[00:00:00] Hala Taha: Young and profiters, welcome to the show. And if you're an avid listener of young and profiting podcasts, you've probably noticed a lot more AI-centered content. And that's by design because AI has taken us by storm. If we want to be young and profiting, we need to learn how to adapt to AI, we need to be prepared and understand the trends that are happening with AI, and we need to learn how to embrace AI in the now.

Today we're going to be talking about how AI is impacting the future of the education space and the workplace. And our guest today is Sal Khan. Sal Khan is the CEO and founder of the Khan Academy, a free online education platform, as well as a huge YouTube channel with over 8 million subscribers. Khan Academy, if you're a former student, you've probably interacted with it.

I know for a fact that the Khan Academy helped me pass calculus in college. So shout out to the Khan Academy and all the work that they do to help students all around the world. And he actually started Khan Academy as a side hustle, which I thought was so interesting and also as a nonprofit. Even to this day, it's generating millions of dollars a year.

It's still a nonprofit. So I was curious about that. We're going to be learning about his side hustle story and how he ended up taking Khan Academy full-time from a hobby to then a full-fledged, very successful nonprofit business. We're going to ask him why he decided to create a nonprofit and not a for-profit business and some of the myths around running a nonprofit.

And then we're going to talk about AI and its impact on education. He's got a new AI bot called Con Amigo on Khan Academy. We're going to learn how that's helping students not replace doing the work, but actually supporting their work. So without further ado, Sal, welcome to Young and Profiting Podcast 

[00:03:17] Sal Khan: Thanks for having me. 

[00:03:18] Hala Taha: I was doing research about you and I was surprised to find out because we all know you as the CEO and founder of the Khan Academy, which I actually used since college. So I've been a user of the Khan Academy for a really long time. And so we know you from this organization, but I found out that you actually were interested in AI early on, and you were thinking about becoming an AI researcher, so I thought we could start there.

Can you first tell us when you first got exposed to AI and how you got interested in the topic? 

[00:03:49] Sal Khan: That's kind of a deep track that you already hit on. You know, it's funny, I'm writing a book called Brave New Words. And it's the very end that I kind of make a confession. I obviously talk a lot about AI and education and I say at the end, it's like, Hey, I actually wanted to be an AI researcher at one time.

So you go all the way back to college. I was a computer science major at MIT and MIT for a long time has been one of the hubs of AI research. And I intentionally sought out Henry Patrick Winston, who'd wrote The textbook about AI to be my freshman advisor. And I took all of the AI courses, et cetera. And I think for someone who is excited about creating data from Star Trek, the character that, you know, sentient robots, the mid to late nineties was not actually a super exciting time in AI, the state of the art wasn't really looking like we were going to get into the realm of science fiction anytime soon.

So. I decided to do other things. I obviously still was interested in software. I've always been interested in education at the same time. So my path took a different turn. 

[00:04:53] Hala Taha: I love what I call organic entrepreneurship stories. So this little seed of an idea that then evolves into some amazing business, as opposed to starting with a business plan and raising capital, you sort of slowly over time create something amazing. And you did that with the Khan Academy. So can you talk to us about how you started your early career in corporate and then found yourself dabbling in tutoring as a hobby?

Which led you to create the amazing Khan Academy that we know it today. Can you tell us that story

[00:05:22] Sal Khan: Yeah, it's interesting. And it almost takes off right from where our last question was, which is Late nineties, 98, I graduated from college and the tech boom was happening. I found out how much you could make as a first year grad, as a software engineer or product manager. So I move out to the Bay area and I do that.

I ended up going at a couple of startups, this, that, and then the NASDAQ collapses. And I remember when that happened, I told myself, because the startup I was at, we'd kind of grown to 40 or 50 people. Then it becomes like survivor. You're just like, who gets laid off on any given day or any given week. And I told myself, you know what?

Entrepreneurship just is not for me. I do not have the emotional fortitude for this roller coaster ride. And I said, well, I'm just going to go to business school and figure out what to do with my life. So I go to business school while I'm there. I just find, Hey, finance is actually pretty interesting.

It's kind of a good combination of quantitative, but it has a huge human element to it. I ended up working at a small hedge fund. It's one of these things, even though I told myself I wasn't an entrepreneur, that first job at that hedge fund in 2003, it was at a one person hedge fund, so, I didn't admit it to it, but I kept gravitating towards these smaller things.

But it was while I was doing that, it was a year out of business school. I had just gotten married 2004 and my family was visiting me from New Orleans, which is where I was born and raised. And it just came out of conversation that my 12 year old cousin, Nadia, was having trouble with math. So I offered to tutor her when she goes back to New Orleans, she agrees.

It works out. I eventually. Got her caught up with her class a little ahead of her class at that point I became what I call a tiger cousin and I call up her school and I say, you know I really think Nadia Rehman should be able to retake that placement exam from last year They said who are you and I'm her cousin and they let her and that same Nadia who?

Initially was put into a slower track was put into a more advanced track Word spreads in the family that free tutoring is going on before I know it. I'm tutoring 10 15 cousins Family friends all over the country. Once again, after my day job at the hedge fund, and by this point, we had moved out to Northern California.

A lot of my friends from business school, like, how are you going to monetize this? What's the business plan? How am I? No, no, no, no, no. This is not a business. This is just, this is a passion project. I'm doing it for my family. I did start to make tools for them. I did start to write software for them so that they can get practice.

