Search the Site

Episode Transcript

Steven LEVITT: When Sal Khan started posting a few free instructional videos on YouTube 15 years ago, I don’t think anyone could have imagined what Khan Academy would become. More than 115 million registered users spread across 190 countries in 46 different languages. And it’s gone way beyond videos. Now there are over 70,000 practice problems, quizzes, and articles, even state-of-the-art S.A.T. prep. All of it available completely free to students and parents.

Welcome to People I (Mostly) Admire with Steve Levitt.

LEVITT: I spend much of my time these days working through my center at the University of Chicago to try to change the world for the better, and I can’t tell you how much we’ve struggled and how little impact we’ve had relative to what I would have hoped for. Sal Khan makes it look easy, but in reality, it is incredibly hard. 

So today, I have two goals. First, I want to get inside the head of Sal Khan to understand what the special sauce is that has made him so incredibly successful in this area so I can do better myself. Second, although he doesn’t advertise it, I have the suspicion that Sal Khan’s ambitions go far beyond Khan Academy. 

I think he has an even bolder plan in mind, a complete rethinking of how education works. I also have radical ideas for the future of education, and my pipe dream is that our visions might align and there could be an opportunity for us to work together. So let’s see. 

*      *      *

Steven LEVITT: Sal Khan, what a treat it is to get to talk to you today

Sal KHAN: Likewise, I’ve always been a big fan, Steve

LEVITT: I’m a huge admirer but I have to say, I think you set a very dangerous example for other people. You had a good job you liked at a hedge fund that paid you a ton of money, and then you had some crazy vision that you would post short lectures on the internet and you’d give them away for free hoping to provide kids another way to learn. That sounds like a well-intentioned plan. It also sounds like the kind of scheme that would fail 999 times out of 1,000. Wasn’t it crazy to think that Khan Academy could actually work?

KHAN: Yeah, on a lot of levels. Well, before I was in the hedge fund world, I kept being drawn to either tutoring directly, like me being people’s tutors, or how can we leverage technology to help people learn ideally at their own time and pace? And so that was always in the back of my mind. Even while I was working at the hedge fund, I would tell my friends, “I like this job, it’s interesting, but I’m only going to do it until I can be independently wealthy and then I’ll start a school on my own terms.”

And then when my one cousin in particular, Nadia, in 2004 needed help, I wanted to help her because she was being put into a slower math track in seventh grade, which has all sorts of implications for her own self-esteem and just her life. So it started in a very organic, family-project type of way. And that same Nadia, who was being put into remedial math class was then accelerated into an advanced math class after I lobbied on her behalf as a bit of a tiger cousin.

Before I knew it, word had spread in my family that free tutoring was going on. So there was about 10, 15 cousins, family friends I started working with. With the background in software, I was always fascinated by, “let me see if I can make some tools to streamline what we’re doing.”

I called it Khan Academy, just as a joke. It was me with my cousins. Although, in the back of my mind, I said, “The cool thing about software, if it works for 15 cousins, it could work for a billion people, eventually.” So I had some delusional aspirations at the time.

And by 2006, a friend suggested that I make videos to supplement my lessons, which I was having trouble doing. And I, initially, thought it was a bad idea. I said, “YouTube’s frivolous. It’s for dogs on skateboards, cats playing piano. It’s not for serious learning.”

LEVITT: Wait, why did you think the videos would be a bad idea? Because you thought that the back and forth was really important? Or just because you thought YouTube was frivolous?

KHAN: It was both. I think technologists often fall into this trap, and I, with a tech background was falling into it. I was like, “Look, I’m writing all this cool software that can adapt to students in giving them as many exercises as they need and giving them feedback and I get all this data and analytics. Videos feel so low-tech. If this was a good idea, people would have done this with VCRs in the ’80s.”

And then there was another notion which was at the time, YouTube did feel the way TikTok feels now. It’s just for these silly little things. But then I gave it a shot and when I asked for feedback from my family, they famously told me they liked me better on YouTube than in-person.

My self-esteem likes to believe that they still appreciated having me on the phone accessible, but they also really appreciated having an on demand, infinitely patient version of their cousin available at three in the morning that they didn’t have to feel embarrassed about if they need to cover something from four grade levels before.

For my cousins, many of them, it was transformational. They went from being students struggling to the top students in their school, or even in some cases in the city. The YouTube videos, in particular, became very discoverable and accessible by people who weren’t my cousins. I started getting letters from people all over the world saying how it was helping them, how it was helping their children.

I’d, frankly, had trouble focusing on my day job. And that’s when I took the plunge. I set it up as a not-for-profit with a mission of free world-class education for anyone anywhere. And a lot of my friends are V.C’s. I live in Silicon Valley at this point and they were saying, “Hey, we’ll write a check. You can quit your job and we can do a social good for-profit.”

