Alpha School 对未来教育的颠覆性构想

All educational content is obsolete. Every textbook, every lesson plan, every test—all of it is obsolete because GenAI is going to be able to deliver a personalized lesson just for you.

Patrick O’Shaughnessy

This is going to be so much fun. I have learned so much from you about what you’re building, and I’m extremely excited to share it with everybody else. I view my role today as the excited skeptic, which I feel like is the role of a lot of the parents considering Alpha School for their kids.

1. What Is Alpha School?

I want to go into deep detail about every aspect of what you and Mackenzie have built. But since most people don’t yet know what Alpha School is, I think you have to start the conversation by explaining what it is and why the two of you are building it together.

Alpha School is a high-end private school. It’s expanding across the country, but at its core, it’s redefining what school can be for kids.

Patrick O’Shaughnessy

Yeah.

I’m a product guy, and so I look at it from a product perspective: How can you rebuild school and make it awesome? We look at 5 things when thinking about how to build a great school. The first is that kids must love school.

Patrick O’Shaughnessy

That’s just a crazy thought, right? Most people don’t believe you can love school.

Our standard is that we want to build it so the kids want to go to school instead of going on vacation.

I’ve been principal now for 3 years. Mackenzie started it 10 years ago, and I came on 3 years ago. 2 months ago, our high school kids sent us a note and said, “Hey, can we not take summer break? We’re loving what we’re doing so much that we want to continue through it.” So it’s sort of a peak.

Patrick O’Shaughnessy

Sounds impossible, but I’ll grant it for now.

I was one of those kids who definitely did not like school. I dropped out of college, and in high school I got in trouble because I skipped school all the time. This is a different world for me.

The second part is learning. Kids need to learn 10 times faster, so we built a learning engine. With all the talk about AI and learning science, kids can learn 10 times faster. In our school, kids spend 2 hours a day with the AI tutor and with screens. We obviously consider it very good screen time, and they learn 10 times faster.

Our academic performance is the top of any school in the country, and we do it in dramatically less time. The key to that is, once you can crush your academics in 2 hours a day, the natural follow-on question is, what do you do the rest of the day? I can tell you, as principal, the 1 thing parents don’t want you to do is send their kids home after 2 hours.

So what do you do for the other 4 hours? That’s where we talk about life skills. In reimagining school, what are all the things that, as parents, we want beyond pure academics? Leadership, teamwork, entrepreneurship, financial literacy, relationship-building, socialization, hard work, grit, storytelling, public speaking—all of those kinds of things.

We do all of that in the rest of the day. Those are big, project-based workshops that are super fun for the kids and team-based, and they’re not screen time. They’re sort of the opposite.

Patrick O’Shaughnessy

The first thing you ever said to me was something like, “Educating kids is the single most important thing we do as a society.” It hasn’t changed in 200 years.

Let me dig into that a little bit more: why it hasn’t changed, how it began, and why it remains the way it is. It’s been the same for so long that, although we’re open to radical approaches in everything in the world, and everyone is excited about the pace of technological change, education is 1 place where I think parents, myself included, are scared. It feels like the ultimate risk to put your kid in something that’s different.

I want to talk about this 200-year legacy. Why was it like this? Why is it like this? And why doesn’t it change?

Every parent wants to educate their kid the way they were educated. I didn’t start the school. Mackenzie started the school 10 years ago. I went to Catholic school, and my kids were in Catholic school. She started Alpha, and it took her 2 years of saying, “Come to my school.”

I said, “No offense, Mackenzie. I’m not sending my kid to your weird, janky school.” It’s weird; it’s different. I didn’t want to do it. That’s the normal parent reaction to all of this because it’s all we know.

Going back 200 years, if you go back to Socrates, Plato, Aristotle, and Alexander the Great, the elite had tutors. But fundamentally, 200 years ago, when we decided everybody needed an education, we created the industrialized model. The way to do it was to put a teacher in front of a classroom.

Whatever the class size is—20, 30—that is actually the model globally. It doesn’t matter where you are in the world. You can be in a poor school district in India, and there’s a teacher in front of a classroom. The building’s not that nice. Or you could be down the street here at a $75,000 private school, and it’s a teacher in front of a classroom in a very nice building.

That’s how teaching has been done. What happened about 40 years ago is that the schools of education—Stanford was 1 of the leading ones, but there were lots of others—took the lead and asked, “How do kids learn and develop?” There’s actually a thing called learning science: the science of how children develop.

They started writing papers 40 years ago that basically said the teacher-in-front-of-the-classroom model was the worst way to teach kids. There are all these ways where kids can learn 2, 5, or 10 times faster, but it doesn’t work in the teacher-in-front-of-the-classroom model, and it’s not scalable. So you can’t do anything about it.

One of the first papers was Bloom’s 2 Sigma Problem, which is a seminal paper. It talks about how, if everybody had an individualized tutor, it would be great, but it obviously wasn’t feasible. There are a lot of other things, too: mastery-based learning, spaced repetition, and all these other concepts from cognitive-load theory. We can go into those, but fundamentally, it just hasn’t been possible.

That’s where I came into the Alpha story. I was an early AI guy—old-school AI, not the new neural networks. In the ’80s, I actually wrote a paper in high school on AI. One of my paragraphs was about neural networks, but they were decades away. Back then, it was all expert systems and ontologies.

Then I went to Stanford, and I was in a class with Ed Feigenbaum, Professor Feigenbaum, who was the father of expert systems. He talked about how these things would be worth millions of dollars if you ever got them to work.

A couple of my classmates and I dropped out and started Trilogy. Our claim to fame in the ’90s was that we were the first AI product to sell a billion dollars. That was old-school, old-school AI.

Then, 3 years ago, when GenAI came out, I thought, “Neural networks are here. They’re here. They’re decades ahead of where they were.” My kids were at Alpha because, 10 years ago, I eventually did send them there. It was a great story of how they got there, but it wasn’t scalable, just like all the other educational things.

