80,000 Hours: How useful are long-term career plans?

By Aaron Gertler 🔸 @ 2019-10-09T01:03 (+11)

This is a linkpost to https://80000hours.org/articles/career-planning/

"Plans are useless, but planning is essential.” — Dwight D. Eisenhower

There are two types of mistakes one can make with career plans: having an overly rigid and specific plan, and having no long-term plans at all. We see both issues in our advising.

In the rest of this article, we give some arguments for and against long-term career planning, and explain how we aim to strike the balance between both types of mistake.

In a nutshell

We find it useful to break career plans into two parts: long-term options and next steps. Your long-term options are the global problems and broad career paths you aim to work on over the next 5 to 25 years. Your next step is what you’ll do over the next couple of years to get there.

How much to plan is a matter of balance. It’s useful to both work backwards from broad high-impact paths you could take in the long-term, and to work forwards from the concrete opportunities that are open to you right now.

We find people often err in both directions: many haven’t spent much time thinking about which problems and careers might be highest-impact to work on in the long term, even though some long-term paths are much more promising than others, and require specialised career capital that can take years to accumulate. We think it’s usually a mistake to have no long-term plan at all.

Others, however, spend too much time analysing different long-term paths when it would be more useful to look for concrete opportunities and submit lots of job applications.

When doing career planning, we usually find it best to start with potential broad long-term options to gain a general orientation, work back to next steps from there, and then to spend lots of effort finding and comparing concrete next steps which you can actually take. We usually spend more time analysing next steps than long-term options.

Ideally, you can also consider whether your bias is towards too much or too little planning, and to try to correct for it.

If you’re early in your career (e.g. under 28), and especially if still an undergraduate or in your first job, then your situation will be changing particularly rapidly, so is reasonable to spend less effort on long-term planning, and more effort on trying out promising paths and building transferable career capital (though ideally still with broad long-term options in mind).

Some arguments against long-term plans

Many of our readers get paralysed thinking about long-term options. It’s easy to see your choice of career as a single decision that you have to “get right” immediately, creating a lot of anxiety. In reality, most of the time you’re only committing to a job for a couple of years, and you’ll have many opportunities to shift course in the future.

Your preferences will also change over your career (more than you think), the world will change (including which problems are most pressing and what the key bottlenecks are), and you will learn a huge amount about your skills and which options are best.

Most people we know who are having a big impact today wouldn’t have predicted they’d be doing the kind of work they’re doing ten years ago.

This means it’s not useful to make detailed long-term plans. Having an overly detailed long-term plan might even cause you to overly fixate on that one path, and miss a great opportunity that doesn’t neatly fit within it.

This is especially true when you’re early in your career. In your first couple of jobs, you’ll learn a lot about what fits you best, and this will probably let you significantly improve your thinking about which long-term paths are best.

So long as you keep taking these sensible next steps — jobs that give you career capital, and help you learn about what fits you best, and that steadily have more and more impact — then you’ll build a good career over time.

In other words, it’s possible to work forwards from the opportunities that happen to present themselves, and by taking steps that put you in a generally better position, even if you’re not sure how it’ll all work out eventually.

Moreover, this approach can be very useful, because the specific opportunities vary so much within a path. For instance, if you apply to jobs in a broad path like “AI policy”, some of the specific positions you’ll find will be far better than others, because there’s so much variation among specific roles, organisations and teams.

The differences between jobs within a path can even be larger than the average differences between paths (e.g. AI policy vs. AI safety engineering), and are not predictable ahead of time: good luck could throw up something unusually good and vice versa.

This means that finding out which specific opportunities that happen to be around is often more important than analysing which paths are best in general.

It’s also much easier to compare specific opportunities as they become available, rather than paths in abstract.

Expanding on this, one of the most useful strategies is to simply pursue lots of specific jobs and make lots of applications. We often come across people who are agonising about different options, whereas if they’d simply made applications, the next step would have become obvious, and they could have avoided analysing options that turned out not to actually be on the table.

Some arguments in favour of long-term plans

All this said, we don’t think that means you should ignore which long term options might be best, and think it’s usually a mistake not to have a long-term plan at all.

In general in life, it seems more useful to have a plan, even if it’s very uncertain. This still seems true even in rapidly changing, adversarial domains, like business strategy, military strategy, or politics.

This is because having some specific long-term plans in mind might help you spot cheap ways to direct your next steps in a better direction. For instance, it can help you come up with better ideas for next steps, accumulate more relevant career capital, and carry out tests of potential long-term options.

Sometimes you can avoid putting lots of time into going down a path you could confidently rule out if you thought about it more.

In fact, we think that for people who want to maximise their impact, it’s likely more important than normal to have long-term plans, and it’s usually a mistake not to have any.

A major reason is that when we look at the highest-impact opportunities, they usually require some measure of specialised career capital. For instance, you normally need some knowledge of the specific global problems you want to work on, such as AI Safety or global priorities research. We’ve argued that the most effective areas are far more effective than the median, but they’re also unusual areas, so you’re unlikely to gain the relevant knowledge and connections by chance.

