Homeroom with Sal & Anant Agarwal - Thursday, June 24

Hi everyone, Sal Khan here. Welcome to the Homeroom live stream! We have a very exciting guest today: Anant Agarwal, founder and CEO of edX. Sorry, I'm messing with my video settings probably at the exact wrong moment, but before I get into that, I will give everyone my standard announcements.

A reminder that Khan Academy is a not-for-profit organization. We can only exist through philanthropic donations from folks like yourself, so if you're in a position to do so, please think about making a donation at khanacademy.org/donate. I also want to make sure to give a shout-out to a whole series of organizations that stepped up when they realized that the world needed Khan Academy more than ever during the pandemic. But also, to support the world, we needed more resources. Our server costs went up; we were trying to accelerate a whole series of content. So special thanks to Bank of America, AT&T, Google.org, Novartis, Fastly, and General Motors for really stepping up during the pandemic and beyond the pandemic.

And to all of our supporters — corporate, philanthropic, all of the above — over the years, to really put us in a position so that we can support as many of y'all learners as we can. Last but not least, I want to remind everyone that there's a version of this live stream in podcast form, wherever you get your podcasts: Homeroom with Sal, the podcast. So check that out wherever you get your podcast.

Now with that, I'm very excited to introduce our guest today for many, many different reasons. Anant Agarwal is the CEO and really the founder of edX. He's a co-member of the Yudhan Council of Luminaries. He is a winner of the Non-Prize in Education. But maybe what most people don't realize is that you were my professor in college! You were my recitation instructor. I believe it was in either circuits or signals and systems; you might remember better.

You know, I think it was in circuit-centric electronics. You know, as they say, this is the one case where I think the professor has learned a lot more from a student than the other way around.

Oh, I know so much about circuits!

Okay, okay! Alright, I'm going to ask you my first question. I don't—I mean, in all fairness, my degree is computer science and electrical engineering; it's much more computer science. My wife rolls her eyes because you would assume with an electrical engineering degree I would be able to be more handy around the house, but I explain I'm very theoretical and I'm afraid of electrocuting myself or others, so I turn to professionals. But give me some software, and I could definitely go to town with that!

Before we go into what you do now, I mean, I’m sure you remembered me when I was your student! You know, in the sea of bright faces, you're like, “That kid!”

Yeah, I had you pegged! When I saw the face, I said, “This person is going to be one of the world's greatest luminaries in terms of teaching and learning.” I had you pegged! And you know, when you went off into the financial systems and so on, that was a total surprise. But what you're doing now—you know, I had you pegged, buddy! [Laughter]

Well, I will take that for what it's worth. I'm pretty much sure that you don't remember me, and I was just a face in the crowd. But let's go into what you do now.

So, for anyone who's been living under a rock for the past ten or so years, what is edX?

edX is a learning platform where we connect some of the teachers and professors from the greatest institutions in the world, like MIT and Harvard, Stanford, Berkeley, Oxford, and others, to learners from all over the world. So it's a platform where learners who want to learn can come in and learn from some of the best professors and teachers around the world and also from companies like Amazon Web Services, Google, IBM, and others.

We have about 40 million learners on our platform, where learners come in and they can learn anything from the humanities to computer science to data science, business. You can even learn how to use mechanical ventilators — a course that we launched with Harvard and MGH during COVID times about a year ago. People can also earn credentials. We have these modular credentials like MicroMasters and MicroBachelors that people can use to upskill in their current job, or if you've been furloughed, you can use that to gain new skills for the future.

So this is who edX is! And you can come in, and you can start learning; you can learn for free on the platform.

Absolutely! How did all of this come to be? You know, you were a professor—a very successful professor at MIT. You had your tenure; you were running departments. You kind of gave that up to do this. What was the evolution of going from professor to edX and making this your life?

You know, Sal, you had something to do with it. I would trace it back to—you know, for me, education has always been something near and dear to my heart. I remember as a physics student at IIT Madras in my first year, I was one of two students out of 300 that failed the first physics exam. And my part completely stopped. At that time, I realized that everybody around me seemed to understand calculus-based physics. There was no Khan Academy then, remember? And I was here; I hadn't learned calculus-based physics and had no idea what they were talking about.

I said, look, hopefully, there'll be one day when everybody will have equal access to education and come up to speed as everybody else. I really got a chance in the year 2000 when MIT launched OpenCourseWare. I was one of the early professors to launch my circuits course on OpenCourseWare.