I told Mackenzie, “You have a great school here. It’s great for my kids, your kids, and all kids who go to Alpha, but this can’t scale.” 3 years ago, with GenAI, I thought, “This is finally the technology that’s going to allow education to scale. We can take all this awesomeness—where kids can learn 2, 5, or 10 times better and faster—and get it out to a billion kids.”

That’s when I became principal. I thought, “Okay, Mackenzie, we could totally make this a super-scalable thing, but I’m a product guy. I’ve got to be in it. I’ve got to see what happens when fifth graders get in a fight, when parents yell at you, and all of those things.”

Patrick O’Shaughnessy

2. Two-Hour Learning

If we think about the school as having lots of components, I want to spend a chunk of time on just the 2 hours of learning—whether that should be 1 hour or 3 hours, and the other aspects of it—and then the afternoon session, the life skills, the motivation, and everything else.

Maybe we could start with the chapters that the 2-hour learning program has gone through. My understanding is that originally it was always computer-based, trying to make it personalized using external tools like IXL and others that people may know about, but then layering in a feedback mechanism—which I know is incredibly important to you in anything that you build. That tight feedback and iteration matters a lot.

Maybe talk us through the various iterations of the software itself—literally what it is, what it does, and what it feels like to use.

Khan Academy is probably the best-known example. Everybody knows that one. Homeschoolers know this concept well, too: most homeschoolers spend only a couple of hours a day with apps trying to teach.

The evolution was very much about using various apps and figuring out how to make them better. 10 years ago, DreamBox was 1 of the leading apps, and we used apps in order to have kids learn. The evolution has been: How do you make the apps better? How do you get better educational outcomes in less time?

There’s a little revisionist history, since I’ve only been here for 3 years, but fundamentally, the design goal our team uses is this: When we put in new software—or anything new—does it increase their academic knowledge or reduce the time it takes to learn?

That’s basically the question: How do I get knowledge into a kid’s head, not cram it in so they’re going to forget it, but make it truly change the schema of their brain? It’s going to change it, and it’s going to be there forever. How do I create thinking and knowledge acquisition? How do I do that in the most efficient way possible?

A lot of learning science says, “Here’s how you do that.” Since I took over, or became principal, and started building out this product, we’ve been asking, “What can you do? What are the principles of learning science that are the foundation for this, so that now a kid can learn and learn well?”

So how does it work? Let’s take a new kid. The first thing that apps are really good at is figuring out what you know and what you don’t know.

Yeah. Right. It's going to be able to ask you some questions. Think of any kind of assessment test; a MAPS test would be a good example. It's figuring out what you know and what you don't know. Once we figure that out, we give you lessons to build up your knowledge from wherever you are.

The first thing you have to understand—and this is very different—is that age grade and knowledge grade are very different things. Just because you're in sixth grade age-wise doesn't mean you should actually be in sixth-grade content. This is one of the big things that's different from a teacher-in-front-of-a-classroom model.

Right. A teacher has to deliver sixth-grade material. I'm a sixth-grade teacher; I'm doing sixth-grade lessons. Right?

The magic—and this is what the AI tutor can do—is say, “Look, I don't know what happened. It's okay. Whatever happened, you were on holiday, you missed those lessons, and you don't know this fourth-grade material.” So I'm just going to give you this fourth-grade lesson, even though age-wise you're in sixth grade.

The reason you've been struggling with sixth-grade material is because you didn't know this fourth-grade concept. If you're doing fractions and you haven't mastered division, you're going to sit there and say, “God, this fraction problem is really hard.” But the real issue is, well, just go back and learn your division, and then the fraction is going to be easier.

Most knowledge is hierarchical. It's based on foundations. Algebra is basically advanced fraction manipulation; fractions are multiplication and division. You can just keep going down the tree, where you have to learn bottom-up and have mastery.

One of the big concepts of learning science is that you need to master material, not just know parts of it and have big, gaping gaps.

What does mastery mean?

Mastery means—think of a sports analogy. Have you mastered the basics? How good are you? We always talk about how, in standard school, there's a whole set of things where 70% means you're passing. You don't know 30% of the material, and they move you on to more advanced things.

It's sort of like in basketball: if you're the point guard and you lose the ball 30% of the time going down the court, the coach isn't going to be like, “Hey, let's work on the alley-oop.” They're going to be like, “Okay, let's get back to the basics and master the basics so you can get down the court.”

So, basically, close to 100?

Close to 100. Our standard is over 90. One of the things that we do versus other apps—and what we had to surround the apps with—is that if you use existing apps at the mastery standard they have, you'll struggle to get the results we get at Alpha, because the mastery standard that's embedded in and used by all the apps is just too low.

Can we spend a minute on the learning science and the cognitive-behavioral stuff, cognitive load theory, and all this kind of stuff that underpins 2× less time, better content, and better scores? Maybe just dump on us the important aspects of those 2 fields that you've assimilated.

Sure. Let's go through them in order. If you take individual, personalized lesson plans, the first thing is that if you need a fourth-grade lesson, it's really important that you're given a fourth-grade lesson, even if you're in sixth grade.

This is like a one-on-one tutor.

It's like the Bloom's 2 Sigma tutor. The second part is that you do need mastery. Going back to fourth grade, go learn it. When kids are struggling as they get older, the real issue isn't that they're struggling with the older material; it's the prior material. So just put them into the prior material.

Then, as you get into cognitive load theory, here's the way to think about that: you have working-memory slots. Every person has a certain number of chunks that they can work on, or registers on a CPU, if you're a technical person. That's what you're doing when you're computing—i.e., thinking. What is 7 × 8? I'm calculating what that is.

The second part is how many reps it takes to take what's in working memory and put it into long-term memory so you remember it forever. There's a certain number, and people are actually different. GT kids need fewer reps than other kids. Fundamentally, it's how many reps you need to put it in.

