This post is a product of the Red Team Challenge organized by Training for Good. It was written by Kayode Adekoya, Lynn Tan and Michael B. We very much thank Cillian Crosson, Jennifer Zhu and Oscar Galvin for their feedback. We hope this post can be a modest but useful contribution to the ongoing question of how to make good career decisions. Please do provide your criticism and feedback in the comments.
- Gaining career capital and doing good are not mutually exclusive. Thus, for making decisions, what matters is the marginal career capital one option offers over others.
- The unit of career capital is hardly defined.
- Even if we say that though a difficult one it is not an impossible unit, it remains hard to measure its future consequences in individual cases because career paths are often messy and complicated.
- Despite these concerns we can still see career capital as a useful conceptual tool for making career plans and decisions. But then we need to treat it like a tool in cooking rather than one in science.
- For exceptionally talented people or those in certain fields (arts, theoretical math & physics, community building or AI safety due to urgency) or those with some exceptional opportunity at hand, focusing on direct work might well be the better option.
- Career capital has implications for community building efforts:
- There are concerns about value drift when people focus on career capital
- Career capital can offer an option valued by the community for people not currently able to do direct work
- Career capital is one of the things we should consider for community-wide coordination
Professional life is like climbing a large cherry tree. You make choices as to the branches you go onto. Each choice opens up new branches to go onto but rules out others. Each opens the path to another group of cherries. Sometimes you can jump over or reach out to an adjacent branch. Sometimes, getting cherries is not all you care about but also safety or the view from a branch. Sometimes the best fruit can be reached right at the bottom though sometimes we have to climb higher or go further out on the branches. Sometimes, you climb high and the fruit there turns out to be bad. Maybe the fruit at the bottom is not quite ripe yet. If we are a group of friends on a cherry-picking excursion, questions of coordination and motivation become important.
That “early on in their careers, EAs should focus on gaining (the right sort of) career capital” is a commonly held belief in EA influencing the career decisions made by young, promising people. It also seems implied in 80000hours’ article on career capital.
In this post, we begin by providing a definition for career capital and then think through some of its implications rather than directly responding to or criticizing any specific article or detail. We adopt the perspective of young individuals trying to make a decision between their available career options as well as the perspective of the EA community by including community building considerations.
We tried to be as generous to the idea of focussing on career capital as we could without becoming dishonest or simplistic. In the end, we think that including career capital as consideration in career decisions can indeed be beneficial. However, in order to be as useful as possible and not be misleading, our thinking and communication about this idea must include additional considerations. Otherwise we might understand career capital to be something it is not: a simple and catch-all answer to the complicated landscapes of professional life.
By the definition we propose below, the right sort of career capital is inherently useful for career goals. But before we get there, we should remind ourselves that
- Anything (project, job, program etc.) you do will have some impact on the world,
- Anything you do gains you some sort and level of career capital,
- Not every capital is useful towards every goal,
- Focusing on career capital can only be seen as useful for EA purposes insofar as it allows you to have more aggregated impact in the future compared to that future self which would otherwise have chosen to focus on direct impact now (including the impact you would have made in the years you spend accruing career capital).
What is career capital?
How might we define career capital? Capital can broadly be characterized as any resource available for the pursuit of goals. In that light, career capital is any resource available to an individual’s pursuit of a career goal. As examples, 80000hours lists skills, connections, credentials and financial resources. These are clearly not of the same kind (human, social, cultural and economic capital). But to think about the topic, we try to convert them to the terms of a sort of career capital equivalent or a unit of career capital.
However, a unit of career capital is hardly defined. Any given configuration (a “portfolio”) of career capital opens an individual to a specific configuration of possible career options in the future (with probabilities of attaining each). In the economic world, pretty much anything can be sold and converted into the universal exchange-enabler of money. With career capital, that is less universally possible. In the marketplace of society, career capital can sometimes be exchanged for some career opportunities but not all types of capital will give you access to all kinds of opportunities. What, then, do we mean by a unit of career capital? The number of opportunities it opens? Their quality in terms of impact? Their safety? A universal career capital equivalent or unit of career capital seems rather hard to define or measure. 