And then, Sal, I've heard you talk about how you were inspired by OpenCourseWare and you launched Khan Academy. I heard about your videos from my children and others. At MIT, a number of us were thinking about what should the future look like? We did OpenCourseWare. What would the next big thing look like? As a confluence of multiple technologies happened around 2008-2009, like social networking, video distribution at scale, cloud computing, mobile computing, at that time, it was very clear that we could do learning at scale.

I launched a program called WebSim as part of, you know, our lab, where I said, if I could teach people how to do circuits labs fully online, then maybe this is something we could do at scale. On an average day, two to three hundred students from all over the world would come in and do WebSim, which is think of it as a MOOC lab. This was being built out in 2003, and this was massively open before a good mapping online questions existed — oh, it just was nine years before they existed in 2003.

Now you could maybe call it the MOOL, Massive Open Online Labs. It was circuits labs, and I figured if I could do online circuits labs, then maybe I could do other things. But then, I did nothing beyond that; I just kept running the WebSim. Then inspired by you, Sal, in 2011, a number of us at MIT got together—the leadership at MIT, Raphael Rife, myself, leadership Harvard, Alan Garver, and others began thinking about how could we take this to the next level.

And that's how edX was born. We said, look, rather than just having course materials as an OpenCourseWare, why don't we take complete courses with certificates, with communities, and also with labs and real hands-on experiences, and make that available to the world? That's when we launched MITx in December of 2011 and it turned into edX when we partnered, you know, with Harvard and MIT in early May and launched edX. That's how it all came together.

And what made you try to make this your life journey? You had a good gig as a professor, head of departments. What’s it like?

You know, it's one of those things where, as far as life's concerned, I'm very much an incrementalist. I don't plan very far out. I sort of take it one base at a time, you know, to use a baseball analogy. So, when the opportunity arose, we decided we'd create a MOOC platform, initially called MITx and then called edX. I was the head of the computer science and AI lab at MIT, and I had a 25-strong research group with PhD students and postdocs and undergrads. So we incubated what would become edX in the basement of CSAIL, the computer science and AI lab at MIT.

I was looking to recruit professors to teach the first course on edX, and when I told people, "Yeah, the course needs to start in two months, and oh by the way, we are building a completely new platform. We wanted an open-source platform that uses Python and Django and the latest technologies and not the old stuff, cloud-based," there was nothing available at that time!

So Professor, my colleagues said, "You're nuts! You can't do it!" By that time, we'd announced edX; there was no one ready to teach something. So, you know, we denounced it and I jumped in and said, you know, look, I'll launch with the circuits course. People looked up to me as if I was crazy: “Why would you launch with a circuits course as opposed to something like machine learning or Python programming?” I said, look, that's the best we can do! I could not get my colleagues to do this quickly enough; they felt we were nuts.

So that's how we launched with the circuits course. I kind of got sucked into it, having to do the first course myself with several of my colleagues: Jerry Sussman, Jacob White, Chris Turman, and others who really jumped in and pitched in to make it happen.

You know, it's a great story! And I do want to ask where this is all going. Where do you see the future? But I'm getting an unusual number of questions coming in through the message board, so I want to make sure I get to some of these.

This is just kind of a fun question from YouTube: Antonine Apollius. What's your favorite thing in math? Maybe I'll expand Antonine's question a little bit: What's your favorite thing in all of knowledge, in all of learning?

You know, I'll answer both questions. My favorite thing in math actually was when I really one day kind of just got it. I really figured out how to solve differential equations, which is that you try not to overthink it. You just kind of do it mechanically, and you apply templates and do it.

I actually remember in my circuits course we used first and second-order differential equations and solutions of those, and I called up my student who had this little gig going in California. I called him up and said, "Hey, look, your website doesn't have a video tutorial on first-order differential equations, so I need you to create something for me." This guy did it, and his name was Khan Academy.

So in my course on edX, I have a tutorial from Sal Khan, you know, on some of the stuff. But in terms of all of learning, I think the thing—my dream is one day to really understand quantum mechanics. I don't understand it! I've never really gotten it deeply. I can do the math and do some of the formal stuff, but my dream one day is to really have the time to sit down and listen to Feynman's lectures and others and really, really, truly rock it. You know, I'm not there yet.

You know, it's interesting, quantum mechanics—sorry—are you getting a little buzz every time I talk? Every time I start to talk, it sounds like there's a little—did you hear that?

No, no, I'm okay. I'm good, good, good.