There's a whole concept of spaced repetition: if you space the repetitions out over time, you retain it for longer. This is the difference between, “I cram,” versus, “Do you remember it?” We have a workshop with the kids where, if you actually take kids and say, “Okay, you're going to do vocabulary words for 10 minutes a day, every day. On Friday, here's your test score,” we have another group that volunteers and says, “No, no, no. I'm going to cram the night before.”

They cram the night before, take the test, and basically get about the same score. They're like, “See, cramming works.” We then retest them a week later, and the cram kids don't remember anything. The people who've done it for 10 minutes a day all remember it. The kids learn, “Oh, okay, I should actually not cram if I really want to learn this.”

That goes back to our whole thing of retention and learning faster. You can teach kids to be these self-driven learners where they learn these concepts and are actually excited to learn 2, 5, or 10 times faster. Kids want to learn as quickly as they can, assuming you give them their time back.

3. Academic Results & Efficiency

Just to put a fine point on it, maybe share some of your favorite outcome statistics. The school has been around for a while, and you've been there 3 years. Some of your favorite examples of what this method has allowed for in the kids? So far, we've been talking about a really radical, cool new way of educating kids, but you measure everything. In our conversations, I've come to appreciate that you're a measurement fanatic, and you care very deeply about the measurable outcomes here. What are some favorite examples?

Our academic results are just crazy based on standardized tests. There are 2 parts. The first is that our kids literally spend 2 hours a day, and that's it. Instead of sitting in class for 6 hours and doing homework, the amount of time they need to spend is so much less than what we spent or what parents expect.

The data you need is that the average kid at Alpha, if they're a year behind in a subject, is only 20 to 30 hours behind that. To a parent, that's like, “Oh my God, that's terrible.” But you can teach an entire grade level of material for 1 subject in about 20 to 30 hours.

Crazy, right?

You're just like, wait, there's 180 school days, and each class is an hour, so it's 180 hours plus homework, and you're telling me it's 20 to 30 hours? And you're like, yeah, 5 to 10 times faster, over and over.

Our AP kids will do AP World History, and all of our kids got 5s because the test came back a couple of weeks ago. They all got 5s, and they spent 75 hours max to do AP. This whole idea of these big AP loads in high school, where it's hundreds of hours, you're like, oh wow, it just doesn't take that long.

This whole concept that we need to have kids spend 6 hours a day plus homework for 12 years—it's just not even close to true in this new world.

The second part is, okay, they're only doing 2 hours. They must not be learning; your scores aren't high enough. Our academic results—the reason we use a lot of standardized tests is, first off, when I talk to the parents, even our parents who don't like standardized testing actually love it with Alpha, because they wouldn't believe our results.

I'm like, you know, if I sat here and just said, “Hey, it's better. Trust me, trust me, those 2 hours are totally working,” they're like, “What's the test?” We do a midyear test, so you come in in August or September and take your MAPS test. We use a third party, and then in January you get your midyear update.

For parents who come in, at first they're like, “Okay, this sounds really exciting.” Then their kid comes in, is doing the 2-hour day, and loves school. Somewhere around December, because they're home at Christmas, family members are like, “Is your kid really learning and all this stuff?” They start to get a little freaked out, and then the January MAPS test comes back and shows they're learning twice as much in 2 hours a day.

And they’re like, “There is no school in the country that can deliver these kinds of results,” because you can take in kids who are behind, right? If they’re learning twice as fast, they’re going to catch up. So you can take a kid at the 50th percentile and get them into the top 10%. Or you can take a kid at the 99th percentile, which is usually capped out, and all of a sudden put them years ahead, right? And so it works on either side because you’re just focusing on the growth, which is: How do you teach the kids more in less time?

Patrick O’Shaughnessy

And you also can just sit there—anybody who, if you talk to the experts in learning science, they’re like, “Well, theoretically, this has all been predicted for 40 years, but no one’s been able to implement it.”

Oh, okay. I guess we’re implementing it, and it’s working.

Patrick O’Shaughnessy

I like the idea that you told me one time that the microscope unlocked biology, and AI unlocks learning science. It seems like that’s just true.

Yeah. This is—I agree. Back to my favorite analogy: the sciences—biology, chemistry, and physics. Let’s take those. They had their days of wandering in the wilderness, lost in the wilderness, where doctors were still bloodletting and chemists were blowing themselves up. Then there was the invention of instruments: the microscope for biology, the telescope for physics, and the analytical balance for chemistry, which allowed more precise measurement.

And that precise measurement is what allowed those sciences to really take off. Learning science has sort of been that way. It’s been in this wilderness period for 40 years, where we still have teachers in front of the classroom, just in a bad model. Teachers are great. The teacher in front of the classroom is the problem, right? The model is the issue, but teachers are awesome.

I want to be really clear, because we should talk about how important the adults in a building are. But back to the learning science: AI allows you to do precise instruction to the student and then, second, measure the results.

If you take the AI out, your effect is dominated by the classroom, right? You don’t know: Is the student motivated? Are they even trying? What’s the prior knowledge of the student? Did the teacher actually instruct it correctly? And so, in the last 40 years of learning science, the environmental variables have dominated.

With ours, with an AI giving very precise instruction and being able to measure it in a closed loop, that closed loop is the magic, to your point. We have a closed loop that no one else has. Our team comes up with an idea and says, “Okay, theoretically, let’s take Dan Schwartz’s Chapter 8 and put it in.” We generate lessons, the students use them, right? How well does this improve teaching? They take a standardized test. We measure how long and how much time they spend. They take a standardized test, we have the results. Did it work? Right? And we keep that loop going.

We’re basically the only place that has that closed loop, which is why we’re able to get these crazy, awesome outcomes and other people aren’t.

Patrick O’Shaughnessy

You also told me the best way to learn is by analogy. And there’s another analogy, which I think is quite cool: to think about this program like an iPhone, where, for a long time, every year they were releasing a new one and it was constantly getting better.