And so that I, as their tutor could keep track of what they were doing. My friends would say, what's the business plan? I was like, no business plan. This is just a family project, but it looks suspicious to them because it had software involved and it could scale. And then a friend suggested that I make videos for my cousins to scale even more.

And my cousins famously told me they liked me better on YouTube than in person. What they were saying, I think it was that they enjoyed, you know, Having an on demand version. They didn't have to feel embarrassed. They could watch it whenever they want. Pause, repeat, double speed, half speed. They still appreciated, I think having me involved in their lives, but then you could imagine other people started to discover the software, the videos, and it started to be at first hundreds of people, thousands of people.

I still told all my friends, no, this is still just a passion project. I still have my day job, which I like, but by 2008, there were about 50 to 100, 000 folks using it on a monthly basis. Which felt like a large number back then. It's obviously small relative to where Khan Academy is now, but I had frankly, trouble focusing on my day job.

Even then I resisted first child had just been born. So I did set it up as a nonprofit Khan Academy mission free world class education for anyone, anywhere. But it took about a year before I finally. Took the leap, which is never an easy thing to do. 

[00:09:07] Hala Taha: It's so interesting. And I'm smiling ear to ear because it's such an inspiring story.

And like I said, it happens so organically and obviously what you were doing was so needed in the world because so many people gravitated to these videos. I remember finding those videos when I needed homework, help and things like that. But nonprofit organizations, there's a lot of misconceptions and myths around this.

And to me, it's really interesting that over the years you've stayed. a non profit when it's so clear that you could have become a for profit company. So can you talk to us about some of the myths and misconceptions? Like, for example, people think a non profit, they're not in the business of necessarily making money or that it's only volunteers that work at a non profit.

Can you talk to us about some of those myths? 

 Yeah, and there's a deep irony here that I even started a non profit. When I was in business school, in business school, they really didn't fail anyone out, but they'll tell you privately that you kind of would have failed this class. And the one class that I.

[00:10:03] Sal Khan: Fell into that category was the class called social entrepreneurship. And I was skeptical in my business school days of what some of these nonprofits would do. It felt like they were more around organizing events than actually curing the disease or solving this or that. Obviously I go work at a hedge fund, which is very for profit, but I noticed a pattern when I talked to these, for the most part, public corporations is that your capital structure and your board dictate everything.

You could have a founder who's very mission driven, but as soon as they don't have full control, and as soon as you have to start dealing with hedge fund analysts like me, who are holding you to the fire on your quarterly earnings, things start to change a little bit. And so when Khan Academy was this thing, my cousins found it valuable, but even more, I got letters from folks all over the world saying how it was helping them.

I kind of started to imagine, well, what if this thing could one day reach tens or hundreds of millions, one day reach billions of folks? Give access to anyone. And this should exist for a very long time. And in the for profit world, you really didn't see any organizations that stay true to a mission beyond their founders.

And so I said, well, what's the structure I should do? And we'll maybe nonprofit. And the main differentiators for a nonprofit are that a for profit has owners, the shareholders. And the board has a fiduciary responsibility to those shareholders to maximize shareholder value. That's their bottom line. A non profit has no owners.

It's considered a public good. So you own as much of Khan Academy as I own of Khan Academy. We both, neither of us own Khan Academy. And the bottom line, the fiduciary responsibility of the board is the mission. In our case, free world class education for anyone, anywhere. And so for me at the time, I said, well, the not for profit structure seems to meet this goal of how does it stay true to its mission, regardless of whether I'm there or not, and it was delusional.

I was just one guy doing this thing. But I was like, what if this does last 50 years or a hundred years? How can I stay, true to that? And the not for profit structure seemed to make sense. Now it opens up a whole other series of questions. How do you raise capital? How do you attract talent, et cetera.

But that's all that defines a nonprofit. Obviously our mission has the word free in it, but lots of nonprofits charge plenty. Uh, most universities are nonprofits, but they're happy to charge you 50, 000 a year. But that money isn't going to any one individual or owners. It's going to the institution itself.

That was my thinking behind it. 

[00:12:38] Hala Taha: Another misconception is that not for profits don't make a lot of money. It's public information. So I saw that you guys in 2022 made 53 million or something like that in revenue. But then you also spent 58 million. And I also always hear that. You want to have a low overhead, not for profit, but that's not actually the case because you need to spend money to make an impact, right?

So can you talk to us about that? 

[00:13:03] Sal Khan: It's different ways of accounting for overhead and in some ways we are a bit of a bizarre non profit. Because your classic non profit would be, hey, we want to get mosquito nets to people in Africa. So we're going to raise money, let's say in places like the United States.

And for every dollar we raise 90 cents of that is going to buy mosquito nets. And then 10 cents of that is for overhead for the overhead of the organization. We're very different in that we raise money. So most of that money, that budget, the year that you looked at it was in the fifties, now it's even higher than that.

Approaching 60 million, 70 million. The great majority of that it's philanthropic donations of all sizes. Sometimes people donating as little as 10 a month. And obviously we have some very large donors. Who give millions of dollars a year, but we aren't just transferring that money from one party to another because of how we operate, we essentially are a tech organization, our way of attempting to give free world class education for anyone, anywhere is to build tools that cost a lot to build.

To your point costs, 50, 60, 70 million a year to build and operate. But the incremental cost for the next person is pretty close to zero. That's one of the beauties of technology. And for that budget of 50, 60, 70 million a year, which is about the budget of a large high school in the United States, we have 160 million something registered users.

We reach a large chunk of humanity. We're in 50 plus languages. And so in some ways you can say that most of that costs, People, it's software engineers, it's designers, it's content people, it's people going out there to partner with schools. So one way you could think about it is it's all going towards the mission because we're building things that then scale.

