connections, fame, influence, and wealth that can be directed towards impactful aims in the future. If there’s a field — or even a hobby — where you think you have a chance of high achievement, put it in your shortlist of options for career capital, even if it’s not immediately related to doing good.¹¹

Isabelle Boemeke was a fashion model, but after wildfires ravaged her home country of Brazil in 2019, she decided she wanted to do something about climate change. She spoke to experts about which interventions were most effective. To her surprise, she discovered many of them thought nuclear power was crucial to providing enough CO2-free energy, but didn’t want to say so publicly for fear of backlash.

She decided to use her glamour and social media following to make nuclear energy cool again. She created a persona, “Isdope,” and made viral videos, initially horrifying her agent. But over time she built a new following, giving a TED talk and eventually publishing a book. Her campaigning helped to prevent the shutdown of Diablo Canyon, one of the largest nuclear plants in the US, and has happened in parallel with an upswing in popularity for nuclear energy since 2020.¹²

6. Learn by doing (good)

When I founded 80,000 Hours, we hadn’t yet come across the concept of career capital. I mainly compared my options in terms of what I thought would be most impactful in the next few years, only thinking vaguely about what lay beyond that. However, if I had been using the concept, it’s possible it would have led me astray. I would have likely concluded that working in investing would lead to better career capital than starting a nonprofit. But looking back today, I gained better career capital by cofounding 80,000 Hours because I learned more, achieved more, and met a lot of interesting people.

Learning-by-doing is often the fastest way to gain a skill, and this can make it the best option for career capital — and even more so if the work you’re doing is directly relevant to having an impact — most often by failing to do so in the way you’d hoped. If you find you lack a certain skill, you can learn that exact skill.

By working on something impactful, you’ll also likely be more motivated, and people will want to help you succeed. You’ll meet people interested in pursuing the same goals as you. And by focusing on issues that are neglected, it can be easier to achieve impressive results. Many people don’t feel like they have the option to do something impactful right away, but if you do, it can be one of your best options for career capital too.

Don’t focus on what might be most impactful in the next year, since most projects take time to succeed. Suppose your goal is building political support for AI safety. Think about how you could build the most support for it over 3–5 years if you aim high. If you think it’s mass media, then start learning communication skills and apply them wherever you can. If you think it’s grassroots advocacy, then just start organising.

This was the path taken by Sneha Revanur. While in high school, she came across a 2016 ProPublica investigation exposing how algorithmic bias was baked into criminal justice software. The findings stuck with her. When California’s Proposition 25 came up — a ballot measure that would have allowed similarly-biased AI algorithms to determine who got bail — she was moved to act.

She started where she was: mobilising fellow students. They organised voter outreach, raised awareness about the dangers of algorithmic bias in the justice system, and built a network of young people who cared about these issues. After Prop 25 failed to pass, rather than let that network dissolve, Sneha transformed it into Encode, now the largest youth movement focused on AI risk policy.

From those early student-organising efforts focused on immediate issues of AI bias, Encode has expanded to encompass the full spectrum of AI risks, for instance, by sponsoring SB 53, a new law requiring greater transparency about the training of frontier AI models, which was recently signed by Gavin Newsom. And this has also given her amazing career capital.

Doing what everyone else is doing is rarely the best path for impact, and now we’ve seen it’s not the best path for career capital either. Common defaults like Google, Deloitte, or continuing into graduate school can be OK, but the best career capital often comes from opportunities that require a bit more work to find: the startup that isn’t yet famous, the policy fellowship you found by crashing on your friend’s couch, or even that modelling career you were considering.

Up next: how can you figure out which option is best for you?

Read next:  Part 10: How to find the right career for you

Continue →

Or see an overview of the whole guide.

Get the whole guide as a book

If you find this guide helpful, preordering the book (especially from a physical retailer!) is a great way to support us.

Preorder the book