I want to understand some of the dimensions of that, with the 2-hour piece specifically, before we get to the afternoon and motivators and everything else. I know you started, like I said before, with tools—you’re starting with a standard set of curriculum that maps to the MAP tests and everything else. Talk about the iteration: What has improved the most since you first started applying this idea? Not only is this core idea of the learning loop better, but each component part of it can probably get improved, like the iPhone does year to year.

4. AI-Generated Personalized Lessons

Yeah. Here’s the magical part, and this gets into the time back of what we’re going to be launching to bring this outside of Alpha: generating personalized lesson plans for kids, for them.

It’s the magical unlock because it affects everything, which is back to motivation and your point about analogies and learning. Let me go to a far extreme. I have a first-grade boy who’s like, “I hate books. I don’t want to read.”

Patrick O’Shaughnessy

And you’re like, “Here’s a book.” Right.

Good luck.

Patrick O’Shaughnessy

Good luck.

Okay, just a second. What’s your favorite movie? Right. Oh, the Avengers. Who are the buddies on your soccer team? Really? Okay. I am going to generate a personalized choose-your-own-adventure story where you and your buddies are saving the world, right?

Patrick O’Shaughnessy

Right?

And it’s going to be geared exactly to your Lexile level, right—your reading ability—which is back to learning science. You want to stay within 10% of where they are, so it’s not too easy and not too hard: the zone of proximal development, right?

And so you’re sitting there, and now I have a lesson that you are going to engage in. You’re going to be reading it, saving the world, choosing your adventure and where you’re going, and engaging. Kids who quote-unquote hate books will spend an hour choosing it. As they learn more, you can move up the Lexile level and such. You can build a personalized lesson just for them that’s dynamic.

You want to learn fractions, right? You can sit there and say, “Oh, you don’t want to do math, but you want to be a fashionista and a fashion designer. Okay, well, how are you going to do that?” Right? We’re going to make this all in rulers and dress patterns and all those kinds of things. All of a sudden, you’re going to want to learn fractions.

Or, I love you—if you love baseball, I’m going to teach you statistics through the tournament of baseball. AI—and this is the generative part of AI that really matters—is that all educational content is obsolete. Every textbook, every lesson plan, every test, all of it is obsolete, because GenAI is going to be able to generate a personalized lesson just for you, right?

It’s going to take into account: What is your knowledge graph? What do you know and what do you not know? It’s going to take in the curriculum: What am I trying to teach you? It’s also going to take into account your interest graph: What do you care about?

Patrick O’Shaughnessy

Yeah.

Right. And so all of a sudden, you can sit and say, “What is my knowledge graph of what I know and what my interest graph is?”

Here’s an example of an AI-generated one. People were doing AP World, and they were coming back from a Taylor Swift concert, and they were like, “I’m having trouble with the start of World War I.” And you’re like, “Okay, well, generate a lesson plan in terms of Taylor Swift.”

All of a sudden, when—I can’t remember exactly—but when she was onstage getting the award and someone came up and took it, and then there was all the controversy and there was a mediator, it literally just created a table of all the people who were involved: the Taylor Swift people and then the World War I people, with likely Archduke Franz Ferdinand and all that. Literally, the kids were like, “Okay, I’m going to remember this forever.”

For AP World History, our high school kids generated this where they go into Sumo, which is an AI music generator, and they literally, just like Hamilton—here’s back to interest level and generating content. Every kid coming into high school these days, everything they know—because we don’t teach history that much anymore in standard school—is basically whatever is in a Hamilton song, right? That’s the—if it’s there, they know it, and if they don’t, they don’t.

Well, taking that theme even further, for every unit in AP history, the students generated a song to teach themselves the concepts, right? They were all just singing the songs, going down, having a great time, right? And they made the material relevant to their interests, right, which massively aids engagement and retention.

When you think about the biggest change, imagine getting personalized lessons, geared for your interests. Those lessons, though, are all based in the concepts of learning science. Think of giving a kid an unending stream of questions and knowledge at 80% to 85% accuracy, which is sort of the zone of proximal development. You don’t want to make it too hard, right? If half the time I’m failing, you’re going to get disengagement.

Patrick O’Shaughnessy

Talk to game developers. They’ve learned about this. Yeah.

Yeah.

Patrick O’Shaughnessy

You’re just like game developers. You use the same theory as game developers: How do I keep them engaged? If it’s getting too hard, let’s make it a little easier. It’s a scaffolding issue, or it’s a prior-knowledge issue. I can do that. I can use analogies that they know. It’s Taylor Swift or baseball or whatever their concepts are.

And I just give you an unending stream of questions that you’re engaged with, that you really like, right? The amount that you are going to learn is 10 times more than you would passively sitting in class, disengaged. Yeah.

And so that’s really what the magic is of being able to auto-generate content for each student based on their interest graph. Now, there is cognitive-load stuff also. When you feel confused and overloaded, that’s usually when you’re overloading your working memory, right? That’s also when careless mistakes come in.

Every student’s different. You tailor the question based on how many slots they have, right? How many chunks they have? That way, it’s not too hard for them, so you’re not overloading them.

A classic example of this is fluency: you can solve harder math problems if you know your multiplication tables and division tables. For students who are struggling in advanced math, part of the problem is just, “Go memorize your multiplication tables,” right? They all can do it, and then all of a sudden their higher-level math scores get better.

We have one student—this is sort of a unique, extreme example, and I hesitate to say it—who had a 740 on the math SAT. In looking at it, they were making careless errors. In some of the problems, they were overloading working memory. For whatever reason, this student had never fully developed fluency in multiplication and division tables.

So, we literally sent her back to third-grade math. We’re the only school in the world that will take a 740 math SAT student and send her back to third grade, and she got a 790.

Patrick O’Shaughnessy

Crazy.