Good! When you mention quantum mechanics, that is one of those things that I also want; I seek to have a deeper understanding of it. But I think anyone, the first time you learn it and you say, wow, all matter has both particle and wave-like properties, that the electron is a wave until you attempt to observe it, and then the wave function collapses and then you know its position. It's so counterintuitive to what we think is our understanding or our perception of the universe.

Actually, I probably had the most interesting insight that affects my understanding of quantum mechanics when I was reading this question and answer recently; this is with an Indian guru who died in the 1980s. His name is Sri Nisargadatta Maharaj, and he wrote this—actually, he didn't write this book. This book, I Am That, is based on conversations with him. People were always asking him these questions, and they keep asking him, "Well, what is the cause of this?" and "What is the cause of that?"

He says, "You are so obsessed with causality! What you call causality is just repeated correlation." Yeah! And I was like, is he right? Is it just repeated correlation? I'm still processing that, but that actually, in some ways, opened my mind to things like quantum mechanics, which would seem to defy our standard intuitions — whatever the intuition is of causality and correlation.

Yeah, I think that's right, absolutely! You know, I did have a really deep understanding of Newtonian mechanics and so on, and maybe that was part of the problem. I really deeply got it! And then when you have to completely rethink things and change your frame of reference, it almost became like one of the Yogi Berra quotes: “Wherever you go, there you are.” You said, "Yeah, I understand that!" But when you really think about it deeply, you say, "No, I don't got that. I have no idea what that means!"

The deeper I began thinking about quantum mechanics and things like that, the more I realized that, well, I could do the math and do calculations – the speed of light and, you know, the compression, you know, time and all of that stuff. I could do the math, but I never really, really internalized and got it. So my dream one day is to really, really get it!

Yeah, completely agree! And it's a—I mean, I could go on into this, but like, you know how much is based on perception. Well, anyway, we could go more into this. Another fun quote from that guru was when someone — you know, he said, you know, you ascribe reality to certain things and you ascribe imagination or dream-like to other things. But imagine if your dreams picked off where they left off the night before – then all of a sudden, you would have two realities!

Anyway, we're going way off! But quantum mechanics makes me think of that. I think it goes much better with the beard in your hand, actually. [Laughter]

Well, not a lot of kids are watching, Professor Agarwal! That's cool! You're a role model — be careful what you advocate for!

So, more questions and really good ones! Several from Badri Ratnam on YouTube. So one of the questions is, “Dr. Agarwal, India in the third world was hit hard by COVID. In one of your talks, you had mentioned that there's a plan to put the entire medical education curriculum online. How is that progressing?”

I did give a talk where we are talking to a number of universities where the idea is, you know, could we put an entire medical curriculum online? The same applies to computer science, data science, and business, and we've been sort of going at it one at a time.

We made huge inroads into computer science, data science, and business on edX. We have over 3000 courses and we're making great inroads in these areas. I would say healthcare is going to be next, where we do have a fair number of courses at this time. The idea would be, imagine if you could learn about, you know, topics like biochemistry, biology, epidemiology, pathology—lots of the theoretical aspects of these subjects. But then you go to college, and you learn, and you practice, and you work with real categories and all that stuff.

You really practice! And look at patients and work with patients and sick people; that would be the way to do it — sort of a blended model of learning where what you could learn online, you sort of learn online, and then you do all the practicum stuff when you come to a physical medical space. To me, that would be something that is very exciting and something that we would be interested in doing in the future.

But again, you know, we started with computer science, data science, and business where we've really built out a pretty deep curricula in those areas. It seems you and I have definitely talked about this in the past. The potential for— I mean, medicine is a big one, considering the shortage of healthcare workers around the world. But there are so many fields that I think, you know, between the edXes of the world and the Khan Academies of the world, I think we're actually just at the very early, early stages!

I have another question here, and this is really — you're clearly an educator reaching millions, so there's that hat. But there's also you used to be the head of the AI Lab at MIT, so you know a thing or two, I'm assuming, about that! You know, from YouTube, Saurav Saha asks, “What are your thoughts about the future of automation in different fields of technology?”

Then I'll add to the question: given your views of where automation and machine learning and robotics are going, where do you see education needing to go or maybe even leveraging some of those trends?

Oh my god, there's so much we could talk about here! Certainly, the computer science and AI lab at MIT is one of the centers for AI and machine learning research. A number of other universities and companies are also where incredible research is happening in AI and machine learning.