When an individual makes decisions, they should therefore consider the specific capitals involved and the specific opportunities those capitals provide access to. Really knowing a position, an organization or a degree is incredibly important. That’s also why Holden Karnofsky suggests taking a junior role first before going to grad school. Asking for informational interviews with people you find interesting could be another low-stakes option. Flexible career capital might seem more attractive and is more intelligible as a unit because it’s more universally exchangeable. But it is important not to settle on the most general options like consulting and to also consider alternatives because the most general option is likely suboptimal for each specific path. Additionally, some estimate on how likely it is that your plans turn out the way you intend them to should be included in your thinking, as well as backup options.
Career capital when making decisions
A Conceptual approach
When making decisions between options, we should think of the marginal career capital and marginal immediate impact associated with them rather than the absolutes. See the following scenarios:
When choosing only between A and B, we should choose A because it offers marginally more immediate impact. When choosing only between A and C, we should choose A because it offers marginally more career capital. But when choosing only between B and C (if A is not available to us), which one would we choose? B offers marginally more career capital while C offers marginally more immediate impact. What is the expected future value gained by choosing the option offering marginally more immediate impact? What is the expected future value gained by choosing the option offering marginally more career capital? Combined: What is higher, the expected value of a unit of career capital or of a unit of direct impact?
Note that it is not impossible that at times any two options could lead to the same expected value (in terms of impact) - in which case, choosing between A or B could boil down to other factors such as personal preference. Arguably, if other factors are negligible, one should choose the option that marginally has more impact now. But where expected values differ significantly, we should choose the higher. In practice, it seems difficult to estimate precisely how much direct impact you will have now or how much future impact you will have from gaining career capital. The uncertainties and indirect effects involved are significant. Thus, these questions become starting points to seriously consider what the specific options you are thinking about entail rather than leading to a precisely measured answer. This is what we mean when we speak of a conceptual tool. The point might ultimately not be to precisely measure career capital or direct impact but to seriously think about the options and make informed guesses as to what they might entail or lead to.
A practical approach
The 80000hours article on career capital says that “People seem to become dramatically more productive over their career. Our impression is that most people have little impact in their first couple of jobs, while productivity in most fields seems to peak at age 40–50.” This indeed seems true in many scholarly fields or sciences like chemistry. It is also supported if we think that in the US earnings somewhat reflect productivity and see average earnings increase by age. But it seems plausible that much of the increase in earnings could be attributed as much to social hierarchies or historical precedence as to increasing productivity or competence.
But the key word in the quote is “most.” There are indeed cases where productivity may be highest or high starting from a young age. This appears especially true in fields and situations like
- Theoretical physics: Einstein wrote groundbreaking papers at age 26, 1905
- Pure mathematics: Descartes conceived of analytic geometry at age 23, 1619
- Music, visual art and poetry: Mozart died at age 34 (though a prodigy child, later works are superior, e.g. Serenade No. 10 (1781/2) written at age 25)
- New technologies or opportunities:
- Sam Bankman-Fried founded FTX at age 27, 2019
- Political activism: Greta Thunberg with Climate Strike/ Fridays for Future movement arguably shifted public and political focus toward mitigating climate change, currently age 19
This does not mean that you cannot make significant contributions later but does mean that it seems at least possible to make them while young. The mean age at which Nobel prize winning research in physics was done was still 37 from 1901-2008.
So, if you are exceptionally talented or have chosen the arts, theoretical math or physics, you may be most productive or creative earlier in your life. Machiavelli sees success as the result when opportunity (fortuna) meets skill (virtu). So, when you have an exceptional opportunity or are exceptionally lucky, you may be able to live up to those circumstances and bring about great things when you are young.