And we do have some earned sources of revenue where, uh, say a district wants to get support and training and things like that. And that costs a lot of money. That's a good bit on a marginal cost basis. So then we do have to charge some nominal amount just to be able to cover our costs there. 

[00:15:05] Hala Taha: You know, I think a lot of us dream about quitting our nine to five job and doing good for the world, just like you did. And I bet you a lot of listeners right now are in corporate or have day jobs and hoping to become an entrepreneur. And this might've gotten them curious about starting a non for profit business.

So can you help us understand some considerations and questions we should ask ourselves before we go off and start our own nonprofit organization and become entrepreneurs 

[00:15:33] Sal Khan: you really have to dig deep and decide what you want to do with your life and how you want to live it.

For me Even before Khan Academy existed, I told myself, and I think this was true. It's like, look, you know, I definitely wanted financial security, but when I really thought about it, if someone transferred a billion dollars into my bank account tomorrow, I wouldn't fight it, but I don't think I would live that differently.

I feel like the life that I was already leading, where, you know, I live in a 2, 500 square foot house, my family healthy, knock on wood and happy. We got two cars in the garage. We can go to dinner every now and then and take vacations. I feel like I'm living the life. I don't really need more than that. And so for me, the optimization is, can I spend my day working on something that I'm really passionate about?

And it doesn't have to be a not for profit. In fact, I think there's many times where a for profit route does make more sense. But for what I was doing and how I wanted to do it, the not for profit route made sense. And even that was a huge, huge risk. I think whenever you're doing anything entrepreneurial, you have to start with some delusional optimism.

And then when you take that leap and you quit your job, and you know, I just told you, it took me many years before I quit my job. It took me many years of really playing this project out. You usually realize that the world does not recognize the brilliance of your endeavor. As much as you did. So there was a tough time of eight, nine, 10 months where we were living off of savings.

As I mentioned, our first child had just been born where I was questioning whether I had done right thing. And I, and I would say on some level, not for profit is harder, especially if you live in Silicon Valley, like I do, because no one really saw a pattern there. They didn't see someone start a not for profit.

And as hard as it is to start a successful, let's say tech company where I forgot what the stats are, but something like 80 or 90 percent fail. I think you probably see even harsher stats for the not for profit realm, because there's an active venture capital community that is looking for the next big thing.

And they're willing to take flyers on startups, the philanthropic community. It takes a little bit more time to get that credibility. So I would say really figure out what you want to do for profit or nonprofit. I would say, get close to the, Problem you're trying to solve, maybe even try to solve it.

That's what I was doing with my cousins. Once again, whether for profit or nonprofit, when you see that it's actually solving a problem, then you know, you're creating value in the world. Hopefully. And then think about how you're going to scale it and then think about what you need to scale it and go down the appropriate path, whether for profit or nonprofit.

[00:18:13] Hala Taha: I love it. Such practical advice,

So I remember using Khan Academy in college, and it's been such a helpful resource for myself as well as millions of students, and you're truly bringing equity to education. And I know one of your goals is to make education accessible to everyone. How do you see AI helping to bring more equity to education

[00:18:34] Sal Khan: I think all people, especially when we're young and idealistic, We want the world to be a better place. And I did that when I was young and I always used to think about, you know, I'm thinking about myself when I'm in middle school and high school and college. And you see all these problems in the world, whether it is inequity of opportunity, whether it's the climate, whether it is wars, famine.

When you really peel the onion on almost every one of these problems in the world, it really boils down to what's going on in people's minds. It really boils down to some form of education. At the same time, I've read a lot of science fiction, and I remember one of my favorite is the Foundation series by Isaac Asimov.

And in this, The protagonist wants to essentially shorten the Dark Ages. It's character Harry Seldon who's able to predict large scale historical movements. And he sees that the Galactic Empire is going to enter a 10, 000 year Dark Ages. 10, 000 years of war and famine and lost knowledge. And he decides to do something about it by collecting the world's or the galaxy's knowledge on the periphery to shorten that Dark Ages by 9, 000 years.

And when I first read that in 7th grade, I realized you can look at a country or a civilization based on how tall their buildings are, or how clean their streets are, or the pillars in front of their Supreme Court, or whatever it might be. But what really defines a civilization is what's going on in people's minds.

If you take a bunch of people who are enlightened, who are well educated, and you give them very little in resources They will start a great society if you do the opposite and you give people all the pillars and the material wealth, but they don't know how to use it, it'll all, you know, over time deteriorate.

So I've always viewed education as that core issue. I'm also the beneficiary of it. You know, my family immigrated to the United States in the early seventies. I was lucky to go to a pretty good public school system in Louisiana and that's why I'm here. That's why I had the opportunity. But I know there are folks.

That look a lot like me, probably related to me, probably third or fourth cousins who are sitting in some village in India someplace, or in Bangladesh someplace, who don't have access to an education, and their potential is going to be squandered. And so, yeah, this idea of human potential has always fascinated me.

I don't want to preach what your framework is for the afterlife. If, just for fun, you say, maybe reincarnation does happen, what can you do in this life? That maximizes the benefit for your next life. I think education is one of them because you or I, if reincarnation happens, are much more likely to be born in a village in rural India or Africa than we are in Silicon Valley or New York City.

 So I found out in the summer of 2022 you got a life-changing email from the CEO and founder of OpenAI, which is the creators of Chad GBT, and it kicked off a whole slew of events. That came next. So can you share that story with us

[00:21:46] Sal Khan: I detail this in this book that's coming out, Brave New Words.