And that sounded good, but it always led to these conversations of, “Well, this is how we can monetize the users.” And when I read these letters from people around the world, I was like, “I don’t know if I want to monetize them. And I don’t know if I want to add any frictions to someone who wants to learn.” So that’s why I took a little bit of the bet and set it up as a nonprofit.

LEVITT: Let’s say you had gone that V.C. direction. What do you think would have happened?

KHAN: I think it would have been a reasonably successful for-profit — maybe super successful, I don’t know. But I don’t think we would have captured people’s imaginations in the same way. I don’t think a for-profit could, with a straight face, have a mission like free world-class education for anyone anywhere. Which is a big mission, but when people started to realize that we actually have a shot of doing it, became a very aspirational mission.

Some of our early funders have been people who come from essentially my two backgrounds, far more successful people than I ever were, people out of tech or people out of finance. They just liked the notions of what this was. This is about scale. This is about personalization. This is about using the internet to solve a core problem that they care about.

LEVITT: From the beginning you made what seems like a really odd choice, which was you never put your own face in the videos. It was always just a blackboard, which I’m sure any kind of media advisor would have told you that’s a terrible idea, but I think it’s worked out for the best. Don’t you think?

KHAN: Honestly, when I did it initially, it wasn’t out of any super deep, strategic-vision thing. We forget back in 2006 cell phones didn’t have particularly good cameras at the time. And with my family we were just figuring out ways to see each other’s writing. I was using Yahoo Doodle while they heard my voice over the phone. And it felt very intimate. It felt like we were sitting next to each other at the kitchen table, looking at the piece of paper together.

In hindsight, I realized that it was probably a good move. There’s actually a lot of research that the human brain immediately fixates on human faces. If the human face is competing with math, the human face is going to win. And so when you focus on the math, or the science, or other subjects, and you’re able to use the visuals well, but you’re hearing the voice, then it’s easier to focus.

There’s other random things that happened that ended up being really good from a neuroscience point of view. YouTube in 2006 had a 10-minute limit. And so we started making videos that are seven, eight, nine-minutes long. And there’s research: that’s about how long people can pay attention before they need a break.

LEVITT: Now, one thing I find interesting about your approach is that it was an educational intervention that circumvented schools completely. I’ve tried to do all sorts of educational things, almost always in partnership with schools, and I’ve failed every time. I see your circumvention as being absolutely critical to your successes. Do you agree?

KHAN: To some degree. In those early days, I don’t think any major system or state board of education would have taken me particularly seriously. In all fairness, back in 2007, there was a teacher who got excited about these tools I was making for my cousins. So that actually did help me have some self-confidence on this project.

But to your point, most of the people in those early days were not in classroom settings. And them seeing the value of it and then the momentum that you got around that, and then frankly, many students started telling their teachers about it. And then in many cases, many teachers were using it for themselves — they had to deliver a lesson that they might not be familiar with. So it allowed the discoverability to be very grassroots and very organic. And in a lot of ways, it validated us.

So then when we did start to reach out to districts in a more formal way, they were like, “Oh, yeah. We’ve heard about this. We use it, ourselves.” And as early as 2011, we did start having some formal partnerships with districts. And today, about 50 percent of Khan Academy usage is just what we would call independent learners. And about 50 percent is what we would call teacher directed, which is teachers telling students, “Use Khan Academy.”

LEVITT: And you could have taken a different path, which is, “Oh, I’m going to go and I’m going to sit down with The San Francisco School District and I’m going to work with the administrators and try to come up with ways to make a change.” And I think there’s a real elegance and beauty in how you circumvented the system.

KHAN: Sometimes inertia and tradition and bureaucracy can slow things down, and just trying to convince people oftentimes will take all of the energy. Or you get to a consensus and then some of the innovation goes away. A similar thing happened when Covid hit. Khan Academy usage went through the roof. But there’s always been a dream of — we have to provide people synchronous learning at scale, with live human beings. And I’ve always dreamt about, “Couldn’t we use a volunteer model or a peer-to-peer model?”

Even within Khan Academy, it was hard initially. And so we started another not-for-profit called to do peer-to-peer tutoring. And so I’m always a big believer in, O.K., start a small and scrappy project that can experiment and move nimble and move fast and can move without permission, so to speak.

And then as it gets scale and it gets traction, you can start to integrate it more deeply with the more established players. Especially, if you’re talking about the school system to be able to have that proof of concept already working. And then once it’s working, it’s much easier to get traction.

LEVITT: I saw you give a TED talk and you used an analogy of education to home building that I found so powerful. Could you tell that story?

