David Epstein: The power of being a generalist

In a kind learning environment, which is next steps and goals are clear, rules don't change, patterns. repeat, feedback is quick and accurate, people get better, reliably with narrow experience. On the other end of what he called wicked learning environments, where next steps and goals may not just be given to you, or rules may change, or patterns don't just repeat, feedback could be delayed or inaccurate, lots of human dynamics are involved, you basically work next year, may not look like work last year. And in those kinds of environments, which we're increasingly thrust into in our work because the workplace changes more rapidly than it ever has before, it's people that have this sort of broader set of skills that allow them to pivot and to create kind of generalized, flexible skills and mental models, who are really able to adapt over time. And that's the situation more and more people are in.

So I think-.

And even when people are staying in the same jobs, the stuff they're actually doing is starting to change at a much faster pace. And they're having to learn new tools and all these things if they want to kind of stay competent. And so I think it's part technological innovation, it's part globalization, but it's just that for most people, that model of period of training followed by period of working has been replaced with relearning and retraining and having to sort of reinvent over and over and over.

The reason it's called the Dark Horse Project actually is because when they brought in these subjects just for sort of to orient them to the study, the people would say, well, you know, don't tell people to do what I did, because, you know, I came through this, I started on this one thing and then I pivoted, and so I was, and dark horse means like someone who came out of its expression, it means someone who came out of nowhere. And that turned out to be the norm. It wasn't everyone, but it was the very large majority had sort of followed this path, where they started down one route and said, okay, you know, this actually doesn't quite fit me the way I thought, but I'm going to take what I learned, and here are my kind of interests and opportunities. right now, I'm going to do this one for, and maybe a year from now I'll change, because I will have learned something about myself. And they keep doing this pivoting until they get to better, what economists call match quality, which is a degree of fit between one's interests, abilities, and the work that one does.

And so not only are they kind of triangulating, a better fit in their work, which turns out to be really important for your sense of fulfillment and for your performance, but they get there with this broad kind of array of background experiences that they bring with them in this broad toolbox. And so I think, having that habit of mind where you're always kind of, what did I learn from this? What can I take to the next experience? And continually triangulating better match quality through the course of your career is a really important kind of habit of mind for someone who wants to be a so-called good generalist.

And you can marshal all of the scientific data you want, right? But if the goat stands up and says, you're an idiot, it's going to be a bad day for you, I think.

. It probably would have been worse if you had to play against her in tennis. But...

So this would have been worse embarrassment for me. Okay. But she stands up afterward and says, like, raises her hand for the first question. I'm like, oh, man. And she says, I think my father was ahead of his time.

He had me do track and field, taekwondo, gymnastics, ballet, learned to throw a football for the overhand motion for a serve, and took me and my sister off the traveling tour when he thought it was getting too intense. so we could also focus on school and family. And I was a senior writer at Sports Illustrated, and I had never heard that story, right? Even though that is what a now mountain of scientific research suggests, that this diversity early actually sets people up to have these generalized skills, what some people call physical literacy, these generalized skills that then scaffold later technical knowledge with the opposite path. The one that I thought that Serena didn't do, I didn't know that she had done all these other things.

There's actually a fundamental trade-off. This gets at what I think is kind of the underlying theme. on every page of Range. There is this trade-off between optimizing for short-term performance and long-term development. And so there are now these just tremendous number of studies in sports that show that the way you make the best 10-year-old performer is not the same way as you make the best 20-year-old performer.

There's actually an inverse relationship when, like national development pipelines are studied, between maximizing junior performance and optimizing senior performance. You want to optimize senior performance, you want this kind of sampling period where people do a variety of things. It causes them to build these generalized models and generalized skills that are then flexible and they then more rapidly learn new skills going forward. It's kind of like people who grew up with multiple languages. They actually have a little bit of a delay, but that delay doesn't last, and they then are advantaged for learning any subsequent languages.