You know, I had been keeping track of the AI world just because I'm a bit of a nerd. And I said, maybe it'll have some kind of implication on what happens at Khan Academy. But Sam Altman, Greg Brockman reached out, as you mentioned, summer 2022, and they said, we're doing our next model. And they wanted, they wanted to meet with me and our chief learning officer about it.

I'm like, Oh yeah, sounds interesting. And so they said, look, we think this next model is going to excite and also maybe unnerve people because it's so powerful. And because of that, they want to launch with organizations that people trust doing socially positive things that have the technical capability to take advantage of the technology.

And by their own words, Khan Academy was the first organization that they thought of. And I was skeptical. I had seen GPT 2, GPT 3 really cool, but it was kind of nonsensical when you really paid attention to what it was doing. But when they showed me what GPT 4 and no one knew about GPT 4 at this point, no one knew about ChatGPT, which wasn't even built on GPT 4, that was built on GPT 3.

5. This was months before any of that happened. But when I saw, and our team saw, that, wow, this thing seems to actually be able to make sense of things, that you could actually have a conversation with it, that it could actually take on personas. If you think about the whole journey of Khan Academy, from me tutoring Nadia until, let's call it the summer of 2022, it was all about trying to scale, using the technologies that existed, that one to one personalization experience.

That I was able to do with Nadia and obviously using on demand video or personalized software, you can only kind of approximate it. But now with and what looked like what GPT 4 was capable of, it got that much further. It didn't take us long before we were able to have tutoring interactions with it that were almost indistinguishable from when Nadia and I were instant messaging each other in 2004.

And so we were excited. We said, Hey, this could work. Could just take it that much further. And then we realized it's not just about tutoring students. It could be about helping teachers with lesson planning and grading and progress reports. And then there could be a whole other class of things that we never even thought of before.

You know, we can have new question types that allow more open ended responses. We could have simulations being run by the AI. So. This was one of those moments that I kind of felt that, okay, this changes everything, but we have to really double or triple down on it. 

[00:24:12] Hala Taha: So when AI first came out, I know that a lot of educators were afraid that there was going to be a lot of cheating going on.

Can you talk to us about how people first perceived ChatGBT and maybe the ways that you thought that it could actually help the education world? 

A lot of people are fearful of AI replacing work and that it's going to enable a lot of cheating and things like that. What would you say to the naysayers

[00:24:38] Sal Khan: You can imagine when we first got access to GPT 4, this is. August of 2022. We were very excited about what it could do, but it was pretty obvious that it could have some negative use cases.

It could be used for cheating. We knew that it had some rough spots. It could make up things. You know, people are calling that hallucinations or fabrications by the AI. It wasn't particularly good at math, which is an issue if you're going to make it tutor. So we had to figure out ways to do that. And obviously we were concerned about under 18 users having inappropriate conversations with it.

How do you handle all these things? So we were under a non disclosure agreement with open AI at the time. We couldn't talk about it publicly, but we were having internal debates about, okay, we can't run away from this technology. It's too powerful and it can have so much positive. So let's put guardrails.

Let's make an AI that doesn't cheat, that provides transparency to teachers and parents that can actively moderate conversations so that it doesn't allow a student to get into something that's. unproductive, uh, that it protects student data privacy that can really feel like a real tutor. So we were working on all of these things in secret.

We were going to launch this on March of 2023 to coincide with the GPT 4 launch. Now in November, At the end, I think it was the last day of November, 2022, chat, GPT comes out and the whole world kind of explodes because of that, and I was pretty bummed. I immediately slacked Greg Brockman, who's the president at OpenAI, and I said, Hey Greg, what's going on here?

You have us under this NDA, we can't talk about anything, and you just launched something. And Greg says, look, we didn't launch anything. We just put a chat interface on top of GPT-3 0.5, which had been out for six or seven months at that point. And the whole world all of a sudden took note. And the reason why I was bummed is this was a thin layer on top of an AI that was not built for education, had none of the guardrails that we were thinking about.

And so it was no surprise that the big headlines in December and January of that year were, this is a cheating tool. We saw, I started seeing school districts banning chat GPT and I was like, Oh no, this is horrible. Like they're going to throw out the baby with the bathwater. The good news, which I couldn't have predicted, was by March when we came out, most of the education system had, to some degree, come full circle.

They had said, look, this technology is going to be a part of students future, and it's very powerful, so students should learn how to use it. But, ChatGPT wasn't built for an education use case. If only someone created a tool that had proper guardrails, proper transparency, proper safety and security, built for education, can't be used as a teaching tool, but can support students, that would be awesome.

And then we were able to come with Conmigo leveraging a more powerful AI, leveraging GPT 4, doing all of that. And so in some ways the chat GPT scare primed the world in a good way to be very ready for what we call ConMigo, which is the AI tutor and teaching assistant on Khan Academy. 

[00:27:42] Hala Taha: So, uh, really interested in ConMigo because I know there's, like you said, the tutoring feature.

Can you talk to us about why tutoring is so powerful for people to learn and why people who get tutoring end up doing a lot better? 

[00:27:57] Sal Khan: You know, in some ways, this is not new news. The oldest best practice for education is to have more personalization, more attention. If we go back 2, 300 years and we go back to Alexander the Great's time, Alexander the Great had a pretty good personal tutor.

