Malcolm Gladwell on Why We Shouldn’t Value Speed Over Power

“Let’s create a safe space for the neurotic tortoise as opposed to penalizing him at the point of entry.”

Malcolm Gladwell, known for his deep inquiries into how the social sciences impact our day-to-day lives, recently sat down for a talk with Adam Grant, a Wharton psychology professor whose latest book, Originals, deals with the character traits that foster creative success. In this conversation, at the Wharton School People Analytics Conference, Malcolm and Adam debated whether our ways of evaluating success are biased towards speed and discussed how colleges and workplaces might make space for growth that takes a little longer.

This conversation has been edited and condensed. To view the full conversation, click the video below. 

Adam: 10,000 hours might be your most widely discussed bit of writing—arguably, the most misunderstood, as well. What would you actually like us to conclude about expertise and deliberate practice?

Malcolm: I was interested in the idea that if it takes you a long time to master something, then that must mean that you need a lot of help, and that you must be in a situation that’s patient.

Want more conversations with the world's great thinkers? Click the shiny blue button!
That’s what interested me—the context. If we’re all naturals, then the context in which we perform what we do is irrelevant. If you’re born being able to be a scratch golfer, then why do we need to spend money developing young golfers? But actually, not only does it take a long time to get good, even if you’re really talented to begin with, it takes an incredibly long time.

Not even Roger Federer could be a great tennis player without a coach, without a place to go and play tennis, without parents who drive him there. Roger Federer, for years, was known for having a terrible temper. At the beginning of his career he was thought to be someone who would never amount to true greatness because he didn’t have the requisite personality. He would have meltdowns, throw his racket, storm off the court. They were like, “Ugh, another one of these people who’s going to squander his talent.”

But that’s just because we were observing him in the middle of his necessary period of tennis apprenticeship. Once he had completed that, he turned into the tennis pro we know, someone whose control of his emotions is perhaps as good as anyone who’s played the game. So even Roger Federer required a patient ecosystem to become truly great. That’s all I was trying to get at. 10,000 hours? It’s a number that has been thrown around by people who were looking at musicians, which I thought was intriguing. But it was never meant to be definitive, nor was it meant to be a statement that talent didn’t matter. I could spend 10,000 hours at any number of things and I will never be any more than mediocre.

Adam: Yeah, and the role that luck and opportunity played in making that possible is a huge part of the story.

[You’ve said] that we should be thinking about slowing down standardized tests like the LSAT, the MCAT, or the GRE. The reason that I want the standardized tests to be short is because we already have long power tests—they’re called grades. If a student spends four years accumulating those grades, why in the world do we need yet another tortoise contest?

Malcolm: LSATs, SATs, and GREs are not knowledge tests, they’re cognitive evaluations. We’re trying to get at something that is different from grades. As a pure cognitive evaluation, why are we biasing in favor of the hare? I’m interested in measuring the pure cognitive strength that shows its face under pure power conditions, particularly for pure power professions.

Adam: I like that. However, in psychology, we have this distinction between typical and maximum performance. Typical performance is how well do you do on your usual day; maximum is how good are you at your best. Those two measures tend to correlate, so that the higher your maximum performance, the higher your typical performance.

“Although there might be some really brilliant tortoises and some less brilliant hares, most of the time [speed and performance] go hand-in-hand, and processing speed is also a proxy for the complexity of information you can handle.”

Which leads me to think that although there might be some really brilliant tortoises and some less brilliant hares, most of the time the two go hand-in-hand, and processing speed is also a proxy for the complexity of information you can handle. If that’s the case, do we really need to tease the two apart?

Malcolm: My understanding is when you move from a speed and power test to a pure power test, you don’t necessarily change the shape of the curve, but you do jumble the rankings of people on the curve. It’s that jumbling that interests me.

Look at, for example, chess. If you compare classic versus blitz chess rankings for the top 20 players in the world, there isn’t a huge difference. But of those 20 players, there’s maybe four who have dramatically different rankings. There are weird cases like Magnus Carlsen, the greatest player in the world, who is marginally the best at classic chess. The more you speed it up, the more he becomes dominant.

There’s two things going on here. One is that when I arbitrarily add in the speed component, I start to lose at the margins. I may have a general sense of what’s going on, but I’m missing people.

The second thing is, it obscures my understanding of what makes someone good. You learn a lot about Magnus Carlsen when you look at his performance under different speeds and conditions. You understand what is brilliant about him as a chess player is that he doesn’t make mistakes, even when the game is going [very quickly].