It's actually like the same exact mechanism, really.

and it averaged 10,000 hours. But I think the issue with the 10,000-hours rule,

which—and I can tell you about after I criticized it, Malcolm Gladwell and I became running buddies, so we have these arguments on our own time, but we've come to pretty much the same ground. The idea was that this is the only thing that matters, your accumulated deliberate practice, and so you should start as early as possible in as narrowly technical practice as you possibly can. And, except for certain very simple activities, there's a mountain of evidence that suggests this just isn't right. Like, there's almost nothing that actually works in favor of it. You really want—again, if you're trying to optimize for the short term, sometimes that might make sense.

But if you want to build both physical and cognitive models that can be flexible for future problem solving, a better rule to think of than the 10,000-hours rule, which is very splashy, is this classic finding that comes out of cognitive psychology that can be summarized as breadth of training predicts breadth of transfer. So transfer is a term psychologists use to mean taking your skills and applying them to a problem that's not exactly the same as the last one, like your ability to adapt to new challenges and new tasks. And what predicts your ability to do that, which is what we, in modern work, rely on every day, what predicts your ability to do? that is the breadth of problems you've faced in training. So when you face a really broad sampling of problems in training or lots of different activities, it forces you to build these generalized models that you can then kind of slot in to whatever the new challenge is, as opposed to just being the opposite, where the 10,000-hours approach often leads to what psychologists call the Einstallung effect, which is where you've solved a problem a certain way so many times, you'll continue doing that even if it's no longer the solution.

But the data also show that they are much more likely to do those procedures when they're unwarranted and to continue to do them even if data has shown unequivocally that they aren't right, that they don't actually work, or they can even be harmful. So it's sort of this double-edged sword. It's not that we don't need those people who have such great motor memory, that they've automated the important parts of the task in their brain, essentially, but it's really a double-edged sword. So I think the challenge is in places where you have, you know, areas where you just need skills that are so thoroughly automated in the brain that someone really, you know, has perfected them, a pilot, a surgeon. How do we balance that need with the Einstein effect, basically, where the inability to see that this isn't the right solution anymore?

Others were, and the students, they make progress fast, they like it, feels good, great, rate their teacher, great, everything. The other groups were randomized to what's called interleaved practice, where instead of getting A, A, A, A, B, B, B, B, B, it's as if you took all the problem types, threw them in a hat, and drew them out at random. So now the students get more frustrated, their progress, initial progress, is slower. Sometimes they rate the teacher more poorly. But instead of learning how to execute procedures, they're learning how to match a strategy to a type of problem.

And the problem is, if you're struggling, that's not a sign that you're not learning. If you're not struggling, that's definitely a sign that you're not learning. And then, when the test comes around, those interleaved, the students who got interleaved practice blew the blocked practice group away. It was like on the order of moving the 50th percentile kid to the 80th percentile, because in the test, everyone has to face new problems. So now they have to do that transfer, where they take some of those skills and apply it to a new problem.

And they were much better equipped to do that, because again, that mixed practice forces you to kind of build these generalized models.

And obviously, it's important for kids to learn how to read. But there's no evidence that learning something that everyone is going to learn a little bit early really makes a difference. So we often... A lot of our intervention programs focus on what's called closed skills, which are things that everyone's going to learn. It's like teaching a kid how to walk earlier than others.

It might seem impressive, but it doesn't matter in the long run. So I think I would make... To make this interdisciplinarity more cohesive, I might organize a curriculum around the progression of human ideas. Because I've found in just my own learning, in any given era, art and chemistry, and biology, and technology and economics, and all these things are interlaced in a given era. And I think I would try to have maybe...

And the way that those ideas progress off one another, I think is very useful for getting a grounding in human progress and human thought. And so I might orient the curriculum around that, about whatever the discipline is in teaching people that progression of human ideas and how we got to where we are today. And that might have some linkage between the disciplines. And so it's not just like, oh, I'm done with math class, now I go to history class, which is a totally different thing. And history and math have nothing to do with one another.

Because obviously that's.

. Disciplines are a necessary evil for making the world comprehensible and giving people career trajectories. Someone has to put the world back together again at the end of the day to solve problems.

. Well, the data seems to support that boys are scholastically kind of properly ready for school later than girls.

Should boys start later in school?

And.

. But, like in Germany, for example, they did some research where they saw that kids, and especially boys, who were born later in the age cohort, so they're effectively 9,, 10,, 11, 12 months younger than kids that they're in school with, were far more likely to get diagnosed with behavior issues or ADHD and medicated and things like that, or to get put down a level if their class is tracked.