He had Aristotle as his personal tutor. And, what Aristotle was probably able to do with young Alexander is if Alexander was having trouble with, let's say, understanding finance, which you needed to do if you were going to be the future emperor, I'm sure Aristotle would have slowed down and make sure that Alexander mastered it.

Or if young Alexander was just really good at military strategy, then the tutor could speed up or give more challenging exercises to work on. And that's always been the. Goal standard for education, but it's always been very resource intensive. So the only people who got that were the wealthy young princelings, future emperors, et cetera.

You fast forward to about 300 years ago. We had a very utopian idea as a society, which is free, mass, public education. But there was no way of giving everyone an Aristotle, so instead we borrowed ideas from the Industrial Revolution, which was happening at the same time, where, okay, let's batch students together, usually in groups of 25, 30, 35.

Let's move them all together at a set pace, aligned to some standards. Let's have the teacher there deliver the instruction, assess periodically, and some of the students are going to do pretty well. They are going to go into what we would now consider knowledge economy jobs. Some of the kids are going to be in the middle.

Yeah. They could be the managers at the factories of the industrial revolution or, or the shopkeepers. And then the students who are really struggling. Yeah. Well, that's okay. They can be the laborers and. The less skilled laborers in some of the factories, et cetera, et cetera. this model of education that you and I grew up in that we've taken for granted, it's not been what education has always been, but it has, I have to give credit for it.

It is what allowed us to give education to everyone. I would have not gotten access to education if not for public education. So it's done many, many, many good things, but now. It's not okay to have a world where only a few, let's say 20 percent of, or 30 percent can participate in the knowledge economy.

We need a world where everyone is able to, or as many people as possible are able to master concepts, get that level of personalization. The good news is that's where technology comes in. Maybe there's ways to scale that. You know, there was a education researcher, Benjamin Bloom, who wrote a study about the benefits of tutoring back in 1984, but he framed it as the two Sigma problem.

Two Sigma refers to a two standard deviation improvement that he thought tutoring could give. But he called it a problem because there's no way you can afford to do this for everyone. The power now is. You can still have the economics of whatever they are in, say, a public school where there's 25 kids in a room.

I would argue the lower the ratio you can get, the better, but that's a public policy budgetary question. But whatever ratio that is, can we now use technology to give more of that personalization that, I was able to do for Nadia or that Aristotle was able to do for Alexander the Great. 

[00:31:08] Hala Taha: A lot of people think AI is this thing that's out in the future, and that ChatGPT is really the only accessible AI tool right now. But a lot of companies, like yourself, are embracing AI. And you guys actually released Conmigo as a part of Khan Academy. It's basically an AI bot that students can use with and engage with while they're learning.

Can you tell us some of the things that Conmigo can do today?

[00:31:32] Sal Khan: From the very get go of Khan Academy, people said, okay, it's great that Khan Academy gives these adaptive exercises, allowing personalization, like a tutor. It's great that you have these on demand videos. If a student has trouble understanding a concept, but what if a student has a question? What if they don't understand a term?

What if they want to understand how something connects to something else? And we kind of were at a loss. I mean, my best answer was, well, use Khan Academy, ideally in a supported environment. And if a student watches the video, looks at the solution and still has trouble, maybe they ask a peer, maybe that's a good time for them to raise their hand.

And the teacher can do a more focused intervention with that student. And that is how it's been used. And it's, we've had really good efficacy studies there, but we said, well, now the AI could be there on top of the video, on top of the exercise to answer folks questions. We then also realized, but it could do more than that.

What if it could act as a simulation of a historical character or a literary character? So it can really act as a, a way to engage and immerse students in something. And then the more you're engaged, the deeper you're going to understand it. What if students can have debates with the AI where the AI takes one side of the issue and the student takes the other.

So we started realizing, in fact, even today, every hour we spend more brainstorming, we realize that we're being too narrow in our thinking. That this can really do things that I wasn't willing to do for Nadia. I wasn't willing to pretend to be Harriet Tubman for a day and do a simulation with her, but that's not possible.

And we're also doing things like writing stories where they, I won't do it for the student, but it can support the student in their writing. We're working on methods for a teacher to be able to work with the AI to develop, say, an assignment, say, a writing assignment, create the rubric, then assign it through the AI, and the AI won't cheat, it won't do it for it, but it can act as a writing coach, giving feedback, highlighting parts of it, and then when the student feels good about it, the AI can submit it to the teacher, and not only will that student be more supported, the teacher will be more supported, and actually they can feel more confident with it.

That the student didn't cheat because the AI can not just give the paper, it can give the entire process. Yeah, we spent four hours working on it together. Here's the whole transcript. And by the way, out of your class of 30, 20 of your students had trouble with a thesis statement. And here's what I recommend teacher that we can do to help them on this.

And here's a preliminary grade. So we think it can really stream support students better in a whole. It's not just in math, not just in science. It could be in writing. It could be in reading comprehension and also hopefully saving teachers some time as well. 

[00:34:03] Hala Taha: This is really interesting because a lot of the times when we're thinking about AI, we're thinking about how it replaces something.

But this is really all about how it's supporting students. How do you think that this could also translate in the workplace or in the private sector? 

[00:34:19] Sal Khan: It is a big, interesting question on what AI is going to do to, to the labor force broadly. I think The meme that has been going around over the last year, year and a half has been, you won't get replaced by an AI.

You're going to get replaced by a human using an AI. And so I think the imperative is, is that almost in any industry, if you learn to leverage these tools to be more productive, you're going to be in a good place and maybe be more productive in more domains as well. I think that's where the education system needs to make sure that students leverage these tools.