Right? A 790. And it’s that kind of thing where, when you talk about the science of learning—and I say it this way—the parent and the student in this case were not excited that their 740 math student was being given third-grade problems. They were like, “That can’t be the issue.”

But it is, right? It is because you’re overloading working memory, and she’s just making careless mistakes. And once she has it, right?

Patrick O’Shaughnessy

Just a quick selfish aside: when will this be available for me? Again, I keep referencing that we’re going to come back to motivation. I’m motivated to learn lots of stuff. Investors in general are motivated to learn lots of stuff.

I’m often facing novel situations where I don’t know the physics or the technicals or this or that, and I wish I could just input, “Here’s everything about me. Here’s all my favorite stuff. Ask me anything you want to know. I’ll tell you anything. Now teach me whatever. Teach me avionics. Teach me whatever. I’ll take tests. I’ll do whatever it takes.”

How long do you think we are from a normal adult who doesn’t need motivating just being able to input the thing they want to learn about and having a system like that teach it to us much faster?

We’re focused on K–12. There needs to be somebody else who does that for adults. But in the next 5 or 10 years, everybody can have this. This is going to be a standard thing for everybody.

And, back to a visionary, I was talking to some ministers of education in other countries, and I was like, “Look, sometime in the next 5 years—I don’t know exactly when—but there’s going to be a tablet that’s less than $1,000 that is going to teach every kid on this planet everything they need to know in 2 hours a day.” That’s coming, right? It’s going to come.

Our AI burn is like $10,000 a kid now, but it’s going to be down. It’s going to be for everybody. It’s going to be on a tablet that everybody’s going to have. That is coming. If you’re a parent, you should wake up and be super excited.

AI is going to affect people in various ways. There are going to be winners and losers, right? I’m in Austin, and we have the robotaxis and the Waymos and all this stuff. I’m like, I would not want to be an Uber driver. This is just not good, right? It is not good for them.

For kids, it is an unequivocal good. This is Mackenzie’s favorite line, and I completely agree with it: this is the best time in history to be a 5-year-old, because we are going to transform their first 12 years. They are not going to sit in class. People do not like school, right? The school experience for most people is not good, and they don’t learn. We’re going to completely be able to transform their dozen years and make it awesome, as opposed to the chore that it is today.

Patrick O’Shaughnessy

How fast does the benefit diminish with age? The best time to be a 5-year-old—what about being a 10-year-old or a 15-year-old? Will that cohort still enjoy these benefits?

Going back to catching up and how far behind someone is, we’ll take anybody into an Alpha School academically if they’re willing to put in the time. When you talk about the time, you’re like, “Wow, you’re really far behind. So, you can’t do the 2-hour learning. You’ve got to do the 3-hour learning.” Okay? You need to take 1 hour of homework, right, or an extra hour in the class daytime in order to catch up.

On our dashboard, we literally have one for when people come in, where parents can look and see how many weeks it will take their kid to catch up if they do 1 hour of homework, no matter how far behind they are. You can take a kid years behind, right? You’re like, “Well, you’re 20 to 30 hours behind per year. If you’re 2 years behind, great—you’re 40 to 60 hours behind.” The number of hours just isn’t that much, and you can catch them up very quickly.

Anybody can take advantage of it. I believe that it’s getting the curriculum in, getting that closed loop, and measuring it.

Patrick O’Shaughnessy

Yeah, well, let’s talk about motivation now. We know that motivation is such a huge percentage of the problem. I think there have been interesting studies. The anonymous parent who wrote about their kids being in Alpha School talked about this idea that they tried to adopt for themselves some of the programs that were underneath early versions of Alpha, and they really didn’t have a lot of success. Their kids didn’t really want to do it. They complained about doing it on the weekend.

There wasn’t, critically, the feedback mechanism to show the kids where they are in a kind of gamified way. Solving for motivation, both in giving feedback on their progress and in what happens in the afternoon, has been a critical part, I think, of Alpha’s success. Just talk me through the entire journey of learning about this and what you do.

5. EdTech Fails Without Motivation

Yeah. So, motivation is 90% of the solution. Edtech is 10%.

Even though I come from a software background, the problem is edtech has failed. We keep putting edtech into schools, and test scores keep going down. So, no, it’s failed.

Edtech is necessary but not sufficient. It’s 10% of the solution. The 2 things every educator will say you need to teach a kid are, first and most important, a motivated student, and then, second, you need to put them in lessons of the correct difficulty—not too easy, not too hard. That’s the zone of proximal development.

Edtech does the second one really well. AI tutors can ask you what you know and what you don’t know, take an assessment test, and then give you those questions. But are you going to engage? Do you have any interest in engaging? If I don’t engage with the app, then I’m not going to learn anything.

Motivation is 90%. Alpha is 100% built around solving the motivation problem. There are 2 stories here. This was back when I moved my kid over, years ago. Brian, who was the co-founder with Mackenzie of Alpha, I asked him, “What’s something that I don’t believe about education that I need to know?” He’s like, “Kids must love school.”

The first commitment of Alpha is that kids will love school. And I’m like, “Oh, school’s sort of like spinach sometimes.”

Patrick O’Shaughnessy

Yeah, right.

And it sort of sucks and all that stuff. He’s like, “You are wrong. Kids must love school.”

That couldn’t be more true. If we’re going to fix education, we have to wake up every day and realize, why don’t we make these dozen years for kids? We spend plenty of money, right? We spent a trillion dollars on K–12 in this country to make kids love school. Right now, we put them in class for 6 hours, give them homework, and make it a very terrible environment. We’re not doing that. But we need to change it so they love school.

And that’s tied into how you have to solve the motivation. If you want them to learn, you have to do that.

And, just back to the edtech one last thing: edtech has 5% or 10% of kids who do well with it. There’s a whole paper out called “The 5% Solution,” where the successes around edtech are basically a set of kids who love academics for academics’ sake and love learning more. They’re self-motivated, high-conscientiousness, often high-IQ kids who do really well in the current existing system, and you give them apps and they do it.