KHAN: Yeah. In a traditional academic system, students get exposed to some concept through a lecture, usually. Then they do some homework, that lecture/homework goes on for about two, three weeks and they take a test. And on that test, you get a 90 percent; I get an 80 percent. Even though we identify gaps, the class moves on to the next subject, probably the subject that’s going to build on those gaps. At some point those gaps become debilitating.

And we’re all used to this because that’s the system we grew up in. But when you apply that same thinking to other domains, it seems somewhat absurd. For example, home building. So if I did the same process with home building, I would tell the contractor, “Hey, we have two weeks to build the foundation, do what you can.”

The contractor shows up, maybe it rains, maybe some of the supplies don’t show up. They do what they can, and after two weeks you bring the inspector in. And the inspector says, “That part’s not quite up to code. The concrete’s still wet over there. I’ll give it an 80 percent.”

And you say, “Great, that’s a C. Let’s build the first floor. Contractor, you have three weeks. Do what you can.” Once again, they do what they can. Inspector shows up in two or three weeks, says, “All right, I’ll give it a 70 percent.” “Great, C-minus, D-plus, whatever, let’s build the second floor.”

And then all of a sudden, you’re building the third floor and the whole structure collapses. And if you have the reaction that many people have in education, you’d say, “Maybe it’s because we had a bad contractor,” or “Maybe we need more inspection.” But what’s clear in this thought experiment, it’s neither of those.

You could have the best contractor on the planet and you could have perfect inspection, but when you inspect and then you ignore the deficiencies and then you build on top of it, you’re doomed to have the structure collapse. And that’s exactly what’s happening in education.

LEVITT: And what you’re describing is referred to as mastery learning in the classroom.

KHAN: This is the reason why so many students struggle, especially in topics like math, is they accumulate these gaps. Because you’re sitting in sixth grade, the class is talking about basic exponents. Even that A student who got the 95 percent, that five percent could have been a careless error, or it could have been a really important gap in their intuition or their understanding. And those gaps keep accumulating over time. And that’s why you see so many kids, especially in classes like algebra or calculus, they start hitting walls.

LEVITT: How long do people spend watching Khan Academy videos?

KHAN: On an average month we see about half-a-billion learning minutes per day on Khan Academy. So we see about 12-billion learning minutes per year.

LEVITT: And how big is the budget for it? And how many people do you have employed?

KHAN: Globally, we have over 200 folks, about 170 of them are based in the U.S. And the annual budget’s about $60 million a year. Every time I say that it gives me a little bit of a cortisol spike because I have to fundraise a lot of that.

But as part of my fundraising, I always remind folks that’s the budget of a large high school in the United States. And our reach is literally trying to educate the world. I never want to make Khan Academy bigger than what it is, but I think now especially in the U.S. people view it as a pretty essential part of the education system.

LEVITT: So I just tried to do the math in my head and maybe I blew it, but going from 12 billion minutes of individual learning per year — and I tried to divide it by your budget. And I think I came up with the cost is about 33 cents per student per hour. And if that’s true, then it just gives you an idea of how incredibly cheap it is relative to traditional education models.

KHAN: That number, that 33 cents per hour — and I’m pretty sure you got the math right — that’s an hour of actual learning. That’s not like an hour of sitting in a class, looking at the clock, passing a note. That is an hour of, “I am engaged on an exercise, answering questions, getting feedback.” 

I never want to say that this is a complete substitute; it’s an “and.” There’s a lot of people around the world where this raises the floor dramatically for them because they don’t have access to courses or a proper school, or there’s not someone around them who can teach the subjects. And then there’s a lot of people, and this is the majority of the U.S., where they might go to a decent school, but they just need help.

LEVITT: I want to challenge you on something. So you just said, “Khan Academy for most students is an ‘and,’” right? And as a parent of children who use Khan Academy, that’s what I thought of it as. But as I’ve thought harder about it and read more about you, I think you’re much more subversive and much more revolutionary than that. Isn’t it true that your actual goal is to completely overturn education, to transform it so it really will be unrecognizable? You hate the current system, don’t you?

KHAN: I don’t hate the current system. I think the current education system, it’s become very fashionable to beat up on it, especially from reformers like myself. But for the most part, it is actually quite powerful, especially when you compare to what existed before the current system, before mass public education.

And there’s definitely moments in my schooling where I’m like, “Wow, this is a little bit boring or I wish it was done this way,” but there’s a lot of other moments in my education where I felt that, “Hey, this is great.” And especially in hindsight, I was like, “Wow, that was an incredible teacher I had that was really able to engage.” So a lot of what I think about is: how do we optimize those good parts?

And to me, that’s where technology can be an interesting angle. If teachers can have more of the students be able to fill in their gaps at their own time and pace, then they don’t have to slow down and focus on remediation as much. More students will be able to be at their learning edge when they show up to classroom.