The highest-impact roles also often require other forms of specialised career capital, such as experience in particular areas of policy, or research skills in specific fields, or connections with people who work in these areas. (Though, as always there are exceptions: management is an important skill that seems fairly transferable, so you could focus on learning that early career and then apply it to pressing problems later.)

Although it could be reasonable to focus mainly on transferable career capital in your first couple of jobs, eventually, if you want to maximise your impact, you’ll probably need to accumulate significant specialised career capital as well, and having a long-term plan helps you do this efficiently. In particular, this plan should include some hypotheses about which global problems you’ll ultimately focus on, and what types of roles you might take to address them.

Having a long-term plan is even more important if you’re focused on an area where most of the impact comes from being near the top of the field, and it seems plausible that many of the areas we talk about, like academic research, entrepreneurship and politics, are often like this.

Making it to the top of a field usually requires years of dedication. If most of the rewards also come from being at the top, and you’re more likely to get there if you spend longer honing your skills, that suggests it’s sometimes worth taking an early gamble on a long-term path that might take off.

For instance, we’ve heard that doing something outside of academia early career can make it seem like you’re not serious about academia, which can be held against you when applying to top PhD programmes. Moreover, the chance of making tenure depends heavily on your publication record at the end of your postdoc compared to your peers, and so taking any time off can jeopardize your chances of being competitive.

In contrast, if your aim were to minimise personal risk rather than maximise your impact, long-term planning is less important. You could simply accumulate transferable career capital which would increase your chances of finding a job that’s ‘pretty good’ in the future.

If instead you’re aiming to keep open the possibility of hitting the peak of a field, you’ll want to try to accumulate career capital in a coherent direction.

Many people also find that having a long-term vision to work towards is motivating, and this also makes it more likely you’ll eventually make it.

All this doesn’t mean you should try to only do one thing as soon as possible. It’s also important to try out different long-term paths to work out which is best, and lots of successful people explore several fields before settling on one. However, it does mean that early career you should ideally be trying to test out long-term options that might be great (so long as you can do so with a decent backup option).

This feels pretty different from a vague plan to ‘build career capital, any career capital’, which is we sometimes come across. (Sometimes caused by content from the early days of 80,000 Hours, which maybe hammered home the ‘don’t worry too much about long-term plans’ message too much.)

Today, we think it’s usually a mistake to not at least have some long-term plans, even if they’re vague and tentative.

That said, there are sometimes exceptions, such as people who are very new to thinking about which global problems are most pressing, are still undergraduates or perhaps in their first 1-2 jobs, or have very limited time to make a decision.

Weighing the arguments

Overall, we find it helpful to both work backwards from the best long-term options, and to work forward from the particular opportunities you face today.

We typically start with long-term plans, because they provide orientation, but when advising someone, we’d generally spend more effort identifying and comparing specific next steps.

As noted, we notice mistakes in both directions, and it’s important to strike the right balance between these the ‘big picture’ and ‘next step’ perspectives. As a reader, it’s worth thinking about whether you tend to plan too much or too little, and adjust where you focus accordingly.

If you lack a long-term plan, and haven’t thought much about which global problems are most pressing, then it’ll probably be a useful exercise to spend some time on that.

But if you tend to over-analyse (as many of our readers do!), then you may need to focus more on just making applications and seeing when turns up, and then taking a job for a couple of years, seeing how it goes, and learning as you go.

The ideal is to end up with some long-term visions for your career that are reasonably stable, and in sync with the concrete next steps you’re setting yourself up for — but it can take a little while to people reach that point.

What should go into your long-term plan?

An ideal long-term plan will focus on elements of your career planning that are relatively stable, but also relevant to your choice of next steps.

We find having some hypotheses about which global problem areas you want to focus on is usually a useful element of a plan (e.g. AI safety, improving international governance or climate change). This is because problem areas are specific enough to help you pick out a long-term direction that’s much higher impact, and therefore to help you spot good next steps and build career capital in a coherent direction. However, problem areas are not so narrow that small changes in the situation will mean you need to change your plan (whereas focusing on a specific organisation or intervention is probably too narrow). We’ve also found that in practice we and many other groups in effective altruism find problem areas a useful unit of analysis. This old blog post is out-of-date, but contains some more thoughts on this topic.

The other most useful element of a long-term plan is something about the type of role or skill-type you want to use long-term, such as economics researcher, entrepreneur or expert in operations. Your choice of problem is about what you hope to change in the world, and your role is how you hope to change it. Careers are often built around a skill type, so this seems like a useful element to focus on. Again, the aim is to pick out a set of ideas for your long-term role that are narrow enough to provide direction, but not so specific that they will change too often.

If you have more time, it can also be useful to write about the moral values and big picture principles that underlie your choice of problem area. These tend to be more stable than your choice of problem, though are too high-level to be applied to most decisions. We’d also encourage you to write out the ingredients you want in a job from a personal perspective.

If all this gets too much, come back to your next step, and make it a good one.