But that’s not 90% of the market. For 90% of the market, it’s, “Okay, how do we motivate kids?” Motivation is the core of the design.

So, let’s start with the first part of this. My first week as principal, 3 years ago, I’m now accountable to “kids must love school,” right? They have to love school so much they don’t want to go on vacation.

I’m talking to some fifth graders. They’re new to the school, and I’m like, “Do you love school? Would you rather go to school than go on vacation?” They’re like, “No.” And I’m like, “Okay, well, what would make you love school?” And they’re like, “Less school.” I’m like, “Okay, well, how much less?” And they’re like, “No school.”

That seems a little light as a principal, to have no school. So, basically, we negotiated. I was like, “Look, how about this? If you engaged in the apps for 2 hours a day—and you’d have to engage; you can’t screw around, you can’t be doing other stuff—but if you really engaged with the apps for 2 hours and then you had 4 hours of the most awesome stuff you love to do, in workshops and team-based stuff that you really loved, would you do that?” And they were basically like, “Okay.”

And that’s when I was back with my team, my software team.

All right, guys. We have 2 hours, and we have to teach these kids as much as we can. All this learning science says it can be done, so let’s go figure this out and create a system where these kids are going to crush it academically in 2 hours, because that is the right thing. You have to first solve it in 2 hours and give them their time back.

The code name of our product—and actually, it’s going to be the product—is Time Back. When you say, “When does everybody get this?”—next year, our goal is to get this out. Our job—my job—is to get this to a billion kids, so I need a software platform to do it. You can’t do it with just the physical schools or just Alpha, so Time Back was the name of the product: give kids their time back.

That is by far the number-one motivator for kids. In that article you’re talking about from that parent, motivation is complicated. There are lots of ways to motivate kids. Our guides and the adults in the building spend all day figuring out how to motivate kids. Sometimes the kindergartener wants a sticker, or some students are competitive and want to be at the top of a stack rank.

But by far the biggest and highest effect size is giving kids their time back, and that drives huge performance. If you took all the other motivations we use at Alpha and said they had to sit in class for 6 hours and do homework, I don’t know that any of those other motivations would work. You have to solve the first problem: you need an engine that teaches kids 10 times faster so you can give them their time back. Once you do that, these other motivations work as well, but time back is by far the most important motivation.

It totally works because kids get it. You’re like, “Do I want to spend 2 hours engaging to then do 4 hours of great stuff?” Almost every kid is like, “Okay, that seems a lot better.”

So what is the great stuff? What happens in the second two-thirds of the day? The parents don’t want the kids back, so you’ve got to keep the kids. What’s going on?

The afternoons are all the life-skills workshops. There are tons of stories around it, but these are project-based workshops where kids are doing stuff they love while also learning life skills.

Think of a life skill that you want. “I want my kid to have entrepreneurial life skills.” Our fifth graders, last year or 2 years ago, ran an Airbnb, and they loved it. They had to learn P&Ls, gross margin, and all these things because they had to learn business, and they were loving it. They had to figure out how to set prices, and they were decorating the place.

This year, that same group was like, “We want to launch a food truck.” Austin has a big food-truck scene. They picked breakfast food. They call themselves Dippers. I was talking to one of them, and I asked, “Why’d you pick breakfast food?” They said, “Do you understand the gross margin on breakfast foods? It’s just enormous.” You’re like, “Okay, this is awesome.”

They have the food truck, and on weekends they go to car dealerships and have it open. They’re making $1,000, and they just think it’s great. When you think of the life skills we’re trying to teach, we have a whole life-skills curriculum.

Common Core is the standard academic curriculum, but parents have a standard of, “How do we teach teamwork and leadership?” Today, teamwork and leadership are taught via sports in America. Most life skills in America are taught in after-school sports. If you survey parents and ask, “Where did you learn your life skills—grit, leadership, teamwork, all those kinds of things?” they’re all like, “Oh, in sports.” That’s why sports is such a big part of it.

You can teach those skills in school. You don’t need sports to do it. School is a great place to teach leadership, teamwork, and all these other life skills. Our big ones are leadership and teamwork, followed by storytelling and public speaking.

If you go to the average 10-year-old and say, “Hey, do you want to do public speaking?” the kid is like, “Yeah, that doesn’t sound that fun. That’s not exciting.” How do you make it fun and exciting so they love it? We say, “We have a postgame press conference. You’re the athlete. Everybody wants to be able to do their postgame press conference.”

These kids all practice their public speaking, and then they go to Austin FC’s stadium. Once they pass, they get to see the game, and then they get to sit in on the postgame press conference with their favorite star. We have a deal where, after the stars leave, a couple of the press members stay, and then the kids go up and get interrogated by them. Every kid is like, “Okay, I want to be able to do that.”

That’s a life skill I want to teach: public speaking. How do I frame it in a way where kids are going to love it and really engage with it?

How do you have the teachers, guides, and coaches do this more systematically? The software really tracks with me. It’s very clear to me intuitively how it might scale because you can program it into the system. It seems less obvious how you could scale something like those examples.

I have a dear friend whose son is super interested, as a young entrepreneur and teenager, in building this little online business. I’m sure if you told his son, “You’re going to get 4 extra hours a day to work on your business,” he would bust his butt on the 2 hours to do anything to run his online sweatshirt shop.

The most amazing version of this I can imagine is that some amazing person who knows a lot about entrepreneurship, online selling, and all this kind of stuff is helping that kid. That matching exercise just seems less scalable and less obvious how a system can get built that allows that afternoon to be as buttoned-up as the morning.

It depends on the age of the kid. K–8, you can build an afternoon, and we have built an afternoon set of workshops and curriculum that’s structured, standardized, and can be rolled out. It’s not custom, where they’re doing individualized projects.