One, to enhance their own learning. At the end of the day, People say how, um, AI can do writing well, how it can do software engineering. Well, the reality is you're still going to need people to be able to put those pieces together. So instead of being the person writing the basic code, you're going to be more of the software architect or the project manager, instead of being the entry level writer, the world is going to need more editors, more people who can.

Put things together, but no one wants an editor or a software architect who can't write or code as well as the junior writers or the junior software engineers. So I think it's still an imperative for people to learn their traditional academic skills. It's been fact maybe better than in the past and maybe the AI can help there and then be able to.

Leverage these tools in whatever they're actually doing. 

[00:35:42] Hala Taha: So I know one thing that you talked about in your book is how AI can potentially supercharge human creativity. Can you talk to us about some of the ways that you imagine it can do that? 

[00:35:52] Sal Khan: This is the other fear that folks have. I mean, I could go on to any of these generative AI and say, Hey, write a screenplay for me or create an art piece in the style of whatever, and it'll bam.

It was just there. And so everyone's afraid, like, Oh my God, this is the end of creativity. I'll say a couple of things. One, this isn't the first time in history, something like this has happened. And I write about this in my book, brave new words in the 19th century, when the camera came out, I am sure a lot of portrait artists said, Oh my God, this is cheating, this thing, you just press a button and it does essentially a real life picture of it.

But all of the artistry is gone. Now we know on one level that didn't happen, maybe the people hiring a portrait artist to paint a portrait of them, maybe that market has declined a little bit because of the camera, but it didn't get rid of creativity. In fact, a whole new field, a new creative field, not only existed, but it democratized art in some ways where more people could do artistic things.

I think you're going to see something very similar happening with AI. And the other thing I emphasize is creativity isn't a zero sum game. It's not that let's say you and I, let's consider ourselves creative people. Each of us by ourselves can be reasonably creative, but if you and I are able to chat about things and brainstorm together and riff together, I think we're each going to become more creative, not less.

I'm not just going to say, Oh, Hala has got good ideas. I'm just going to check out. Okay. I'm going to say, Oh, I love Hala's idea there. And well, what if we did this too? I think any of us who consider ourselves reasonably creative recognize that our most creative times in our life were when we were around other creative people.

I think AI is going to democratize that, where there could be a young girl in Afghanistan someplace, and she's not even allowed to go to school. But if she has access to this, she could brainstorm. She could riff ideas. She could test ideas. Now it'll be even better if it could be with the AI and other people around, but you might not have that.

And so I think AI is going to actually be an enhancer for creativity. I also think it's going to lower. I didn't even allow myself to think that I might be able to be become a filmmaker one day. And I was like, who gets to make a film? They cost tens, if not hundreds of millions of dollars. You got to know the right people, et cetera, et cetera.

But now AI is going to make that much more accessible where you can start to do movie quality. production for a hundredth or a thousandth of the cost that you might have before. Now, once again, I don't think it's going to put creatives out of work. It's going to allow more people to have creative expression, and that the good stuff is going to surface.

In fact, it's a lot like podcasts or YouTube. These were both democratizing. And yeah, they have in some ways threatened the traditional media establishment. But they've been good overall because there's a lot of creative people who couldn't break in to the traditional media establishment before, but now they can self publish on a podcast or self publish on YouTube and the world discovers them.

[00:38:54] Hala Taha: I want to read a quote from your book that's really related to this. You say, the best ideas will come not from the AI creating for us, but when the AI is creating and riffing with us. Much like poets hanging out in a cafe in Paris, humans and AI can augment each other and inspire a mutually creative process.

So I loved that. And I know that entrepreneurs also have to be super creative. What kind of opportunities do you think that AI is going to bring for entrepreneurs? 

[00:39:21] Sal Khan: I think if I wasn't looking to start it and I, you know, whenever I talked to young people, I I've got a lot of ideas for businesses that I don't have the time to work on.

AI is such a general and transformative technology. It's very tempting for people to get enamored with the technology itself and just want to make a better AI. I think that's going to be a game for about six or seven big players are going to be working on that, like hardcore R and D. But the real opportunity, I think, especially for new entrepreneurs is go into any field that you have an interest in and get to know that field well, and then think about how you could leverage things like generative AI to create products for that field.

And probably the further that field is away from technology, the more opportunity there will be. Let's just imagine I'm thinking of something that might, you know, it's very far removed from technology. Let's say that you are a 

[00:40:13] Hala Taha: farmer, 

[00:40:14] Sal Khan: a farmer, a farmer, right? If you understood all of the tasks that a farmer does, I'm sure a lot of it is around like, Oh, I got to order the seeds.

I got to look at the weather. I have to use the weather to figure out this. I have to then go and negotiate with these people for whatever my fuel costs. Okay, could you potentially, the same way that Conmigo is doing that for teachers, where we're helping them with lesson planning, we're helping them with grading papers, we're helping them with writing progress reports, are there tasks like that that farmers are spending 10 hours a week doing?

Automate that you'll be the only, I probably, I guess maybe more than one person might hear this idea, but it's not going to be a crowded space. And if you're out there in the next couple of years, creating the farming management system, using AI, you might get a lot of traction. It might even be more powerful.

You might even be able to use generative AI. Maybe the generative AI can read the newspaper or recommend to you, which crops to plant. So that it improves your yield and improves the market prices that you're able to get, or your ability to market into the distributors or whoever else. So, you know, you just threw out a random industry farming, but you can see that there's probably a lot there that it's not as obvious, but if you take these really powerful tools, there's, there's probably an opportunity.