LEVITT: I still think you’re not being a hundred percent truthful because it strikes me that the classroom model we use makes zero sense. It was derived in a time when technology wasn’t available and in a world of technology, it seems like the right thing to do is to revolutionize the classroom. And doesn’t that require a total rethinking of the technology of the classroom, of the production function we use for learning?

KHAN: Oh yeah, absolutely. I believe so strongly in what you were just referring to that I started a lab school in 2014 to do exactly this, to re-imagine every aspect of the school. Teachers don’t lecture at Khan Lab School. They sit alongside students, have Socratic dialogue with them, run simulations, teach the students to then teach themselves, or to teach each other and have a lot more experiential learning. Students learn at their own time and pace. School is not the only place where learning happens, it’s the base, but then they’re going out into the world and doing interesting things.

The reason why I started Khan Lab School, to your point, is I wanted to show what is possible when tools like Khan Academy, when tools like exist and are accessible. It gives you the opportunity of re-imagining the school, but it reimagined it in a way that every teacher who’s either worked at Khan Lab School or who has visited Khan Lab School says, “This is my dream school to teach at.” And the students like it as well. So that’s my aspiration, is that the whole world looks like Khan Lab School in 10 or 20 years.

You’re listening to People I (Mostly) Admire with Steve Levitt and his conversation with Sal Khan. After this short break, they’ll return to talk more about the Khan Lab School.

*      *      *

LEVITT: Hey, so two things caught my attention in the discussion so far. First, it is incredible how cost-effective Khan Academy is as a teaching device. In the U.S.A., the government spends about $15,000 per student per year on K-12 public education. Kids spend about 180 days in school and the school day is roughly 6.5 hours long. So doing the math, that equates to $13 per student per hour. At Khan Academy, in contrast, it costs about 30 cents per hour. That’s 40 times cheaper. The opportunities, if we could rethink our teaching model, are just enormous.

The second thing I found really fascinating about our interview is how reluctant Sal Khan was to say that he had a radical vision for overhauling education. Did you notice how I just kept on asking him the same question over and over in different ways before he finally opened up to it? 

I’m so curious to learn more about the Khan Lab School idea he referred to. Could it be the basis for rethinking our educational system? And I also want to dive into the specifics of why he’s had so much success and what people like you and me can steal from him in that regard. Or maybe “borrow” is a better word than “steal.”

LEVITT: Can you tell me more about Khan Lab School? How’s the evidence coming in of how successful it is?

KHAN: We started the school in 2014. By that point, Khan Academy had become reasonably well-known. I had written a book called The One World Schoolhouse where the first third of the book was a bit of the history of education. How did we get to where we are? The middle third of the book was my personal narrative. How did I fall into all of this that we’ve just talked about? And then the last third of the book was: let’s imagine a world where we can think from first principles — and that’s, mastery learning, personalization, the future of credentials, what are the subjects that we really do need to learn versus the ones that we don’t necessarily have to learn as much anymore? What’s the future of higher education?

And then, I’ll admit selfishly, my oldest was four-and-a-half-years-old, and he was going to go into kindergarten. And I had already been making the rounds and giving TED talks about mastery learning and personalization and I said it would be hypocritical for me not having my own child immersed in this type of a school.

And so I started talking to frankly, other employees at Khan Academy who felt the same. And I put a slide deck together, and I convinced families of 30 kids, ages five through 10-years-old, to go into this experiment. We were able to get a space from Google in an office building that we were in to start the school, reclaim a little bit of the parking lot for a playground, and we got started.

This isn’t a school where we had an I.Q. test to join. And we had kids who started off the school who were in the bottom quartile on standardized test scores, all of those kids — we haven’t published this as a study, so this is pseudo anecdotal but it’s based on real data — all of those kids in the bottom quartile are now operating in the top decile. Because, I would argue, they had a chance to learn at their own time and pace, fill in those gaps.

When you get an 80 percent in a traditional system, you’re told you’re a C student, when our students get an 80 percent, they’re like, “You’re at an 80 percent; keep working on it to get to 100 percent.” So it completely changes even how they perceive themselves. The expectations are fundamentally different.

We are also seeing really powerful things with the mixed age environment. There’s actually some research about this, that when older students are with younger students, all sorts of good things happen, which is, frankly, in line with human nature.

For most of the 300,000 years of human evolution, we lived in these multi-aged clans or tribes, but especially, I think, the positivity happens for the middle school or the high school students, because they now feel responsibility. And I think it takes a lot of that negative energy that often goes into bullying or all of that in middle school angst and it takes it into a positive place.

We actually have our first graduating class this year. And we’re seeing that they have been able to navigate what is otherwise, especially in Silicon Valley, a very stressful college application process. And they’re getting very good college placement. It’s really what the colleges have been asking for, kids who have authentic interests, who actually have mastery of content, which is exactly what the school is focused on.

LEVITT: How long is the waiting list and how do I get on it?

KHAN: The rate limiting factor to your point has actually been real estate. Silicon Valley real estate is not a trivial thing. And so we did have the school limited pretty significantly but now we’re moving, actually, the upper school to co-locate with a community college where these kids are going to be able to have this really great journey but also finish a good chunk if not most of college before they graduate high school.

And then we’re moving to another campus. The whole school K-12 right now is about 200 students. We think we’re going to expand to about 300 students.

LEVITT: What prevents you from doing five of these schools or 100 of these schools or 10,000 of these schools?

KHAN: Well, that’s part of the idea. We’ve been trying to stay relatively under the radar for the last seven or eight years. But now that we are showing that it’s working, that the kids are healthy and happy. Over the next few years, you’re going to see us spread much faster.

One of the things that I keep pushing the team is we’ve got to show that it can be done in an economic model that could fit most places. It’s now at the scale that it’s about — it’s pretty close to the average cost per student for the local public schools, if you factor out the cost of real estate, that we have to pay real estate and a lot of them don’t.

LEVITT: One of the potential challenges to this model is that there’s a real efficiency from a cost perspective of having one teacher teach 30 kids the same thing, whether or not it’s the right thing to teach the kids at the same time. It’s obviously not, but it’s an efficient way to do it.

But what you’re saying is in the Khan Lab School, you’re actually finding that because of probably cost efficiencies that come from other sources, like the teachers don’t really have to lecture anymore. Instead, the kids are largely working independently, I would suspect most of the time, you can actually make this on par in terms of costs with the traditional teaching model.

KHAN: Yeah. And it’s not just the kids working independently. In fact, they are oftentimes spending time with teachers, but that time with teachers is much more focused. So it might be a small group for 15 minutes that is focused on exactly what those kids need at that moment. While another group of kids are learning at their own time and pace.

And we’re also seeing this type of phenomenon in mainstream public classrooms, but the school is able to rethink everything. So even how we do scheduling and use of space and all of these other things that enable these ideas that, hey, your age shouldn’t be the indicator of what you’re getting exposed to. That if you’re ready to move ahead or if you need to back up, either way there’s no shame or no stigma, you should just be able to do that.

When I meet educators, they all agree, class time should not be a passive lecture. It should be interactive. You have to differentiate for students. This is always what’s been the gold standard, but it’s just been hard to implement in the traditional Prussian education model that we developed from 200 years ago.

LEVITT: So you’ve hit on this model that really works and you’ve proved it to yourself, at least, with the Lab School. Do you not have impatience for making broader transformation? I would think you would bristle at the fact that you see these answers and it’ll be, how long — decades at best — before these are available in a widespread way? Does that frustrate you?

KHAN: Oh, yeah. Don’t be fooled by my relaxing tone of voice, which, maybe I’ve fine tuned over thousands of videos on Khan Academy. But no, I’ve had to develop a practice of meditation so I can sleep. I’ve been known to be so obsessed and agitated and impatient about things and want to make them happen fast because — yeah, the real wholesale change, that is a long game. That’s going to be decades.

In the whole scheme of education, if we can get a major change in 10 or 20 or 30 years, that is worth it. Khan Academy, we did surveys at four-year colleges, and this was a few years ago when Khan Academy was at a lower scale, and the question that we randomly sampled students: Has Khan Academy had a meaningful impact on your education? And we thought it’d be lucky if 10 percent of college students said “Yes.”

It turned out that 60 percent of college students at the schools that we surveyed said, “Khan Academy had a meaningful impact on my education.” So that tells me that we’re already making a dent, we just have to make more and more of a dent between adding more subjects, making the interface more engaging.

And then there’s another piece that I’ve been impatient about around credentialing. We’re realizing: wait, if someone is able to get mastery on Khan Academy, if they can record themselves, both their face and their screen, while they take the unit test, while they explain their reasoning out loud, and it’s peer reviewed, that then certifies them to go on a journey to become a tutor. But it also shows that they know the material.

And University of Chicago has been using certification as a factor in college admissions. We’re talking to universities about, “Look, if someone can tutor pre-calculus or someone can tutor biology, why can’t we give them college credit for that tutoring?”

To your point, you’ve been pushing me to say that I’m being more bold than my tone let’s off. These are big ideas. I want to create a world where for close to free, you can essentially get everything that you need to know to work, to get a job, to go to grad school. Now for some kids that might be all they need, but for other kids, they’re going to need an in-person community. They’re going to need support. In fact, I would argue for most kids. And so that’s where you want to interface well with the existing system.

LEVITT: So the test prep industry, firms like Stanley Kaplan, the Princeton Review, that’s a multi-billion dollar industry. Now Khan Academy comes along and provides completely free, amazing S.A.T. prep materials. Do you think you’ll cause that entire industry to unravel?

KHAN: College Board, the folks who administer the S.A.T., they reached out to us saying, “Look, the College Board has just ignored the test prep industry because we never wanted to validate them, but at least the perception of inequity and maybe the reality is real. We have to address it.”

When Nadia was preparing for the S.A.T., I went through the College Board 10 S.A.T.’s book, and I did every math question in video form. And College Board said, “What we loved about those,” which apparently they were looking at, “is that you never did what the test prep industry did, which is like, ‘O.K., this is hard. This is how you guess.’”

I always said, “Look, you might think this is hard, but this is really just an application of what you learned in Algebra 1 where if you think about it this way, or if you’re having trouble with it, why don’t you go review this other video or do these other exercises on Khan Academy?”

And they’re like, “That’s how it should be. It should not be gaming the test. It should be becoming more college ready. And then the test will measure that.” And that’s when they said, “Would you want to partner to make the world’s best test prep in conjunction with each other that happens to be free?” And we said, “Yes, absolutely.”

LEVITT: So, obviously, the aim was to democratize the S.A.T., but a few times in my own research, I’ve set up plans where I make things available to everyone with the goal of hopefully having high take-up on the bottom end and maybe less high take up on the high end. But oftentimes, I’ve found it’s just not true. It’s the opposite. Are you able to see in your data whether you’re hitting you know the middle and the bottom more than the top? Or how does it pan out?

KHAN: It’s a really good point. And this is something that both us and the College Board wanted to really understand. What we have seen over the years is the proportion, if you go by race, it’s about the same as the proportion of test takers of the S.A.T. We are slightly over-weighted on Black students and Latino students, but it’s roughly the same proportion.

Now on income and on parents’ education level, we have seen some of the phenomenon that you’re talking about, where we have a higher proportion of children from higher income and kids with more educated parents. With that said though, you don’t look at over-proportion/under-proportion to understand whether it’s driving equity or not. You have to say: what does the world look like with this resource? And what does the world look like without this resource?

So the world without this resource, the affluent, they’re going to Kaplan. They’re going to Princeton Review. They’re getting the help. The poor kids aren’t getting the help. You add Khan Academy, now the rich kids, some of them are staying with Kaplan and Princeton Review, but some of them are saying, “Wait, this Khan Academy thing is better.” So they’re switching to Khan Academy not because it’s free, they’re switching because it’s better.

That’s actually validating of the platform. It would be a problem if the rich kids stayed with Kaplan and Princeton Review and only poor kids were using the new platform. It would continue that perception that free isn’t as good as paid.

And now all of a sudden you have a huge swath of kids from historically under-resourced groups who are actually able to get test prep. The analogy that I draw is a museum, or a library, or park. In many cases, museums or libraries or parks might be disproportionately being used by affluent people, but if you take those parks away, the affluent are going to go to their backyards, or they’re going to form private parks. Or if they take the libraries away, they’re going to buy the books. They’re going to have access, while the poor are not going to have access.

LEVITT: So I’ve been spending a lot of my time lately fighting a different educational battle than the one that you’re focused on. I strongly believe we are teaching kids the wrong things. And I believe that much more time and effort should be put into teaching students about data analysis, understanding the difference between correlation and causality, how to work with spreadsheets or data visualization, computer programming. What’s your view on this?

KHAN: Yeah, I agree. And I would add, personal finance or just general finance in there, probably a little accounting, a little bit of law. I will say as someone who — I love calculus, I teach calculus. I’ve made hundreds of videos on calculus — statistics and data science and all of these things that we just listed are more useful than calculus for pretty much everyone.

There is a system, it was decided essentially in 1890 by the Committee of Ten, which was 10 university presidents in the U.S. that sat down and said, “What is important?” And they’re the ones that actually determined that you learn chemistry in 11th grade and biology in 10th grade and physics in 12th grade. And obviously, they were doing that before we even knew about DNA or we had the modern Federal Reserve or an interstate highway system, and we really haven’t been able to revisit these things because it has so much inertia to it.

LEVITT: I also think that we should put much more effort in schools helping children to understand their own emotional lives. So mindfulness and conflict resolution, dealing with stress and trauma. And these are things that every adult I know spends an enormous amount of time on, belatedly. What do you think about bringing that radical change into school curriculum and how might you do it? If you agree that would be a good idea.

KHAN: Oh, I’m a huge believer in it. I often shy away from talking publicly about it because it’s not measured in the same way that your math scores are, that it doesn’t get as much attention. But at our Lab School we don’t just have P.E., we have outer-wellness and we have inner-wellness.

The students meditate; they’re learning to self-regulate; there’s some hidden cognitive behavioral therapy that we’re working with the kids where they learn to become more aware of their own thoughts and aware of their own feelings. We have epidemics right now, globally, but especially in the U.S. around stress, anxiety, loneliness — how can we address these? Our Lab School, I would say about 30, 40 percent of the energy is on that. Because the academic stuff is happening in a lot of ways, easier, we’re able to put more energy there.

I have published on Khan Academy, some guided meditations for students and things like that. Personally, I started meditating religiously for the past two or three years and it’s changed my own life. I continue to be impatient and angsty about all of this change happening, but I wake up in the middle of the night a lot less.

LEVITT: When I talked to schools, people in education, about data science, the answer’s always, “Yes, that makes a ton of sense. How can we do it?” And when I talk to them about emotional wellness, usually it’s complete silence. I think the schools just feel completely unprepared for taking that step.

KHAN: They’re feeling overwhelmed because even on the things that are measured, which is essentially math, reading, and writing, we know the data is not good. And so I think a lot of these systems feel so overwhelmed. The reality is if kids actually learn to write at an eighth grade level, if they learned to read at an eighth grade level, and they truly master eighth grade math, they’re going to be pretty empowered.

Because the sad reality is very few high school students have actually mastered reading, writing, and math at the eighth grade level by the time they graduate from high school, many times by the time they graduate from college, they really haven’t done that.

LEVITT: Almost nobody uses 10th or 11th or 12th grade math ever in their everyday life. I agree with you 100 percent that a great goal would be for everyone to understand ninth grade math at the end of 12th grade. But the way the incentives are set up now, and the way the high stakes testing works, schools are incentivized to piecemeal as best they can 10th grade, 11th grade, 12th grade learning, and it doesn’t leave any time to go back and actually get the mastery of ninth grade.

KHAN: I would argue that they could be incented right, it’s just it takes a little bit of a leap of faith. The same is true of the higher education system. My hedge fund days, I remember talking to a C.F.O. of a company, and I remember like, “Well based on my calculations, it looks like you make about,” I think it was like a jeans company, “it looks like you make about $20 per unit on your jeans.” They’re like, “How did you know? That’s not public information.”

You published your quantity you sold and you published your profit in two different periods. It’s called a system of equations. I learned that in eighth grade. And this is someone who probably went to fancy business school and they learned all that stuff. They probably got A’s in it, but they didn’t realize that this is just basic algebra. If you really master 12th grade, or first year college biology, chemistry, physics, and math, you are dangerous in the world, in a good way.

LEVITT: You’ve had so much success. Do you have trouble reining in your ego at all?

KHAN: That’s why I meditate. No, well, look. For me, I’m a big believer in: your ego creates most of your suffering in the world and most of the suffering in your own mind. I’d like to believe that I’ve never taken myself seriously. I’ve taken some of what I’ve worked on very seriously. When I worked at the hedge fund, I was driven. “O.K., I have to do well.” Obviously, part of my ego was like, “I want to be ‘successful’ in this field.” But a lot of it was, “I want to be able to support my family and do things that are intellectually stimulating.”

But then when you transition into a mission like Khan Academy’s, you’re like, “Wow, this is important. If you fail, Sal, maybe millions of people aren’t going to be able to learn as much or whatever.” And so that creates a certain form of stress. I guess you could say it’s somewhat ego-driven in that it’s associated with your identity. It’s just, “Sal, you’ve tied your identity to this mission. You’ve got to make this succeed.”

And one thing that’s helped me not be as stressed or anxious, although I still am sometimes, is to say, “Sal, you’ve got to let go of that.” Your identity, your persona, your — these are just completely human constructed things. And trust me, I have more than my share of failures and times where I’ve been knocked down and things aren’t looking good.

And so I also need to remind myself in those times that, look, “This isn’t about this thing called Sal, this physical body, this persona that you identify with. This is about just trying to do, every day, what you think is right. And then the chips fall where they do.” This goes back to things like Buddhist and Vedic philosophy of, “Don’t get attached to the results. That’s what’s going to cause suffering, but do what you think is right.”

LEVITT: Is there anybody in your organization who will tell you you’re wrong?

KHAN: Is there anyone who wouldn’t?

LEVITT: So, that’s good. I mean, so, you cultivated that because that’s really hard. You’re such a good talker, and you’ve had such great results, and you’re such a visionary. It’s really hard for me to believe that the people who would work for you would feel the ability to tell you’re wrong.

KHAN: I’d love to hear from them, but what you just said, it’s very nice and very flattering. But I think, if anything, I almost feel like the people who work a lot with me almost go into an engagement with me saying, “Well, Sal can be convincing. So I’m almost going to put extra guards up to be extra ‘play devil’s advocate.’” Obviously, you want to cultivate a place where it’s not about being right. It’s all about the mission.

LEVITT: So it’s easy to say though that you want to have an organization that is mission-driven, but are there particular things you’ve done at the Academy — are there norms around failure, or around telling you you’re wrong — that makes it easier for people to do that than at typical organizations?

KHAN: One of the big questions when we started, especially at a not-for-profit, is: how are we going to attract top talent, especially in a place like Silicon Valley, where people could go work at Google and Facebook and get all the stock options and whatever else?

I remember reading a study back when I was a hedge fund analyst. And it said that beyond a certain level, pay and things like bonuses, actually they’re not that great of an incentive or they can even be a disincentive at some point. And what really people care about, they definitely need enough pay to feel valued and feel like they can support their family at a reasonable standard of living.

But beyond that, it’s all about having mission, that you feel a real purpose; feeling that you’re working around other people who are alongside and are invested in you; and having intellectually stimulating work that really leverages your skills.

When we give job offers to let’s say an engineer or a designer, most of them already have job offers or are coming from many of these tech companies that pay a lot. And we get 70 or 80 percent of them. Not only are we attracting these people, we’re getting some of the best people in the world joining us because I think some of the best people in the world are drawn to purpose.

LEVITT: There are a lot of people who are thinking about careers, and they’re not sure whether to follow a for-profit path or a not-for-profit path. Do you have advice for people about how to make a wise decision in that case?

KHAN: When I meet people who are in college or just graduating from college, I tell them just to ask a lot of questions and don’t be afraid to ask nosy questions. And then, based on that information, find a path that, especially early in your career, is going to grow you and challenge you the most. Where you’re going to get invested in, you’re compensated fairly, you have great managers who are willing to keep giving you more and more, frankly, as much as you’re willing to take.

But at the same time, make sure that you have space for your interests and your passions outside of work, that you don’t define yourself, your persona, by just that one job. I give the example, the hedge fund I worked at, it’s non-stereotypical. Dan Wohl who was my boss — I was ready to work 80 hours a week. Because I thought that’s what you had to do at a hedge fund.

But I remember after the first week or two was like, “Sal, why are you still working?” It was like 6:00 or 7:00 p.m. And I was like, “Oh no, Dan, I’ve got to analyze these companies.” He’s like, “Go home.” I was like, “O.K., I’ll analyze them from home.” He’s like, “No, no, no. Go home and don’t do this.” And I was confused and he was like, “Look, Sal, our job as investors is to avoid making bad decisions and make a few good decisions every year. If we just work ourselves silly, you’re going to end up making bad decisions.”

And so he’s like, “You need to have other interests; you need to recharge. And when you come to work, be there 110 percent.” I was like, “All right, Dan, I will do that.” And that’s what allowed me to tutor my cousins and write software for them and all of this.

A lot of my peers who went on a similar track to me. They weren’t lucky enough to have a Dan Wohl as a manager. By the time they were in their 30s or 40s, they were making a ton of money but they’ve completely forgotten what they wanted to do with that. They’ve completely forgotten their passions or they’ve gotten the golden handcuffs.

One of the good things is I’ve always lived quite frugally. So I’ve never been attached to a certain standard of living. Part of it is fear. When I grew up, my family wasn’t well off, but that also gives you a certain form of freedom so that you can sometimes pursue these other things.

LEVITT: Yeah, I’m glad you added that because I was going to say that I give almost the exact same advice that you just gave, but with one addendum, which is: if you do take a job that pays you a lot, don’t get addicted to the money because then you’re stuck. And those golden handcuffs are brutal.

LEVITT: I love the ideas underpinning his Khan Lab School. And I was only half-joking when I asked about getting my kids on the waiting list. My wife and I actually talked about moving to Silicon Valley for that reason, but we decided it would be better to get Sal to start another Khan Lab School here in Chicago. So that’s what’s at the top of my to-do list for today.

*      *      *

People I (Mostly) Admire is part of the Freakonomics Radio Network, which also includes Freakonomics Radio, No Stupid Questions, and coming soon Sudhir Breaks the Internet. This show is produced by Freakonomics Radio and Stitcher. Morgan Levey is our producer and Dan Dzula is the engineer; our staff also includes Alison Craiglow, Mark McClusky, Greg Rippin, and Emma Tyrrell. All of the music you heard on this show was composed by Luis Guerra. To listen ad-free, subscribe to Stitcher Premium. We can be reached at That’s P-I-M-A at Thanks for listening.

KHAN: Trust me. I have many low self-esteem moments, so I will call you, anytime. Remind me of how dynamic I am. 

Read full Transcript

Episode Video