And the problem was, just as sports coaches mistake early maturation kids for potential just because they're bigger, because they're 10 months older or whatever, in school, when kids are 7,, 8, or 9, or 10, being 10 months younger is actually a big deal, and it shows up in their behavior. And so we mistake this sort of talent or potential or whatever for just biological maturation. And so I do think we should try to do something about that. I don't know how we would fully correct that. One, I think, and I think there's some countries, particularly like Finland, does a great job of making sure that when kids are young, they don't get tracked down.

So there may be a kid who's relatively younger, but they put a ton of emphasis. They don't put as much emphasis, I think, as maybe the research would suggest on tracking people up when they're ready for it, but they put a lot of emphasis on not letting people get tracked down, which then puts them on this trajectory where they're kind of stuck forever. And so I think some of that emphasis, more oriented toward not letting those kids, who are often boys, get tracked down until they can kind of catch up maturationally, would be really important. But also, we know that movement has a huge impact on kids who have trouble sitting still, and especially boys, and especially young boys. And so I think we only view some of their behaviors as maladaptive when we ask them to sit still for 10 hours in a row.

And so I think we should kind of reorient some of the school environment to what we know about developmental psychology.

And I think that's increasingly ill-suited, I guess, to the reality of the workplace. So there's this psychological finding called the end of history, illusion. It's this idea, this finding that when you ask people, hey, have you learned a lot about yourself and what you're good at and what you want to do and what you're bad at in the past? And people say, oh, yeah, of course. And then you ask, well, will you learn more in the future?

And people say, not so much. Now I'm pretty much done. And at every time point in life, people say, I learned a lot in the past, but now I think I've arrived. So we're like works in progress, constantly claiming to be finished. And this continues throughout our life.

But the fastest time of personality change, which is, again, what you think your strengths and weaknesses are, what your values are, what you like in friends, et cetera, is about 18 to the late 20s. And that's the time when we're telling people, you have to have sort of figured out. And that's not to say people shouldn't dive into stuff. You should dive into stuff because that's how you get a signal about whether you're doing the right thing or not. But I think we need to be prepared for change and kind of keep dabbling and keep doing some of that experimenting.

There's a fascinating study that came out too recently for me to get into my book. It was by researchers at Northwestern. It's by Dashan Wang, who just does this. incredible... People will do a study of 20 careers, and Dashan Wang will do a study of 20,000 careers.

His work is just on another level. And so he and his colleagues did this analysis of, I think it was 26,000 career trajectories. I think a lot of them were artists, film directors, scientists. And what they found was that there was actually this kind of universal... First of all, they found that most people do their most impactful work in their life in a so-called hot streak.

Their best work tends to kind of cluster. Now, sometimes those clusters can extend for quite a while, but there tends to be a cluster. And that doesn't mean that they aren't doing failures at the same time, but the hits tend to be sort of clustered. And sometimes people have two, sometimes people just have one. But what they found was reliably before a hot streak, people spend time in this period of experimentation.

And it really jives with this big body of literature on what's called the explore-exploit trade-off, where explore is searching for new knowledge, trying new skills, trying new projects, trying new collaborators. Exploit is taking stuff that you already know and people you already know and digging down on that and making the most value out of it. And what they found is that this explore phase, it always proceeds like a successful exploit. And when I've talked to Dashin, I think part of his concern is that in the race to kind of find a specialty, because we all specialize to one degree or another at some point or another, unless we just become complete dilettantes forever. What his concern is that people are rushing past the explore straight to the exploit, which undermines the chance that they'll ever have.

Or for artists, sometimes it was experimenting with a bunch of different styles, and then merging something from those different experiments into a new style that they then really drill down on. I think it's this sense of, as I think of it, of sort of prototyping your possible lives and careers, and taking a little bit from these little experiments and saying, all right, now it's time to kind of drill down, put more resources, often putting more of a network. In some cases, when the careers were in business or in science, actually then gathering a much larger group of people to then work on the project.

And so I think, there's not dispossessive evidence of this, this is me speculating, but I think one of the reasons that the creating so much more is kind of a hallmark of these great creators is that that's actually really the only way you learn, that introspection, I think, has been oversold as a method for learning about yourself, and doing stuff has been undersold. Doing stuff and then reflecting.

That said, I think that people are inherently more curious if they don't think it's disincentivized, right? Like, I would even talk to some of the researchers I was interviewing for my book, who were disappointed at their careers in academia because they felt that they had gotten into this world where it would be the life of the mind, and yet then they were told... Like, I remember interviewing this one researcher who she said she felt frustrated, that she felt like she was maybe being held back in her career because she wasn't more specialized, because she wanted to explore more stuff. And I went and looked at her list of papers. Every single one, it was like 48 papers or something, every single one had Aristotle in the title.

Like, from my perspective, she was very specialized. Every single one.

And so I think people want to be curious, but they often feel that it's disincentivized. And then sometimes people get to a level where they're sort of executives, or in Silicon Valley, and then they get to be free and really curious, and I think you see some of the kind of flourishing when people feel like that's incentivized. But I think specific curiosity is important and underrated, and that we need to encourage people sometimes to dive into problems that interest them or that they think are important, that maybe are not exactly in their view every day. There's a guy named Ed Hoffman, who was a background source for me, who was the first chief knowledge officer at NASA. That's basically when NASA had some accidents, they created this position.

that was for a psychologist to make sure that they had institutional memory, make sure that they were learning from their mistakes, and that those learnings were spreading across the organization. And now he consults with a lot of organizations, and one of the things he does is he goes in and he asks people, when he's just sort of first doing his listening tour, what's something you're good at that we're not using? And what he finds is that people will often bring up some problem that's in their office, that they basically know how to fix, but they don't feel incentivized or even allowed to do it. So I think there's like a lot of innate curiosity that if leaders or managers are willing to underwrite the risk of those curious people in taking some risk, that would be basically unleashed. Whether or not we can just create a lot more innate curiosity for people, I'm not sure, but I think there's a lot that is suppressed that we could just kind of liberate.

So it's this, they're almost trying to valorize this idea that, hey, we took on something that was worth taking on, whether it worked or not. Like we had to figure this out. We had some lessons from it. And so you see these examples of people who then recover and go into other projects.

Census Bureau in Northwestern that found that the average age of a founder of a fast growing tech startup on the day of founding was 45.

. Right. And we never usually...

. But we usually don't, right? We usually don't pay attention. We, I think we put much more attention to the stories of like Mark Zuckerberg, who famously dropped out of Harvard and he famously said, young people are just smarter. He was 22 when he said that, right?

You don't hear him saying that anymore, surprise, surprise. And so I think it's sort of counterintuitive.

And really a lot of this great innovation in people's career moves comes from being opportunistic and seeing, learning something about an area, seeing an opportunity and then pivoting to take advantage of it, as opposed to being part of some like perfect long-term plan. I've yet to meet an entrepreneur who said like, yeah, things went as planned.

How do you think about range in an organization?

And what predicts the number and breadth of analogies is the diversity of the people in the group. And so in many cases, when you have everyone with a similar specialty, it's not that much better than having just one brain with that specialty. And some of these findings can even be quite like, almost absurd. So, based on this research, you know, there was research looking at teams, okay, if you throw in someone with a different specialty, does that help problem solve? Okay, yes, it does.

What if you throw in someone from a different area who's like not really that competent? Still helps some. And then I said, well, let's take it to the farthest extension. We'll look at remote teams working on problems and we'll toss in AI bots programmed to behave randomly. You know, they don't know our bots because this is just a remote team.

And even that improved problem solving, because it apparently gets people to de-anchor, right? Because usually the first solution that comes to mind is the one that we sort of get stuck on or anchored, the first or second. And these random behavior would actually kind of knock people off of those anchored solutions, because it turns out we usually anchor on a first solution and it's usually not our best one. So there's this thing called the creative cliff illusion. that's like, we think good ideas come either quickly or not at all, when in fact actually they tend to come later in your thought process.

So I think mixing cross-functional teams and I think saving some room for people that seem a little different. Like some of the, I think sort of most kind of progressive companies I've been around, in the sense that they disrupt themselves. instead of waiting to get disrupted, we'll put some effort at least into hiring. Sometimes they just need to put a square peg in a square hole today, but other times they'll save some efforts for like sort of looking around and saying, okay, here are things we're really good at that we can teach people how to do.