On the other hand, there are some roles that are perceived to need certain skills and experiences in order to be done well. Holden Karnofsky writes that “Career capital building for longtermist causes will likely swamp any donations you can make, because this area is primarily constrained by people with senior leadership abilities such as: entrepreneurs, managers, research leads, project leads, hiring lots of people, new ways of spending funding (without lowering the bar for funding).” The idea here is that some important roles can only be filled after candidates have gained a lot of relevant career capital.
However, we should hire someone based on whether they can do the job. This may demand a lot of formal career capital. But it may also mostly depend on some sort of inherent talent, education or other qualities. How people are hired depends on the organization. Older, more bureaucratic organizations will likely have different practices than younger, more start-up-like organizations. Work at younger institutions (or founding your own) may well be the type of work which both offers a lot of career capital and could have high expected value, e.g. Charity Entrepreneurship and their incubated charities.
Further, Some career capital or direct work comes at little cost and can help reduce the margin between any two options. For example, when pursuing a degree you can try to choose meaningful thesis topics (i.e. the idea behind Effective Thesis). When working at a company to gain career capital, you might be able to volunteer some of your time or work on interesting side projects. Networking is another great low-cost career-capital activity.
At bottom, some of the question of gaining career capital or not seems to stem from the implied demandingness of EA. There are people and animals suffering all around us right now. AI doom might be just a few years ahead. So, all else being equal, it is only reasonable to try to do good sooner rather than later. Another consideration to that point could be compounding benefits from some earlier work. The question of career capital then becomes one of how willing we are to delay our moral contribution to the world in order for that contribution to be bigger - or how much in a rush we actually are (i.e. AI timelines).
Implications for community building
The danger of value drifting when focusing on career capital is significant and underemphasized or underexplored. There are multiple estimates about value drift, including More empirical data on 'value drift' - EA Forum. When someone begins to take on roles outside of EA mostly with the aim of gaining career capital to later have more impact, they might in the meantime change their aim. Even if the extent to which this occurs is not huge, or if in most cases it is not severe, one would think that it is better to reduce value drift if we aim to capture ‘exceptional talent’ due to power law distributions. Also, the ideal time to switch back into EA is hard to gauge and will always include making a potentially risky step. That might lead to people never making it or making it later than we would hope. We should do more to mitigate such problems by encouraging people who focus on gaining career capital to take steps to stay connected with EA. That could be building a network within EA, being involved with community events or doing side projects within EA.
Focussing on career capital as a valued option can help reduce frustration in EAs. Getting jobs at explicitly “EA” organisations (or in roles relevant for EA) has historically been seen as hard and led to frustration in some community members. If gaining career capital outside of EA is seen as a valued option by the community, that can help reduce such frustration. We also need to consider that for many EAs there might be additional considerations which currently prevent them from taking the most risky or long-shot options. If the community sees “career capital jobs” which can help build reputation, skills and financial security as valuable, that can help us keep such people connected to the community. Gaining career capital outside of EA and then switching to work on what we currently think is best also diverts resources towards EA causes.
Not only the impact of individuals matters but also that of the community. From the collective perspective we can look at neglected paths, skills or pursuits. According to the 2020 EA Survey, around two community members prioritize career capital per person prioritising immediate impact. Todd writes that “Overall, [his] guess is that we’re only deploying 1–2% of the net present value of the labor of the current membership. This could be an argument for shifting the balance a bit more towards immediate impact rather than career capital – though this is a really complicated question. Young people often have great opportunities to build career capital, and if those opportunities increase their lifetime impact, they should take them, no matter what others in the community are doing.” An alternative way of changing the current allocation might be more corporate or professional outreach, increasing our average age and amount of career capital.
There is an underlying question here about what sort of community we want to be. Are we in a rush? How urgent are the things we care about? How do we handle on the one hand aiming to maximize collective impact while on the other hand valuing individual life journeys? AI timelines and the moral demands of current suffering are important considerations. Short AI timelines demand direct work now while longer ones allow for more building of career capital. Are we a community where you need to become “more and more ‘involved’” or is it okay to circle in and out of EA, periodically checking new ideas or positions that could be a good fit, now that you once again gained more experience?