Once you get to high school, we have what we call the AlphaX project, which is your 4-year, super-passion project. You have all afternoon for 4 years to work on what you love, and that is very custom because every kid is different.

We have one student who’s trying to get a Broadway musical produced. It’s the first all-teen-produced and developed Broadway musical. She’s spent years building it out. She sourced all the people from TikTok—all the singers—and she can fill 1,000 people into a venue now. To your point, there’s no guide or coach at Alpha who can help her.

Our high-school curriculum very much is about how you become a self-driven learner who can use resources, including AI, which really gives you superpowers, to do these projects on your own. That’s the difference. In K–8, we have a life-skills curriculum that the kids do, designed and managed by the guides and coaches. In high school, they transition to taking ownership themselves, and then they get feedback on how they’re doing from the market.

Yeah, right.

From the public.

Are you selling stuff online?

They learn how to reach out, and they DM people on Twitter. One thing we’ve found is that super-successful adults overinvest in ambitious teens, so all these kids are getting help from adults in order to do it. You can’t hire a teacher staff who can do all that.

What about the stuff that goes into the K–8 bucket specifically? I could imagine that if the carrot in the 4 hours was an amazing field day, that’s different from needing to memorize some poem and give a public speech. The amount of actual motivation can vary a lot, and it’s student-dependent.

What have you learned in the 3 years about the sorts of things that kids actually love doing as the reward for doing their 2 hours in the morning?

Here’s the most controversial thing I think I’m going to say in the talk right now. There are 3 things that, if I could wave a magic wand, I would get every parent in America to believe, because they don’t believe them today.

If this were the outcome of the podcast, it would do more to change education than anything.

The first thing is: your kid must love school. We’ve talked about it, and they should want to go to school instead of going on vacation. If every parent believed that, it would be critical to changing education.

The second is: your kid can crush their academics in 2 hours a day.

You cannot subject your kid to 6 hours a day plus homework. We need to stop that—2 hours a day.

And then the third—and this is the controversial one—is that we've learned that the key to your child's happiness is high standards. That's the controversial part. Kids want to do awesome stuff. Kids think they're awesome. Kids are awesome. Parents are very worried, naturally. They care about safety, and they worry about their kids trying and failing.

The workshops that drive excitement and make kids love school are not easy workshops. The kids want to go do hard things, and they want to struggle and fail. They don't want to struggle and fail, but they want to struggle and fail on their road to success. They need to be supported by a caring adult. You need high support and high standards.

When we talk about our guides, we should go into that. Fundamentally, kids love it. One of the things you'll see in the videos is Mackenzie's Instagram, which is basically all the videos of these workshops. All she does is go to school and film them. There are half a million or a million people who love it because they're like, “Wow, this is what school should be like.”

We have kindergarteners who climb a 40-foot rock wall. When you hear that and you're not in the Alpha environment, you're like, “Oh, that is so bad. That's so whacked. That's too hard.” But when you actually see the kids go through it, they love it.

We mic up the kids as they're going up the wall, and you'll see somebody who's just like, “Okay, this is scary, but I have a growth mindset. I can do this.” Then they achieve it. It's all safe, and they come down beaming, like, “Oh, this was great.”

Our guides, who create these workshops, have a whole library of the ones we've developed. We're constantly asking, “What are the ones that really work, and what are the new ones we're trying?”

Last August, our 2nd-grade guide, Faith, came to me and said, “All the 2nd graders can run a 5K in less than 35 minutes.” I was like, “This seems too hard. That seems too hard. The parents are really not going to like that if it doesn't work.”

She's like, “I know they can. I was an athlete at UT. They totally can. Let's do it. Let's do it.”

She said, “Here's the key. In kindergarten and 1st grade, we teach a lot of growth mindset: You should think you can. You guys can do anything. In 2nd grade, we've got to start teaching them how to do anything—not just think it, but how to do it.”

She said, “We're going to teach them Atomic Habits—the book Atomic Habits—which is basically about getting 1% better. Most adults don't learn that until after college, when they read the book. But you can teach the concepts in that book to a 2nd grader. Everybody gets it.”

She got all the kids together and said, “We're going to run a 5K. Who thinks that's impossible?” Every kid was like, “Okay, that's totally possible. I can't do that, Miss Faith.”

She went out to the track and said, “I've signed you guys up. In August, I signed you up for the Jingle 5K at The Domain at Christmas. We're going to run it. Here we go.”

But on day 1, they were just going to walk the track. All the kids walked the 5K, and she said, “Okay, so we've established that everybody's going to finish the race.”

Yeah.

The next time they went out, they ran a quarter lap and walked the rest, then half a lap, and so on. They incrementally built up until, by December, these kids were able to run the 5K in 35 minutes.

You look at the finish line, and the kids are crying. It's so awesome. But going back to the life skill that you taught, it's not just about the heart. They love that it was difficult, but they sit there and they're just like, “Mom, Dad, I know how to do anything now, and I have those skills that I'm developing.”

As a parent, you're like, “That's the life skill that you want your kid to know how to do.” The way to do it, though, goes back to how the whole system has to work. You have to have a system where the kids love school and where it only takes 2 hours, so they have these afternoons to do this. Then you have to have those 3 beliefs, including that high standards are the key to your kids' happiness.

It's an amazing lesson. Can we talk about these guides? I think everyone listening will imagine an amazing teacher they've had or their kids have had. Maybe they'll also imagine a terrible teacher they've had or their kids have had. It matters a lot.

I can cite my favorite teacher, looking back to 3rd grade or whatever. What have you learned about who these people are, who they should be, what skills they should have, and what role they should play? Same blank sheet of paper, first-principles approach here, I'm sure. Walk us through that part of this.

The first thing is, when you hear about AI tutors and all this stuff, that's a complete misconception: Adults in the building are critical.

Yeah.