There's some other things we want that we're not good at or that we couldn't show people how to do. So let's go, look for people who have those things, bring them in, and then we'll teach them the stuff that we're good at teaching, right? So the, I think the farthest extreme of this that I encountered was Bailey Gifford, which I think is like the most successful investment company in Scotland, where they would not hire anyone with an MBA. I think that's a little overboard, but maybe part of their culture. But they would say, let's just go, look for things that we want in here and then we can teach them the finance stuff.

And so I think that's kind of extreme, but it comports with, like the work of Abby Griffin, who studies so-called serial innovators, these people who make repeated creative contributions to their organizations. She finds these people are, you know, just to almost directly quote from her work, they have a wide range of interests. They read more and more widely than their peers. They have a need to communicate with people with expertise outside of their own. They connect information from different domains.

They repurpose, like old things that are already available. They're systems thinkers. All this stuff.

I talked to Adam Grant, I think our mutual friend, about this a little bit. And I think he said, well, can you create these serial innovators? And I think the answer is not sure. My intuition is probably not that you can't just create them. But I think absolutely you can stifle them from becoming if you don't allow that sort of broad networking and broad thinking.

And one for scientific research called Site.

ai. And so everyone should turn on a little bit of their technological anthropologist brain and be using these tools to figure out where do they go right and where do they go wrong. Because my experience has been they're very often wrong and yet very useful. So it's interesting to get used to a tool that's very often wrong but also useful. But in the larger scheme of things, I think whether technological innovation of this magnitude leads to shared prosperity or increasing misery has a lot to do with whether people have opportunities to, and are able to adapt to, sort of new roles that move them to a bigger picture scale.

So, to give a sense of what I mean, like when I was looking back, I was kind of doing some research on the introduction of the ATM, the automatic, like a cash machine in the US. And it first came online in the early 70s. And the news coverage at the time was really apocalyptic. It's like there were 300,000 bank tellers at the time said they're going to go out of business overnight. And instead, over the next 40 years, instead of, as there were, more ATMs, there were more bank tellers because they made each branch cheaper to operate.

So banks opened more branches. So fewer tellers per branch, but more tellers overall. But, even more interestingly, it fundamentally changed the job of one from someone who's doing repetitive cash transactions to someone who's like a marketing professional or a customer service representative or a financial advisor, this much broader mix of more strategic skills. And I think that's kind of emblematic of when technological disruption sort of goes well for society at large, is it moves people back to this more strategic level of thinking and takes over stuff that, you know, it might be scary that it's taking it over, but it's like less of the place where we kind of add value. And so I think people need to be experimenting with these tools and starting to think about how we can liberate them to spend more time on the strategic side, the bigger, the bird's eye view picture of their work.

So you'll have this real echo chamber. You should be constantly pruning and adding people who are more different from you to diversify the sources of input. And I think you should be doing purposeful experimentation. So I keep something I call a book of small experiments, where every, you know, borrow from the scientific method, say, here's something I want to learn or here's something I want to learn about. Here's my hypothesis for how I'm going to do it.

Here's how I'm going to evaluate the outcome and do that over and over and over and over. So, you know, to give a concrete example of something that's relevant to my life, like I felt like I was stuck in a rut with writing. I said, I need to learn some new ways to structure writing, because structuring information is the major challenge for me. So I took a beginner's, I said, well, where's there more structure? Okay, fiction.

So I took a beginner's course in fiction writing and it was a total revelation for me. Like there was one exercise where we had to write stories with only dialogue and with no dialogue, and the no dialogue one was so much better and I realized I had been overusing sort of dialogue. I went back and changed it. You know, like basically every page of range. And so I'm constantly doing these experiments where it's make a hypothesis, find a way to test it, reflect back on what you learned and take that forward.

So I think more purposeful experimentation mindset is a really crucial one, for, you know, and may come naturally to some people who are just like incredibly curious, but not to most people. But I think it's really crucial in a world where, like you, should not be feeling entitled to stability of the thing that you're doing for years on end.