Adam: I can see the rationale for that. It just seems like in most complex jobs, it’s not quite as independent. In general, the more expert somebody is at a job, the less you have to rely on slow, system two thinking, the more that your fast, intuitive, visceral heuristics are accurate. Shouldn’t you just assume that the experts are going to be the fastest?

Malcolm: Hold on. I chose a legal professional for a reason. That is one example where that relationship between speed and performance starts to break down. I absolutely do not want a speed-reading lawyer.

We have to understand that the cognitive profile of the profession is different. My father was a mathematician. There’s no upside for him being fast. A great mathematician might publish 10 great papers in their lifetime. Why would we want to reward a mathematician who wrote his paper in six months as opposed to two years?

Adam: The more output you produce, the better your shot at stumbling onto greatness. So won’t we want to reward speed to get to quality?

Malcolm: Not in a lawyer. I’m being very specific about lawyers. We do not want the high output, lots-of-errors lawyer.

Can you imagine a lawyer who said, “Here’s the contract, take a look. If it turns out it’s not right we can just go back and do another version later.” Are you kidding? That’s a disaster. In the financial crisis, someone put the comma in the wrong place and ended up paying $20 a share for Lehman and not $2 a share. Who was the person who read that document at two in the morning? The hare.

In any kind of high-stakes job where the penalty for error is high, you can’t afford to have hares. What I’m objecting to is trying to make a general set of principles about selection—I think you can’t. You have to be much more specific.

“We can’t argue for the perfect form, because the perfect form happens once in a generation.”

There are parts of the law where I might want a hare. I want a legal professional to say, “For this kind of law, I want the neurotic tortoise. For this kind of law, I want the hare.” Let’s create a safe space for the neurotic tortoise as opposed to penalizing him at the point of entry.

Adam: [But couldn’t] there be a really conscientious hare, who’s fast, who executes, and is also careful on the back end?

Malcolm: That’s like saying, can’t we have all basketball players who resemble Michael Jordan? We can’t argue for the perfect form, because the perfect form happens once in a generation. If you want highly conscientious, highly neurotic people, they’re going to be tortoises, by and large.

Adam: Let’s say we could profile every job, then end up in a situation where we’re really good at selecting for individual job performance and terrible at selecting for the qualities that would make for a high performing culture. What extra layers do you want to add, as we think about not just optimizing my own individual contributions, but also the sum of the parts?

Malcolm: Why isn’t it good enough to say that if I have an environment that allows individuals to maximize their potential, that will ultimately be for the best? Let’s take an example of a university faculty. As a writer, my principal observation about why other writers fail is that they are in too much of a hurry. I don’t think you can write a good book in two years. You may disagree, you have done that, but you’re an anomaly. Most of us can’t write books that quickly, and we need to be a little bit more tortoise-y and a little less hare-ish.

The problem is that the world wants you to be a hare. Your publisher says I want it now, you’re under pressure, you have a one-year sabbatical where you try to cram and finish, you’ve got a teaching load, etc.

But one thing that almost all of the professional writers I know do is write drafts and then put the book in a drawer for six months. Then they come back to it, turn themselves into tortoises, force themselves to slow down. That, in a sense, harms the system in that amount of output is lowered. But I don’t think the problem with writing in America right now is a failure of output. I think it’s a failure of quality.

So there’s a case where the overall system could use a lower level of production and some higher production values, and having individual writers who write better books makes us all better. Is the New Yorker a better organization if writers write fewer articles in a year, but those articles are very memorable? If you did a systematic analysis of the financial health of the New Yorker you would learn that the New Yorker is a hit-driven enterprise, that probably eight articles a year account for 90% of people’s interest in the product, and so to the extent you could encourage people to write fewer hits, you’re better off.

Adam: The academic cheat on this is by the time you write a book you’ve already been working on the topic for five or 10 years, and so a lot of the tortoise work happened up front.

You were harsh on standardized tests because of what they’re missing, key skills that might be relevant for a job. What else is being rewarded that we shouldn’t be measuring? Like, can you afford to take an LSAT prep class? And how do we get that out of the system?

Malcolm: A really good way to get it out of the system would be to dump standardized tests entirely. A lot of other countries don’t have standardized tests. It’s not a given that human beings need to conduct their entry to elite institutions this way. I don’t understand why people are so obsessed with them given the fact that their actual predictive usefulness is small.

“If we were starting the American education system from scratch tomorrow, would we have the SAT? Of course we wouldn’t. So why are we persisting in the charade?”