So just get to know a space deeply. And start making stuff. And if you start to see that there's value being created and getting traction, then I think there's going to be people who want to invest in it. 

[00:41:42] Hala Taha: So I noticed that you've been saying a lot, get specialized in one area. Why do you feel so strongly about getting specialized in one area in preparation for the AI revolution?

[00:41:53] Sal Khan: I'll say it two ways. I would say, especially if you're an entrepreneur, I would say get to know an industry well, but not to necessarily be narrow in your skills. I think if anything, Get to know industry. Well, we'll allow you to make tools for that industry that actually create value for the people in the industry.

So that's why you should know that. But I do think AI is actually going to have the opposite effect on the skills that we need to develop. I think the time where you say, I am only a software tester. I am only a designer. I am only a technical writer. I think that's a very troubling mindset because the AI is going to allow you to do more and more of the pieces.

You know, I thought it was funny, you know, this whole screenwriters guild strike that happened many months ago in Hollywood, where they were negotiating with the studios to say, you are not allowed to use generative AI to write screenplays. I think the screenwriters guild had a too overly narrow view of it.

It should actually be the other way around because today, a really good screenwriter, screenplay writer has a great sense of story, creativity. They write a screenplay. And if they're lucky, someone's going to take it up and produce it. And they might get paid, I don't know, at best tens or hundreds of thousands of dollars for that screenplay.

And then a hundred million dollars are going to be spent on that movie. And then if it's successful, it's going to generate hundreds of millions or billions of dollars. The screenplay writer who actually created the story, they get a very small cut in the future, let's say five years, 10 years in the future, not that far.

If you are really good screenplay writers, you don't have to stop at the screenplay. You'll be able to create the whole movie. With the help of AI, in order to do that, though, like you're going to have the tools to be able to create the scenes, do that, but you have to be more of a generalist, you're gonna have to think about the business, the distribution, et cetera, et cetera, but all of the tools are going to be at your disposal.

So I would say, get to know an industry well, so that you can help people in that industry, but broaden your own skill set so that you can manage the AI to do more and more pieces of the puzzle. 

[00:43:53] Hala Taha: And are there any specific skills that you feel are going to become more and more important to have? 

[00:43:58] Sal Khan: Yeah, I've always been a little bit of a traditional, even when things like Google came out, a lot of people would say, Oh, well, kids don't have to know their facts anymore.

They don't even have to know math anymore because there's calculators. I've always disagreed with that because it's very useful to have that on demand in your mind and it allows you to form connections and see patterns in the world that if you don't know those facts, if you don't have those skills, if you can't.

Have that numeracy. You're just not going to see. And it also, even in the world of pre AI and the internet, someone with a knowledge of history is going to be able to make much better web searches about history than someone without a knowledge of history. And so I think the same thing is true now with AI for my own kids, for everyone I know, it's even more imperative that you should be able to not just check out and let the AI do the work, but you should be better than the AI so that you can manage the AI.

So the three R's, which has always bothered me because arithmetic does not start with an R, but reading, writing and arithmetic, I guess, plus things like oral communication, I think is going to be more and more important because people are not even going to know where that written communication came from, being able to manage relationships.

And this, you know, you mentioned about creativity. I think creativity is going to be King because in the past, creativity has been bottlenecked by gatekeepers. As we said, the internet has helped YouTube has helped podcasting has helped, but now AI is going to be the ultimate. opening of the gates where anyone is going to be able to find creative expression.

[00:45:28] Hala Taha: Something else that you wrote in your book is that AI could revolutionize the way that we actually hire people. Can you talk about how hiring can change? 

[00:45:36] Sal Khan: Yeah. And this is a sensitive topic because when people imagine dystopian scenarios for AIs, one of them, they imagine, Oh, people are going to use AIs to look at resumes.

And then what if the AIs have biases? And no one knows about it. All of a sudden the AIs are distorting the world and the European union. In fact, some of the early regulations have explicitly called out, like, Hey, don't use this for hiring, et cetera, et cetera. The point I make in the book is there's nothing that's going to be bias free.

And in fact, there are some biases you want. If I'm hiring someone, I have a bias towards people who are going to show up on time. I have a bias towards people who are going to be pleasant to work with, who aren't going to yell at people who are open minded, those are positive biases. Now, it should not bias on gender, or race, or religion, or age.

So those are the ones you have to watch out for. And what I argue in the book is, no AI system is going to be perfect. You shouldn't measure against perfection, you should measure against the status quo. And I can guarantee you that the human beings who are right now looking through stacks of resumes and interviewing people are incredibly biased.

You know, you can do bias training and all of that stuff, but at the end of the day, They've been studies where people looking at the same resume at 5 p. m. They have a different point of view when they, then when they look at it in the morning and they're fresh, what's interesting about AI is it can give more energy to the process than most human beings can, you know, most jobs at Khan Academy for every job opening, we get a hundred or 200 resumes.

Our screeners at best filter it down to about 20 people that they'll interview. When they're looking at those 200. And we have 30 open job openings. So they're having to sift through 3000. There's no way that they can spend more than tens of seconds per resume. So they're indexing on just a few keywords, this, that, and then moving on.

We're probably losing a lot of great talent. So if AI could one, make sure we're getting a deeper look at more of them. And then AI can also be audited in ways that a human being can't, I can't go to one of our hiring managers and kind of like, I don't even have the bandwidth to do the search. Where we give them 10 equal resumes, different genders, ages, ethnicities, religions, whatever, and see whether they are statistically equal through our process.