Much as computing and the internet changed the world in the late '80s, machine learning and AI and automation are completely transforming industries today and in the future. The downside is that people talk about roughly 50% of today's jobs will be transformed into other jobs that will require new sets of skills where a whole planet—half the planet—needs to be upskilling and reskilling. That's one of the areas that edX focuses on very heavily.

We offer professional certificates and MicroMasters and MicroBachelors and other credentials in these topic areas where people can come in and re-skill with short courses and programs in things like machine learning and AI. But certainly, we're seeing transformations in industries where in the past, I'll just give you a simple example. Let's say for example, in translation, when you speak, you know, usually humans would translate what you said into some other language at the U.N.

You've seen some of these—you've probably seen the movie Interpreter where, you know, Nicole Kidman was out there translating in real time what people were saying. Now imagine that today this technology is there where even as you speak, a machine could be translating you into one of any number of languages. Machine learning technology is key to things like this.

To your second question, Sal, I think the same thing—huge impact will happen in the learning field as well where first of all, let me start by saying the teacher is always going to be critical! I think people need to connect with people and get inspired, you know. They're looking for mentors and so on. I, too, had some incredible mentors in my school and college.

I'm hoping we can talk about some of them. But at the same time, machine learning and AI can take over many of the tasks. So, for example, I'll give you one example. One of the things we worked on at edX was a shifting framework, but imagine based on the solutions that people were developing for problems. Imagine that you could create a hinting framework where if a student got stuck in a problem, they'd say, “Okay, give me a hint,” and the computer gives them a hint based on all the solutions they've seen in the past.

There are so many things you could do! Another example is some amazing work that Berkeley did where they did research using edX learner data. Looking at the progression of learner activity, they could tell when a learner was going to drop out of a course. Using machine learning, based on that, you could actually create interventions and mentorships that happen just at the right time to enable someone to get more motivated and to stay in the course.

These are just two examples. Georgia Tech has done and IBM have done research in some edX courses where the teaching assistants in the discussion forums—the live teaching assistants have been experimenting, in some cases, replaced by AI technology, where some of the questions were answered again using machine learning as a basis for building up the repertoire of questions on which to train the machine learning model.

So I can see many, many areas where learning and teachers and universities and schools can benefit tremendously from AI and machine learning helping learners do a lot better.

Imagine to me one of the big ones is personalized learning, where imagine if every learner could have a personalized tutor that could teach them problems in certain areas at their own pace in a way that was completely customized to them. To me, that's one of the big holy grails of teaching and learning!

No, I couldn't agree with these notions of personalization! But I am curious, you know, how do you think this is going to affect—and I agree with you also that this will never replace the teacher! That human connection actually is probably going to become even more important over time as other things get automated. But what do you think this is going to do to labor markets?

Are there going to be, you know, there's kind of these techno-optimists who say, “Well, the industrial revolution, yes, a lot of jobs were lost, but a lot of jobs got created.” Do you think that's just going to be the same thing? We just have to have some optimistic faith that will happen? Or do we think this time is different because of the nature of the tasks that machine learning and robotics can take on?

If so, how do we get the positives while mitigating the negatives?

I think we're seeing some of this happen today where jobs get transformed into new jobs where, you know, just imagine that let's say, for example, you're driving a truck, and you know autonomous vehicles, self-driving cars are, you know, almost there. I think there's still some tweaking that needs to be done, but they're almost there.

You could imagine that you'll think, “Oh my goodness! You know, we'll never have people driving cars anymore!” For trucks that rumble from the west coast to the east coast or wherever in the world, you won't have drivers. But just imagine: you'd need other skills. You'd need people that maybe—you'd need an IT technician or an IT expert in the truck making sure that all the IT and support systems are working well, making sure that there are fail-safe models, and making sure that when you come to areas where the AI or machine learning doesn't quite cut it, you know, I can imagine if you're trying to drive a Vespa scooter on the small streets of Moradabad near Delhi in India.

I promise you, I promise you, you know, I don't think that you can have any machine learning technology do a self-driving car or something on the streets of Moradabad, where you know any random molecule or a particle or biological entity is coming from every direction. It's insane!

I think you'll need people for that, but I think you'll need more people who are generalists that can do driving as well as some other skills. So I do see that jobs will transform into newer jobs, which is why this whole concept of upskilling and reskilling is so important in marketing.

You know, a lot of the marketing we did earlier has now transformed to digital marketing. Analytics has taken off, and so today, you know, data science teams of companies have grown. They do machine learning, data science, statistics, and so on, and those teams are becoming bigger.