Most of the value of the idea of career capital could come from people taking a strategic approach to their lives and careers. It is useful to think through different options, their implications and the access they offer. The oversimplified statement “Gaining career capital is good” in a first iteration could lead to people thinking that career capital is unambiguously what any young individual should prioritize (over other considerations like impact). In a second iteration, they might discover more specific options which gain similar transferable skills but also more specific capital allowing work in a specific area later and perhaps even already do some or a lot of good. Our aim should be to push people further down the additional considerations.
80000hours is quite explicitly including these additional considerations by shifting the question from “should you focus on career capital?” (which might have led to the strong focus on flexible career capital in the past) to “how best to invest in yourself?” From the 2017 career capital article Which jobs put you in the best long-term position? - 80,000 Hours to the current one, there was a shift explained in Before committing to management consulting, consider directly entering priority paths, policy, startups, and other options - 80,000 Hours (an excellent article on the topic).
We should also further emphasize the community building implications of career capital. On the one hand, career capital is an idea that might help new people become accustomed to EA principles while pursuing something which provides backup options and safety. On the other hand, career capital is also a vital consideration from the collective perspective: what types will we need in the future?
Additional questions to explore could be:
- How large is the danger of value drift when focussing on career capital in non-EA environments? If it is - why or how is it a danger? How can it be mitigated?
- What patterns concerning career capital do we see when we look at people about whom we think that they had or currently have an outsized positive impact on the world?
Cooking here is an example for an artistic, creative practice and science for a rigorous, technical, precise method. The transition from one to the other might be gradual and fluid.
80000hours writes in their career guide that “If you don’t feel like you have much leverage now, another option is to build what we call your ‘career capital’: skills, connections, and contributors to your reputation that will let you have more influence in the future.” In their separate article, career capital is “a key strategic” and “potentially vital consideration” (This is your most important decision - 80,000 Hours; Career capital: how best to invest in yourself - 80,000 Hours ) while they say about the job board that it lists “current job opportunities that can let you have a big impact, or build career capital for high-impact roles.” (This is your most important decision - 80,000 Hours )
According to the sociologist Bourdieu who builds on Marxist analysis.
If someone has a good idea, we’d be excited to hear it! There might be sensible ways to think of a unit of career capital but we’re not quite getting there.
Holden Karnofsky, My current impressions on career choice for longtermists - EA Forum
80000hours substantiates this notion in footnote #5 in This is your most important decision - 80,000 Hours. The claim is based first on average earnings by age in the US and second by analyses of achievement in different fields. However, the key insight that different fields significantly vary in peak age is unfortunately not mentioned in the main article.
B. Jones, B. Weinberg, Age dynamics in scientific creativity. Argues that the more important factor is time rather than field since training, institutional and scientific patterns have changed over time within fields. Theoretical work is still done at younger age than empiricist work but in physics, theoretical work is currently less prevalent than in the early part of the 20th century.
And, importantly for the EA context, he might at times suggest that skill can have the upper hand. Machiavelli’s ideas on these topics permeate The Prince. If you think it tasteless to mention Machiavelli because of the negatively connotated adjective machiavellian, you should read the book.
Gaining career capital outside of EA may divert resources to EA causes. When people build career capital, that often costs economic capital. For example, if someone is hired by a company but first has to learn the skills they need, that might cost the company time and money. The mistakes they might make early on will also have costs. So, if EAs gain career capital outside of the EA community and then re-enter, they actively divert resources into the community. But if EA is talent-constrained, it might currently make relatively more sense than in the past to develop talent within EA rather than outside and “pay” for that development ourselves.
Thank you for writing this article, especially the section with implications for community building. I think I have been underestimating the effects of value drift. But I can imagine that if someone gets a tech job and is earning six figures, they might be happy to, let's say, stick with ETG instead of jumping through hoops and changing careers. Thinking out loud here — I will give this more thought.
On a different note, probably good to strike this out as an example: