How Khan Academy Will Help Find The Next Einstein

Earlier I wrote an article, Five Lessons from Salman Khan on the Future of Education, that included core insights from Sal that I distilled from our full discussion.

Sal kindly shared his thoughts on what it means to be creative, intellectually precocious kids and the Khan Academy, how he enjoyed gifted programs as a kid, how Khan Academy impacts the role of theory in education, and why he thinks it is socially acceptable to be bad at math in America today, among many other topics.

JON: You mention in a USA Today interview that your “mission is to have every precocious 13-year-old in the world have access to every bit of information they could ever want.” Is Khan Academy serving as a haven of sorts for talented kids from all around the world?

SAL: We haven’t done any rigorous research, but my best guess is that it seems like that is the case. It’s not just a haven for precocious youth, but I suspect our penetration among precocious kids is really high and a high fraction of our users are kids who have a love of learning.

Your personal story shows that you were clearly academically precocious, taking graduate level courses at the University of New Orleans while still in high school and earning two bachelor’s degrees and one master’s degree from MIT in just four years. You essentially accelerated your own education. How do you think Khan Academy can help talented people from around the world accelerate their educations? And what will the impact of this be?

We’re still in a primitive state in terms of the tools that we’ve created, but our vision for the future is that everyone will be able to learn at their own pace and it should be competency based. Once you feel like you know something you can prove it, and the world respects that, and maybe you have to maintain that knowledge state, it’s not that you just have to prove it once and not have to worry about it. In a world like that I think that you are going to expand the number of people who are these precocious self-learners, and we can engage far more people because they are going to have a chance to take ownership of their learning and they’re going to have a chance to not be bored or lost. But, yeah, I can imagine a world where there are students who want to go deep and go fast and they’re getting it at a deep level and they’re getting it quickly, who knows what might happen. One dimension is that you can empower that student, hopefully where one day Khan Academy will be taken very seriously where someone will say “Wow, so you did that on Khan Academy? You’re ready for the next level now.”

Some kind of certification?

Yeah, a certification or a micro credential or whatever you want to call it. And then there’s the whole issue of talent identification when there’s someone off the charts. We’re going to get the data, the analytics, on all these kids and we don’t just get these snapshot SAT scores or whatever else, we get these data narratives. And so you could almost identify talent as well, I would think.

Yeah I think you can. You could see how quickly they go through certain lessons on the Khan Academy. Math in particular.

I would say math in particular and the one dimension that I wrote a lot about in my book that is lost in traditional assessments is the most important which is the dimension of creativity. And there might be a correlation between these precocious kids and their ability to create novel things. And we’re not quite there but we’re about to start launching things where people can subjectively assess other people’s work. And so you can start to have assessments that are not just objective, but that could also be used to identify talent subjectivity. For example in computer science, if someone makes a really out of the box way of visualizing the Pythagorean Theorem, or writes some code that does something amazing, that can’t show up on a test like the SAT but we could have the crowd assess that type of thing.

That actually brings up another question. Some of my research shows that math and spatial ability which are measured through traditional tests longitudinally has an impact on performance in science, technology, engineering, and mathematics (STEM) fields. You talk in your book how you are concerned why a lot of people don’t see STEM as creative, or engineering as creative. Could you talk about why you think people don’t view it that way?

I think there’s a couple of dimensions. One, the unfortunate way that STEM or math is presented in early grades is very very mechanical. Even when there’s an opportunity to make it less mechanical once you start getting into algebra, unfortunately it is disproportionately taught in a very formulaic way. So you don’t really get into the creative side of it. The other dimension is that I think our culture has been defined in a way that creativity is a cool thing. Nobody wants to say that “I’m not creative.” It’s cool to be creative.

Yeah it’s very “in” right now to be creative.

It has never not been in to be creative. At any given point in time you are perceived to be creative if you create neat things and are a good artist or a musician or a singer, which I think people associate with being creative. And I think the people who are creative in the way that people tend to think of as creative right now—the singers, the writers, the dancers—they oftentimes have trouble with math. I’d argue that’s because the math was taught in such a non-creative way that they didn’t respond well to it even though they could have. Sometimes the creative people look like they are different from the math people. But then there is an overlap, and then you say “Oh that person is a math person and a creative person.” And you’re like “Oh, they’re well rounded,” somehow implying that those are two different dimensions. If we could get those creative people—whoever they might be—to say “Hey I’m using these math tools to further my creativity” this would help. I think people are starting to appreciate that, and that is what computer science has really done. Even a young student who is interested in visual arts or music or design, they’re starting to realize that there are all these amazing tools that can bring them into engineering, into programming, and into mathematics. So I think that is changing.

That makes me think of building an App, for example. Now that’s cool. If you build the App that goes to number one on the list, you’re pretty cool now right?

Exactly. If you just build any App, even if it’s not number one and you’ve created something in high school that will help you decide where you’re going after the football game, that’s a cool thing to do.

According to the National Association for Gifted Children website, federal public funding for gifted education is .02 percent of the entire K-12 education budget. This means kids who are smart but who come from low income backgrounds are likely being underserved because they rely on such programs. Where do you think the Khan Academy might help in this regard?

Well the great thing about the gifted programs—which I was lucky enough to go to starting in 2^nd grade—is that you show up one day out of the week and you have these two amazing teachers with Ph.D.s in psychology and they ask you “What do you want to do?” and you just start doing stuff and you’re playing and some of my best experiences come from that. And before you felt you had to separate those gifted kids because everyone got that Prussian model thing and frankly there wasn’t enough resources to give everyone this gifted thing, even though everyone could have benefitted from it to some degree. What I’m excited about is that the tools are emerging and the mindsets are changing—and my book is essentially about that—where everyone is going to get the gifted program. Any system should make sure that everyone is well served. I see the reality where in the same classroom gifted kids can be served very well and a kid with a learning disability right next to that student being just as well served.

Is that with the blended learning model?

Putting technology aside, if you could have a classroom where everyone goes at their own pace, there’s a lot of peer to peer tutoring going on—where the gifted student is tutoring others or even the non-gifted student is tutoring others—which I think questions some of the definitions of gifted. A learning model where it’s not lecture based but focused on small group activities with peers and teachers and their projects. That I think is the right model, a competency based model, where I’m going to learn at my own pace, show you that I know something, and then we’re going to move on. And I’ll retain that knowledge state as well. I think this is a good idea, it’s what my book is about, and that’s what a lot of the research has backed up. Historically that has just not been easy to implement. The technology in blended learning makes that a reality. That experiential thing is the goal and blended learning is a means to that goal. For me it’s about getting to a self-based competency based world.

Richard Snow was a famous educational psychologist who came up with a program called QUINTILIAN. This program tested to see where a student’s strengths and weaknesses were across different abilities, and then matched a curriculum to their profile of strengths and weaknesses. I wonder to what extent Khan Academy is naturally doing that as the student is going at their own pace. Is that true?

Yeah I think we are doing it in a very primitive way right now. Right now some people look at the videos and figure out if they can do it or not. Precocious students are going to be better at that. We have some assessment and practice on the site. What you’re going to see on Khan Academy over the next six months to a year is much more rigorous deep assessment diagnostics on the site. This will really help people fine tune the activities that are going to be the most appropriate for them at any given time.

In The One World Schoolhouse you note that “We can now craft more particular and individual solutions than ever before, thanks to the availability of data from millions of students on a daily basis. This is not theory and this is not the future. It’s happening in the real world and it’s happening now.” In the age of “Big Data” such as the kind you are collecting on a daily basis, it appears that theory is not all that important anymore because we have enough data to find out what actually works. What are your thoughts?

In The One World Schoolhouse you note that “We can now craft more particular and individual solutions than ever before, thanks to the availability of data from millions of students on a daily basis. This is not theory and this is not the future. It’s happening in the real world and it’s happening now.” In the age of “Big Data” such as the kind you are collecting on a daily basis, it appears that theory is not all that important anymore because we have enough data to find out what actually works. What are your thoughts?

I kind of wrote about this in the book. You have all this education theory, and people try to make larger statements than maybe what their data would back up because they’ve done these small experiments that are tied to a very particular case with a very particular implementation. I wouldn’t say that strong of a statement, theory definitely matters, but I think dogma matters less. We can say, well “The current established theories say we should be conscientious of this but let’s just test it again.” Maybe the results were only particular to that time. Now we can run very similar studies with much less pain and perhaps the theory won’t apply. That’s what’s exciting. Things like Khan Academy can be very powerful in this context. Theory tries to be very broad, it tries to make these laws of learning. But they’re almost always too general. And if you made it too particular it’s almost useless. For example, “If Jonathan teaches these types of kids, it’s better when Jonathan uses this technique won’t work because it won’t pertain to Sal.” Now the results will be particular to the Khan Academy implementation, but they will be impacting millions of students.

I’m very empirical and am not easily impressed by theory that isn’t built on empiricism, which is probably uncommon for an educational person. But I feel you can actually build a theory based on empiricism now, right?

Yeah, absolutely. You kind of have to ask yourself, what’s the whole point of theory?

Exactly. If you have a black box and you know you put in this and you get out that, and you know it works, the why comes after that.

Yeah, if it works it works right? And the whole point of theory is to generalize so that you can scale or for the advent of scalable systems. So now I think theory is nice if something seems to be working we can say we think this is why and we can explore the problem space more in that direction and perhaps the theory will be helpful for future experiments. But at the end of the day, the experiments are what they are and we just need to keep exploring the space.

In the Khan Academy factsheet, you have a question to address concerns about how Khan Academy may be “enabling rote learning.” This is a common criticism of Asian educational systems. Yet in academic international comparisons, those Asian nations are outperforming America, particularly in math and science. Bill Gates has said that “You need to understand things in order to invent beyond them.” Do you have any thoughts about this?

I think most of the criticism is targeted at the videos. And I tell people that they should watch ten videos for themselves and see whether they are more focused on rote learning or whether they are more focused on conceptual reasoning and understanding. The reason the gifted students of the world like Khan Academy is because we don’t say “memorize this formula,” but say “let’s try to derive it from core principles” or “I forgot my trig identity, so I’m going to just try to prove this to you.” At the same time we don’t talk down to students. Sometimes I think people confuse rote learning with traditional conceptual instruction. A good traditional conceptual instruction is what I got from my better professors at MIT. They would be at a chalkboard, and they would literally be explaining something and working through a problem, but it wasn’t rote. They were explaining the underlying theory and processes and intuition behind it. And some people just assume that if you have an equation on the board or you are working through mathematical symbols that it is somehow rote, which is not the case. We do also have a lot of worked examples on the site, in fact probably six or seven hundred. And perhaps this is where people come up with the perception that these are just a bunch of step by step problems, but I personally think it’s valuable to see worked examples because this is how I liked to learn as a kid, but at the same time the worked examples are very seldom memorized. There are the rare instances where I have to give a formula without proving it but I say “I hate memorizing things and you should not have to memorize things, but every now and then it might be helpful.” So I think there’s a press narrative now that a lot of kids like Khan Academy and it’s really popular. And because of this there’s probably a small minority who say, “Oh, it must be superficial then because it’s popular,” or “It must be Cliff’s notes.” And they haven’t even looked at the videos. And when they do, they’re looking to confirm what they already think or that bias. But there are many teachers including those at the university level who are saying that it’s not rote. The videos are focused on conceptual understanding. But they are also focused on having a lot of problems which is a little more like a traditional model. And we do have exercises that I wouldn’t call “rote” but would call “deliberate practice.” The difference is rote means just memorizing the formulas. Deliberate practice is when we say “Look, we’re going to have you solve a bunch of equations. And solving these questions will help you get into the rhythm and logic.” I personally believe that most of the time if you’ve worked yourself enough, your brain starts to draw connections. Sometimes the connections happen before the problem solving, sometimes the problem solving happens before the connections.

You have to have a base of knowledge before you can make connections right?

That’s exactly right. There are basic rote things that people do have to know, such as the multiplication tables and basic addition and subtraction. And even once you get into higher mathematics, you’re more likely to be able to engage at a more fluent level if you do have some key things that are in rapid access memory.

You were saying in your book that in math and science in particular if you miss something you’ll just fall through the cracks at higher levels.

Mastery or competency or proficiency is a different thing than rote. In terms of rote your Kumon math worksheets may get you there and it’s probably better to do them than not. But my main thing is to make sure that students really deeply understand the conceptual underpinnings. In regards to your point about the Asian comparison, I would say that the kids who do well in math in the US actually have the best of both worlds. They’ve done a lot of practice and different types of problems, but at the same time they have thought more critically and more out of the box about things. I talk to my relatives in India, and they went through pressure cookers, but I feel like me and my friends over here understand math a lot better. It’s both ways. If you do look at Singapore math they are actually less rote than the American system. The books are very small and they’re focused on hitting the concepts on multiple dimensions. They’re not about doing as many mindless problems but doing pure problems that are more interesting. I think Asian families are more likely to say “Go do five hundred math problems right now” but the Indian curriculum as I understand it is not that good and I don’t know about China and places like that.

I do know that rote learning is very common in Chinese culture because of learning the language, for one thing. I think the interesting thing—to connect this back to the whole international comparisons—is not at the testing level of students, but at the level of our graduate student population. A large number of our STEM graduate students and Ph.D. scientists are from other countries. So the thing is: they are being creative, so maybe the rote learning is not a terrible thing?

Yes, some large fraction is from other countries. India and China are probably the bulk. But these people are still a very small fraction of their home populations.

Yes, they are very select.

Exactly. So, the US in terms of the fraction of its own population is probably doing better for the number of people in STEM graduate programs. There is also a cultural dimension of whether it’s cool to be in STEM or not, and then there’s the dimension of opportunities available to you. So if you graduate from Harvard in math with a 4.0 GPA, yeah you can go to grad school in math if you wanted to, but you could also get a job at McKinsey or become a quant trader at a hedge fund and make a lot of money. You could go work at Google, or go to law school or medical school. There’s really a lot of interesting things you could do. In most Asian places there’s more cultural emphasis on going into STEM and that’s often the only outlet for becoming upwardly mobile.

Why do you think it is socially acceptable to be bad at math in America today?

People who are looked up to in America will often say “I don’t get math.” And they’ll often say it like they’re being humble but they seem almost proud of it because it’s acceptable to act this way. And frankly they were probably trained to say that even if they are good at math because around middle school they got beat up if they said they were good at math. It’s just a cultural artifact. Barack Obama has never said it, but I’ve heard politicians say, “Yeah, I’ve never really been good at math.” People in the media and press often say they’ve never been good at math. It might be that people that consider themselves creative didn’t consider themselves good at math or didn’t find math interesting at those early stages. And those creative people are disproportionately represented in those influential roles.

Verbal roles, like the media.

They’re writers, they’re media. And I think these people who are off the charts in their respective fields I think could easily have been very good at math or even amazing in math. But maybe they failed a test at some point which made them think they weren’t good at math.

Do you think the Khan Academy can help us find and develop the next Einstein?

I hope so. Einstein was definitely a unique character. But you’ve got to think, okay, given some natural gifts that someone like Einstein had, what’s the probability that they will have access to an education where they can cultivate that brain or raw talent? Right now it’s one in X, I don’t know, but it’s not a high number. And then, what’s the probability that they decide to go into physics, research, or STEM, depending upon whether you want an Einstein in physics or if you want an Einstein in biology or medicine. My sense is that in Einstein’s generation there were probably a hundred potential Einstein’s that were just completely squandered. They turned into some crazy person on the side of the street or something. Who knows? Now we can go into a reality where that could be nurtured in a more scalable way and be identified in a more scalable way. And this is the beauty of the internet. There’s a classic example of the genius who is not so appreciated by the non-genius people around them, and then they get told that they’re an idiot and so they think that they’re nothing. But the internet now allows that genius to get that exposure and hopefully other geniuses around the world will recognize that genius for what it is.

© 2012 by Jonathan Wai

You can follow me on Twitter, Facebook, or G+. For more of Finding the Next Einstein: Why Smart is Relative go here.

Also see: Five Lessons from Salman Khan on the Future of Education