Career Choice Should Be Treated as an Empirical Search for Impact

Benjamin Todd starts from a simple premise: a typical career contains about 80,000 working hours — 40 hours a week, 50 weeks a year, for 40 years. That makes career choice one of the largest decisions most people will make, especially if they care about their effect on the world.

But Todd’s argument is not that everyone must solve the same problem, or that a good life requires discovering a single perfect vocation early and sticking to it. He says the name “80,000 Hours” can be misleading in that respect. A career is “ultimately a series of next steps.” People can build a good career over time if each step teaches them something, does some good, and helps them learn more about what fits them.

That matters because much career advice, in Todd’s view, is too thin for the decision it is meant to support. He says the advice people receive is often little more than slogans: “follow your passion” or “keep your options open.” There is more guidance on how to apply to a consulting job than on whether consulting is a path worth pursuing in the first place.

Russ Roberts presses the point from another direction. He says the standard high-status routes — law, consulting, finance, big tech — can be more than merely uninspiring; some may fail to make the world better, or may even impose costs. Todd’s own opening case is that many students from top universities cluster in precisely those areas, often working very hard in jobs that can be boring, while more meaningful and more useful paths remain underexplored.

Todd’s broad recipe has three stages. First, explore promising paths. Second, build valuable skills. Third, use those skills to tackle problems that matter while also getting what one personally needs from work. That framework rejects both panic and passivity. The young person does not need to know the final destination before taking the first step. But neither should they assume that career choice is only about preference, prestige, or pay.

The core question Todd asks is not simply “What do I want?” It is closer to: What problems are big, neglected, and tractable enough that my work could matter? And among the possible ways of helping, where might I have unusually good fit?

Todd’s criticism of “follow your passion” is not that intrinsic motivation is worthless. He says the true part of the slogan is that being intrinsically motivated is good. The failure comes when the slogan is interpreted as a prescription to match a career to a preexisting hobby or interest.

In surveys Todd cites, young people’s leading interests often cluster around sports, art, and music. That is unsurprising, but it creates a problem if people infer that a fulfilling career must be built directly out of those interests. Someone who loves literature may conclude that they must become a writer. Someone with a biology major may assume they must work in biology. In Todd’s view, both are unnecessarily restrictive conclusions.

The restriction is not only psychological. It also pushes people toward some of the most competitive fields. Todd notes that only a small share of jobs in the economy are in art, sport, and music — “a couple of percent” — while many people are passionate about the same small set of domains. The people who succeed in those fields then become the visible examples. They tell others to follow their passion, but they are the outliers who made it. Roberts adds that the people who did not make it usually do not get the platform.

Todd’s alternative is to look less at job titles and more at the conditions that make work satisfying. Roberts summarizes the predictors of job satisfaction from Todd’s book: freedom to decide how to perform one’s work; clear tasks with a well-defined start and end; variety in the nature of tasks; and feedback that tells people how they are doing. Todd adds that fulfilling work helps others, is something one is good at, and is done alongside supportive colleagues.

The important implication is that satisfaction often depends more on context than on the nominal field. Todd gives the example of financial administration. It may sound dull, but in the right team, with clear tasks and a meaningful goal — for example, in a charity one cares about — it can be engaging. Conversely, someone may love motocross racing and still be miserable if they have a bad boss.

Roberts draws out the management lesson. A leader who gives people autonomy, clear tasks, variety, and feedback is not merely being kind; they are shaping the conditions under which people can do better work. He describes part of his own role as president of Shalem College as trying to ensure that people use their “superpowers” rather than spending too much time on tasks that drain them or do not use their strengths. Todd connects that to “job crafting”: designing a role to make it more engaging and meaningful.

The broader point is that personal fit is discovered rather than divined. Todd’s advice is not to abandon interests, but to generate a short list of paths that could be good for the world and fulfilling, then choose among them partly by asking where one is likely to be best.

Roberts introduces a classical economic intuition with caveats: under certain assumptions, salary is related to the value a worker creates. He immediately qualifies the point. The assumptions often do not hold, and the job that pays the most does not necessarily create the most value. Still, he asks whether Todd’s focus on big problems undervalues forms of work that delight, comfort, or solve smaller problems for individuals.

Todd agrees that there is some correlation between income and positive social impact, but he thinks it becomes weak when asking which work could have the most impact. The most neglected issues are often outside the market altogether. He points to future generations and factory-farmed animals. Future people do not participate in today’s markets. Factory-farmed animals do not either. So there is little reason to expect that simply earning more in the market will address their interests.

Roberts sharpens the economic caveat: markets reward people for giving people what they want. Sometimes those wants are unattractive. Sometimes they impose costs on others. Sometimes they are shaped by regulation or perverse incentives. Sometimes they are addictive or self-destructive. He is less comfortable than he once was, he says, with simply defending the satisfaction of preferences when what someone wants may harm them.

The tension becomes concrete in Roberts’s example of Jeff Bezos and Amazon. Roberts says he feels gratitude for Amazon’s early role in making books easier and cheaper to acquire. If that were all Bezos had done — if he had made books less expensive and more accessible without saving lives — would that be a life well lived?

Todd’s answer is not dismissive. Making the world wealthier is better, he says, and saving lives is not the only way to contribute. But he raises two reservations. First, in winner-take-all markets, the person who gets there first captures enormous value, while someone else might have done much the same thing later. That makes the counterfactual important. Second, the money earned from a successful business can itself become a major source of impact if used effectively.

Roberts remains interested in defending smaller, human-scale contributions: businesses that provide meaningful work, products that bring pleasure, institutions that give people opportunities to use their skills. Todd does not deny that these matter. His claim is comparative: for many people, there may be options with far greater impact than running a conventional business, especially when markets systematically neglect those affected by the work.

One of Todd’s more counterintuitive examples is medicine. Roberts calls it a “contrarian” argument: many people assume being a doctor is among the most obviously valuable careers available. Todd clarifies that he does think medicine has an important positive impact compared with most jobs. His claim is that someone capable of becoming a doctor may be able to have much more impact elsewhere.

Todd gives several reasons. First, he says health care has increased lifespans by a couple of years, which is valuable, but most historical gains in life expectancy came from public health, sanitation, and wealth. The additional years attributable to health care are not the work of doctors alone; they are shared with nurses, hospitals, equipment, and the broader system.

Second, Todd emphasizes diminishing returns. A graph of doctors per population against ill health, he says, shows that the first additional doctors in a population have enormous impact, but the curve quickly levels off. Rich countries are largely on the flatter part of the curve. Todd notes that Robin Hanson has argued that in the United States marginal health care may even be slightly counterproductive because of side effects, though Todd does not go that far. His own position is that marginal doctor work remains positive but subject to large diminishing returns.

Third, he considers replacement. If one person does not go to medical school, someone else will likely take the place, especially where the number of places is fixed. That does not reduce the marginal impact to zero. If the admitted candidate would have been a better doctor than the next person, the field improves somewhat. But it attenuates the intuitive picture in which “I performed these surgeries” means “I saved those lives.” Many crucial surgeries would have happened anyway.

Todd says the estimate is uncertain, “probably within an order of magnitude,” but the book arrives at roughly one life saved for every 10 years of work as a doctor in the United Kingdom.

Roberts accepts the force of marginal analysis but pushes back on what it misses. He tells a story of his wife visiting the NHS after injuring herself in London and receiving reassurance from a doctor. That interaction may not count as life-saving, but it mattered. More broadly, Roberts argues that even where the structural impact of a job is smaller than people imagine, individuals can still change the experience of those they meet: by being kinder, more effective, more understanding, and more humane when delivering bad news.

Todd agrees that being a nice person in day-to-day life is important. But he returns to scale. There are civilizational challenges, he says, and there is a lot people can do about them. For him, the point is not that small acts are worthless. It is that a career strategy focused only on immediate visible help may miss much larger opportunities.

Todd’s framework depends on the claim that impact is difficult. Wanting to help is not enough; interventions can fail or backfire. His example is Scared Straight, a U.S. program in which young people who had committed misdemeanors were taken to prisons to be shown the consequences of criminal life. The hope was that the experience would scare them onto a better path.

Todd says trials found that participating youths did commit fewer crimes afterward, which made the program appear successful. But compared with a control group who had not gone through the program, the decrease was smaller. Relative to the counterfactual, the intervention made outcomes worse.

The reason is uncertain. Todd says perhaps the program normalized prison life, perhaps the youths looked up to the criminals, or perhaps prison was not as bad as expected. Roberts suggests another possibility: they may have received useful criminal advice from prisoners. Whatever the mechanism, Todd says the program persisted for decades, received large amounts of funding, and had supporters who believed it worked. It even had an award-winning documentary.

The example is important because it separates sincerity from effectiveness. Many institutions and programs can point to apparent improvements before-and-after. Todd’s claim is that the counterfactual matters. The question is not only whether things improved after an intervention, but whether they improved more than they would have without it.

Roberts sees the value of this discipline in the effective altruism movement: trying to quantify impact and insisting that feelings are insufficient. But he also stresses how hard the project is. Large cultural and social problems do not respond simply to the desire to make them better. He gives climate change and COVID-era public-health messaging as examples in which the identity of the messenger may change how people respond. A politically salient advocate can mobilize one audience while alienating another. The desire to change the world can collide with polarization, trust, and human psychology.

Todd’s response is not to make impact seem easy. He says difficulty is a theme of the book. But the fact that many interventions fail does not imply that none work. It implies that choosing carefully matters.

Against the examples of failed interventions, Roberts highlights another theme from Todd’s book: people who were not celebrities, not household names, and not necessarily geniuses in the popular imagination, but whose work made enormous differences.

David Nalin is one of Todd’s examples. Todd describes Nalin working in India, in refugee-camp conditions where many people were dying from cholera and diarrhea, largely because of dehydration. The standard treatment was intravenous fluids, but IV treatment was difficult to provide at scale in refugee camps. Nalin helped show that a solution of water and salt in the right concentration, sipped orally, could rehydrate people far more effectively than ordinary water and reduce mortality almost as much as full intravenous treatment.

Todd does not claim Nalin deserves all the credit for the subsequent global impact. Someone else might have discovered or scaled the intervention eventually. But when the intervention is associated with reductions in child deaths from diarrhea and dehydration by millions per year, even speeding it up by a small amount has enormous value.

Roberts connects Nalin to Ignaz Semmelweis, who recognized that doctors should wash their hands after going to the morgue and before delivering babies. Roberts notes the tragic delay: Semmelweis’s findings were dismissed in part because he was difficult and because, seeing the conclusion as obvious, he did not make the case rigorously enough. Women continued to die from puerperal fever. For Roberts, this is a cultural failure as well as a scientific one: societies should spend more resources honoring such people.

Todd adds that culture can even invert praise and blame. He says when someone like Mark Zuckerberg spends $50 million on a house or yacht, little is made of it. But if he donates large sums to medical research, he receives substantial criticism. Todd acknowledges objections to inequality and billionaire philanthropy but argues that people should still be encouraged to do more good when they can. When Roberts asks why such giving is criticized, Todd says many people dislike billionaire philanthropy as undemocratic and suspect ulterior motives or reputation laundering.

Karl Landsteiner provides another example of low-fame, high-impact work. Todd describes Landsteiner’s discovery of blood groups. Before that discovery, transfusions could save a patient who had lost blood during surgery, but incompatible blood could also kill them. Recognizing blood groups made transfusions safe by allowing matching. Todd says that enabled many millions of operations each year that would otherwise have been too dangerous.

The examples function as a corrective to both celebrity culture and simplistic career prestige. The people whose work most changes human welfare may not look like the culture’s heroes. Sometimes the central achievement is a cheap method, a classification, a sanitation practice, or a way to make an existing treatment scalable.

Todd’s framework asks people to think broadly about how they can contribute. Direct work in a charity, social work, medicine, or policy is not the only route. “Earning to give” is the idea that some people should enter higher-earning careers, ideally not harmful ones, and donate a substantial share of their income to fund work that others are better placed to do directly.

Todd is careful about the moral condition. He says he would not advise someone to earn to give through a harmful career. But if the work is morally neutral, or positive because it builds a useful business, then donating can enable others to take jobs in high-impact organizations that lack funding.

His example is Fred Mulder. As a student at Oxford, Mulder wanted to make an impact and considered becoming a social worker. He thought he would be a poor fit. While in Paris, he found a chest of Picasso prints and realized they could be resold in the United Kingdom at a significant markup. He used his graduate stipend to buy them, sold them for a profit, and eventually became an art dealer. Over his career, he donated money and helped create the Funding Network, a giving circle that encouraged others to donate. Todd says the network raised many millions of dollars for charity and enabled many people to take charity-sector jobs they otherwise could not have taken. Mulder believed he had more impact that way than he would have had by becoming a social worker himself.

Todd’s advice for charitable giving follows the same problem-first logic. Ask which problems are biggest and most neglected. Then ask which organizations are doing strong work on those problems and have room for more funding — for instance, organizations that have people willing to work for them but lack the money to pay salaries.

As a lower-bound example, Todd mentions GiveDirectly, which transfers money by mobile phone to some of the world’s poorest people, who he says typically live on about $500 a year. The average U.S. college graduate, he says, earns about $70,000 a year over their life. Donating 10% of that would be enough to double the income of several people or households each year among the world’s poorest. Because money goes further for the very poor than for people in wealthy countries, direct transfers can have substantial effects.

Roberts presses the obvious behavioral puzzle: if this case is so strong, why do so few people do it? He says he tries to give 10% of his income, partly because of Jewish religious obligation, and notes the historical resonance of tithing. But most people do not give 10%, even when they would endorse the idea of helping desperately poor people whose lives could be transformed.

Todd thinks social norms matter. In cultures where tithing is normal, many people tithe. He also notes a curious asymmetry: people often accept large implicit salary sacrifices for socially valued careers, such as teaching, but react with alarm to explicit charitable giving. Someone might take a two-thirds pay cut to become a teacher and be seen as virtuous; yet when Todd told his mother he planned to donate 10% of his income, she worried that it would be a real hardship.

Roberts proposes several explanations: distance, lack of feedback, skepticism that the money will be used well, and ordinary self-interest. If a donor saw the recipient every day and received visible gratitude, the act would be emotionally different. Giving to a faraway person through an institution requires trust, and the donor may not know whether the money reaches the right person or helps in the right way. Roberts also argues that human generosity is often powered by connection: family, neighbors, religion, culture, community. Asking people to help distant strangers who provide no feedback is “impossibly bold” because it asks them to extend concern beyond the usual human mechanisms.

Todd accepts that many charities may not work or may be hard to evaluate. But he says if at least one works, then it is genuinely possible to make a difference. And because neglectedness is central to impact, the highest-impact causes will often be precisely the ones that do not pull the heartstrings: abstract, distant, or invisible. Helping future generations is one example. The beneficiaries cannot thank you. Yet if the work matters, the absence of feedback does not make the people helped less real.

Todd’s practical response is to separate sources of meaning. People who work on abstract causes can still have fulfilling careers if their daily work is engaging, their colleagues are strong, and the tasks use their abilities. One can take a big-picture approach to impact while making the everyday work psychologically sustainable.

Todd’s final synthesis returns to the career sequence: find work that builds valuable skills, apply those skills to something that helps others, and make sure the work also provides the personal conditions one needs — engaging tasks, good colleagues, and sufficient salary rather than maximum salary. Before committing, explore. Especially early in a career, it is hard to know what will fit without trying things.

He uses the phrase “career capital” to broaden the idea of skill-building. Career capital includes skills, but also connections, character, and reputation — the assets that make someone a more effective actor in the world. The point of accumulating career capital is not simply personal advancement. It is to make later contributions more valuable.

Todd also gives his own view of the most pressing problems. He thinks the key issue “right now” concerns AI, particularly the possibility that AI systems could do AI research and development within the next couple of years. That could change society and create downstream problems: loss of control, concentration of power, engineered pandemics, and wholly new technologies.

His neglectedness comparison is stark. Todd says there are around 10 million people working in climate jobs around the world, though he adds that ideally there would be even more. By contrast, the AI-related and adjacent civilizational problems he lists are staffed in the thousands. In his framing, that is roughly a thousandfold difference in people working on the issues, despite the possibility that these are among the biggest challenges civilization could face.

Todd also mentions factory farming and global health as other problem areas discussed in the book. The ranking is not meant to erase personal fit. Once a person has identified important and neglected problems, they still need to ask where they can contribute well. The right path depends partly on what they are good at, what roles they can realistically enter, and what evidence would change their view.

Roberts closes on the importance of using time well because life is short, while also noting that thinking alone does not always solve the problem. Todd agrees and offers a concrete method: “get off the armchair.” Rather than trying to determine the entire career in one’s head, write down what could change the ranking among options. What information would be decisive? Then go get it. Often that means talking to people in the field or simply applying to many jobs. For Todd, empirical career choice is not a slogan. It is a practice: form hypotheses, test them, learn from contact with reality, and revise.