You’ve got a system where people are hiring coaches at enormous costs in order to improve their score on a test that doesn’t really matter. We are at a level of absurdity with this particular game, why not just call a halt to it? I’m sick of trying to fix the system. I think it’s time just to dump the system. If we were starting the American education system from scratch tomorrow, would we have the SAT? Of course we wouldn’t. So why are we persisting in the charade?

Adam: I have a hypothesis: we’re in this charade because it creates an illusion of certainty. It allows those of us who make selection decisions to believe that there are more deserving and less deserving candidates, which I think is largely an illusion. How do you tackle the more fundamental problem of people having to admit that what they do in sorting and selecting applicants is basically throwing darts, that there’s basically a lottery running there?

Malcolm: Why not just run the lottery? Say I’m Penn. In order to apply for Penn, you must be in the top 10% of your class and do one interesting thing on the side. Then we’re going to throw all those names in a hat and pull them out. I can tell you with 100% certainty the freshman class at Penn under those circumstances would be infinitely more interesting than it is now.

Adam: What that surfaces for me is there’s still some arbitrariness in where you draw the cut-off, so you’re never going to end up at a complete lottery. If you were going to design your own selection system, what would do you think is less arbitrary than the alternatives on the table today?

Malcolm: I’ve thought about this, actually. The conversation is too one-sided. If you read the literature on what makes for a meaningful college experience, almost all of the literature stresses the way the student interacts with their institution: when I show up on campus on day one, how do I behave? Do I seek out the most interesting professors and take their classes? Do I willingly throw myself into the experience or do I smoke dope in my room? The variable is you, not the institution.

If I’m an institution, what I’m really interested in ought to be, what does the individual want from me? Instead of writing an essay that talks about what happened in your own life, and the institution says “I like that,” flip it. It should be, “What do I want from my college education?” If you’re MIT or Harvard, someone who is applying to your school should be able to say with a certain degree of specificity which professor they would like to study with and why.

If you’re not at that stage of intellectual development, then don’t go to those schools, because that’s what those schools are for. If you want to join a frat and party, you should say, “My principal interest is joining a frat and partying. As a result I would like to go to Duke.”

Institutions ought to have clear personalities and ought to recruit people who are interested in that. If I’m Carnegie Mellon and have the greatest robotics faculty in the world, what I want to know is, if you’re applying to Carnegie Mellon, have you read stuff by the professors at Carnegie Mellon? If so, what did you think? That’s relevant. What are you going to do when you get here is the question I’m interested in.

Adam: Do you want companies to do the same?

Malcolm: I do. I think they do a better job than colleges at that. This is one of many areas where education can learn from the private sector.

“Spend six months in an institution, and if you and the institution say it’s not working, that shouldn’t be a black mark on your resume.”

Adam: If I’m an employer and I want to make sure that I’m not privileging speed over power, where would I start?

Malcolm: Apprenticeships or trial periods are a version of power over speed. I’m removing the time constraint, saying, “If you would like to come, why don’t you just come, and why don’t we see how things go over a period of weeks or months.”

Spend six months in an institution, and if you and the institution say it’s not working, that shouldn’t be a black mark on your resume. That should actually be positive. It says that you’re brave enough to experiment, to go out in the world and try stuff.

Adam: You’ve written recently about how engineers think. We’re trying to make organizations more evidence based, more data-driven, and that’s what engineers do for a living. Is there anything we can learn from how engineers think as we think about making HR and the world of people more data-driven?

Malcolm: I don’t know whether you want to make the world resemble engineering culture. I think you want to find better ways for these two very different cultures to speak to each other.

We absolutely need engineers to think like engineers, but we absolutely don’t want everyone to think that way. Nor do we want non-engineers to shut down the engineers—we want to have both at the table. I’m more worried about the hiring process becoming too dependent on analytics than I am about it not being dependent on analytics enough.

I wish there was a little more humility about what can and can’t be measured. I follow this most closely in sports. You can’t follow the analytics revolution in, say, basketball, and not be simultaneously thrilled at what we can know and deeply humbled about what we can’t know. There were two European players playing for the Denver Nuggets earlier this year. Neither were playing very well, and the consensus was that one or both of them was going to wash up.

Denver traded one of them. It is now the case that both are playing unbelievably well. If you can find any analytic that helped you predict that outcome, be my guest. It was an intangible. They weren’t happy together, and apart they’re fantastic. That just tells you that there’s an awful lot that we can’t easily understand about human performance.