So even as certain kinds of teams are diminishing, you see the burgeoning of IT teams, you see the burgeoning of data science teams, and I think the same kind of trend will keep happening.

When we look at computing, when data centers went away to the cloud, people said, “Oh my god! You know, so many jobs will be lost!” But guess what? There's more jobs in computer science, more jobs in data centers. We need the ops people, we need support people that understand how to do load balancing and security and privacy and all of that stuff.

So really, jobs get transformed and you know, you need new kinds of skills or adjacent skills that people have to really become experts on.

It's interesting; one of the examples given to me recently which made a lot of sense was, you know, something like online commerce. We imagine, I mean, it has in some ways disrupted a lot of place-based retail, but it didn't remove—although maybe machine learning and robotics might change some of it—but still, someone is picking things, putting them in a basket.

Actually, it has created the new job of bringing it that last mile to someone's door, which wasn't there before because no one had the expectation that they would be able to click a few things on an app, and then their groceries show up at their door or, you know, the book they want to buy shows up there at their door within 24 hours.

It is going to be interesting! I hope that we create more jobs and actually that the jobs are even higher value. But there is this notion of technology being hyper-productive, which means that you can get the same output with less labor. So, you know, but my gut sense is we—edX and Khan Academy, we've got a role to play to make sure that most of the jobs are going to be higher up the value chain—more kind of knowledge economy jobs and so we need to make sure more people are prepared to be there in this world.

What’s interesting here, Sal, is that I actually see this maybe it goes back to quantum mechanics. You know, I see sort of a cyclic thing happening here where as we think we are climbing up the ladder and learning the newer and higher tech skills, what we're also seeing within enterprises is that more and more employees and companies are doubling down on what they call soft skills—really, it's better to call them power skills or human skills.

Empathy, teamwork, collaboration, creativity, design thinking, empathy—these are all becoming a lot more important where machines don't quite do that. So I think we will see kind of going back to the future approach, or maybe more like future to the past or whatever, where the human skills will be emphasized, whether in companies or in teaching and so on. The human skills will be emphasized, where people can do what they're good at, which is being human.

I couldn't agree more! I think there's something about technology, and as we get some of our material needs met by more and more productive technology, we realize that some of these core ideas that we've had for hundreds of thousands of years of human history—the community, being part of a village, being part of a clan, being part of a tribe—that we've gone through a first phase of technology separating us.

We see anxiety, depression, loneliness are really epidemics, disproportionately epidemic, it seems, in the western, more materially successful world. I hope you're right that as we, as more things get automated, we start realizing, hey, that frees us up to be the humans in the village and connect with each other emotionally.

People talk about healthcare! There's probably going to be a lot of jobs just to make sure people have companionship!

Make sure, you know, there's a lot of studies that the best doctors—who cares what the medicine they're giving is? Are they—does the patient feel like they're being heard? Do they feel like they're being cared for?

So, uh, and this goes back; this does go back to the quantum physics and Sri Nisargadatta Maharaj around perception is often reality. This feeling of connectedness, a feeling, is a real thing that we can't understand exactly.

Well, this got way more philosophical than I had expected. I thought we were just going to talk about democratizing education, which is a big idea in and of itself!

But, you know, you have a lot of folks, young people, people of all ages. You know, I mean, I'll let you take us home, Ana, Professor Agarwal. Thank you so much for joining us!

I'm excited to go on this journey with you because I think it could make a big dent in the universe!

Well, fantastic, Sal! Always great chatting with you. You've been an inspiration. I would not have been doing this had it not been for the work that you did. So just keep up all that you do, and hope we continue to collaborate and do great things in the future. Thank you so much!

Well, thank you! Thanks, everyone, for joining! As you could tell, very fun conversation! Myself and Anant go way, way, way back—probably before he remembers it! I don't remember all my professors, but I definitely remember him!

I had no idea that our lives would converge in this way, but you know, every few months we get together when one of us is in, when he's either in the Bay Area or I'm in Boston, and we end up spending hours chatting about a million different things on a million different dimensions!

So, really, really interesting guy, and edX, I think, is doing some incredibly powerful work, and it really picks up where, to some degree, Khan Academy leaves off! So I think between these types of organizations and others, there's a real chance to provide really rich resources for learners and for educators around the world.

But thank you so much for joining—thank you for being fired! I'm sorry for all of y'all I couldn't get to all of your questions! I will try to be better about that next time! But thanks for joining us on this live stream! Onward!