That's very hard to do, but you could do that with an AI. In fact, you could use AI to generate these statistically equivalent resumes. And then filter them in a, you know, the scale of thousands or millions to see how the AI evaluator evaluates them and says, okay, at least on that dimension, it does not seem to have a bias and for sure it's a better bias than what human beings have.

So I actually think this used well, it's going to be a chance to go deeper and richer, but the only way we're going to have a chance to use it if people don't hold it to too high of a standard. The standard really should be, is it better than the status quo and can we measure it? Because nothing's going to be perfect.

[00:48:36] Hala Taha: So interesting. Well, thank you so much, Sal. I always end my interviews with two last questions that I ask all my guests. You don't have to relate it to the topic of the episode. The first one is, what is one actionable thing our young improfiters can do today to become more profitable tomorrow? 

[00:48:52] Sal Khan: The one thing I would do is start understanding problems in the world that you might discover you can solve and then use whatever tools at your disposal.

We've talked a lot about AI. AI might be part of it. In fact, I suspect it will be part of it. But it could be something very easy too. It does not have to be technology for technology's sake. But if you know a space well and you start solving problems and you have traction with that, then I think you're off to the races.

[00:49:17] Hala Taha: And what is your secret to profiting in life? And this can go beyond business and financial. 

[00:49:22] Sal Khan: I think it's very important you change as you get older and you realize things, but audit what's important to you, what's truly important to you versus what you think is important to you, what your family is projecting onto you, what your friends are projecting onto you.

We all remember being in college and you have these late night conversations out in the hallway in your dorm room and everyone's saying, Oh, if I only had a million dollars or a billion dollars, I would do this, I would do that. And now I have friends who have actually made their tens of millions or hundreds of millions or billions of dollars.

Some of them are pursuing their passions, but a lot of them have forgotten what they got into it in the first place. Or they might not fully appreciate that at some point, once you have your financial security, that actually the most valuable thing. I mean, I remind myself this every day when I'm working extra hours or I'm like, wow, I have a limited amount of time on this planet.

I have a limited amount of time with my wife, with my kids. I have to embrace that. I have to really enjoy that. I have to. Put energy, at least as much energy as I put behind work, arguably more energy to make sure that I get the most out of that and they get the most out of it. So I think it's just, yeah, really auditing what you care about.

I'd argue in most cases, you know, not to be judgmental, but I think money matters. You need to get to a level of security, but once you get to that level of security, the number one things, or actually even before that level of security, the number one things are always. Your health and your relationships, the money can help give security to your health and relationships sometimes.

But after a certain point, that's what I think most people should be optimizing for. 

[00:50:59] Hala Taha: I think that's really good advice. And where can everybody learn more about you and everything that you do? 

[00:51:05] Sal Khan: I'm all over the internet. Obviously anyone could do a web search for Khan Academy. Find those resources. This is the second book.

The first one was One World Schoolhouse, this new book, brave, new words about generative AI and education and work coming out, people can do a search on that. And there's plenty of me on YouTube. 

[00:51:23] Hala Taha: Yeah. So he's really easy to find. We'll stick all the links in the show notes. Thank you so much for joining us on Young and Profiting Podcast.

Thanks for having 

[00:51:31] Sal Khan: me, Hala. 

[00:51:32] Hala Taha: I love speaking with entrepreneurs like Sal Khan, who have made such a big splash in the nonprofit space, because we don't hear about that often. And I never think about entrepreneurs being in a nonprofit, but it totally works. And you can totally have a huge innovative company. That makes a lot of money and you can make a great living by running a non profit while also doing something you feel passionate about.

And that's just so cool, so I love this conversation. And I also found Sal's own passion for human potential to be really contagious and it was fascinating to hear what it was like for him to be able to kick the tires on ChatGBT in the early days, before any of us had access. We tend to think of AI now as something that will help students by doing their work for them.

But Sal and his team saw AI, and they saw that it could revolutionize teaching by helping students, not doing the work for them, by helping them work through problems. Sal was also optimistic about how AI will help us unleash creativity across the board, among students and professionals alike. In the same way that podcasts, social media, and YouTube have democratized content and fueled a new generation of creatives, AI may supercharge that productivity, providing additional ways to succeed in the future.

Entrepreneurs could be extremely well placed to benefit from this revolution, according to Sal. After all, we entrepreneurs are by nature generalists and creatives. We have many different skill sets and we excel at putting resources together to create value. And the more you broaden your own skills and learn to incorporate AI in everything that you do, the better placed you'll be to ride this AI wave in the future rather than sink beneath it.

Thanks for listening to this episode of Young and Profiting Podcast. We'd love you to help this podcast ride into the future too. Every time you listen and enjoy an episode, please share it with your friends and family. Spread our podcast by word of mouth. And if you did enjoy this show and you learned something new, then please drop us a five star review on Apple Podcasts.

It just takes a couple minutes and it's the number one way to thank us. Apple Podcasts, Spotify, wherever you drop a review, I'd love to read it. And if you prefer to watch your podcast as videos, you can find us on YouTube. Just look up young and profiting and you'll find all of our episodes published on there.

You can also find me on Instagram or on LinkedIn by searching my name, it's Hala Taha. And I did also want to shout out my amazing production team. I am so grateful for all that you do. And today I want to give a special shout out to Maxi, our audio engineer. He's been doing such a great job. I appreciate all the hustle that he puts into his job.

So just thank you for all that you do, Maxi. And thanks to the whole team. I've got the best team in the world. This is your host, Hala Taha, AKA the Podcast Princess, signing off.