People have this image of robot terminators and technology. We actually have a lot of discussion around the terminology. Should we call them teachers? They don't do academic teaching. If you are an adult in our buildings, you do not do academic teaching.

That's why, when we started, we were like, “We shouldn't call them teachers, because it gives the wrong impression.” So we call them guides and coaches. But then you're like, “Well, people think they're teachers, and it's a new role for teachers.” Anyway, we'll see. We'll figure it out.

Fundamentally, think about this: Every adult has 1 or 2 teachers who transformed their life. Our guides are that for our students. We literally have a survey question that we ask every 8 weeks, which is that exact question to kids: “Every adult had 1 or 2 teachers who transformed your life. Is your guide that for you?”

That is the standard we hold our guides to. These adults—these guides, coaches, and teachers—their job and what they're accountable for is to deliver the 3 commitments to the kids: that they love school, they learn 2X, and they learn life skills. They're accountable for that, and they are seen by the kids as, “You are transforming my life.”

So how do you know who's good at that? The first part gets back to academics and the teacher in front of a classroom. Whoever the teacher is in your head, the reason they transformed your life is not because they graded your 7th-grade science quiz really well. It wasn't that. It's because they met you, motivated you, understood you, and probably held you to a high standard and convinced you that you could do a lot more.

Going back to the key to your happiness and success being high standards: Great teachers are high support and high standards. There's a great book—your readers should read it—by Dr. Joerger called 10 to 25. It's a great book, and we work with him a lot on all this stuff. Every guide at Alpha lives that philosophy.

That's what you're looking for. You're looking for individuals who can hold high standards but also provide high support. It's different at different ages. If you go into a kindergarten classroom, you're going to get Charity or Mallory, and they're going to wrap your kid in love, because at kindergarten, that's what you want.

But as you get up to middle school and high school, they're going to be able to sit there and bond with you over, “Okay, let's talk about the high standards. Look at all these great things that you can do.”

It's various things, but because you don't have to teach 7th-grade science in our school, we actually hire former coaches and athletes. We have a sports academy, and we literally have former NBA coaches, former NFL players, and former MLB players. They're awesome at holding high standards and providing great motivation for the kids.

When you don't have to do the academic side, you bring in all those other things and focus 100% on, “I'm going to hold high standards for the kids. I'm going to be able to motivate them.”

We're probably the only school where you have a system that lets you go in for every kid and see every motivation that's ever been tried by any guide at Alpha to motivate them. You can see what they really love, how you engage this kid, and what their interests are.

Our guides spend time getting to know the kids. During the 2-hour session, for example, the kids do 25-minute Pomodoros. The guide will pull a kid out one-on-one and say, “I'm not going to do the math stuff today. Come sit down and talk to me.”

The guide will say, “How was your weekend? What happened? Did you win the softball game? What happened?” They get to know the kid. “You seem to be not as engaged anymore.”

What’s going on? And they talk through all of that with them. So they get to know the kids. We talk about Mount Alpha, which you’re climbing up, and your guide is there to help the students get up the mountain. People who love to do that are the ones you want to hire.

You know, we pay our guides a lot more at Alpha than standard teachers.

How much more? What percentage more?

Our minimum pay for a teacher is $100,000. In the Austin market, that’s basically double the pay. You’re going to get great talent, and you’re going to get great talent not just in teachers. We can go get people who are the best teachers in any teaching role, but we can also get people who aren’t traditionally teachers. That’s very much what we believe.

The core of it really is that students want these great teachers, and you don’t have a class-size problem. The economics of a school are based fundamentally on how much you pay the teachers and what your class size is. We have some other economics, like our AI and our workshop costs, but fundamentally that’s driving the economics of schooling.

At the lower grades, we have low student-to-guide ratios, so you’re going to have 5 students per guide. But as you get older, what you actually find is students don’t want that. By the time you get to middle school, they’re like, “Mr. Liemandt, I’m an adolescent. I don’t need all these guides around.” They want more freedom, but the ones they want, they want really high-caliber, awesome people who are going to know them, motivate them, and hold them to high standards.

All right, I want to talk about feedback loops. This is something that we’ve referenced many a time. It’s a huge part of your success outside of Alpha and leading up to Alpha that has allowed you to fund it and be such a force behind it. In your life, you seem to have learned a tremendous amount about the power of well-enforced feedback loops.

So maybe just give us the foundation of this belief, and then I want to talk about all the many ways in which you’ve been able to deploy these learnings into the Alpha School setting to give kids feedback—not just motivation, but to show them their progress, show parents, show coaches and guides. Where did this deep belief in the power of iterative feedback come from in the first place?

There are 2 parts to it, I guess: the individual level and then the systemic level, which is part of the feedback loops that are really important. The biggest yet-to-be-proven part is that we need to scale this. We need to get this out to a billion kids. So how are you going to measure success? I always look at myself as sort of a product person, but then: How do you put in the systems that allow it to scale?

The key to the systems scaling is the feedback loop, so that you know what’s going on. When we open a campus, how do I know that the kids love school? How do I know that the parents are super excited? How do I know they’re learning 2x in 2 hours? Did we hire the right guides? Are the guides awesome? Is my guide in New York City going to be as good as my guide in Alpha? How are we going to be able to measure it all?

The key on all these feedback loops is, if you want to scale it, you have to measure it, and you have to provide that loop if we’re going to be able to say we’re going to keep this quality standard at scale. There’s sort of all that that we can talk about, and that’s a big part of what AI enables. It enables scaling at a level of quality that, I believe, 10 years ago you couldn’t have achieved.

The second one around feedback loops is receiving feedback. One of the most important life skills to teach is that if you are somebody who actively seeks feedback to get better, that is one of the most important life skills you can teach somebody. Receiving feedback is one of the skills in our life-skills stack. For all our life skills, we have a test that you have to pass, because you have to be able to measure it.

At kindergarten, the test is: Can you receive feedback without crying? By 8th grade, do you know that the key to getting better is actively seeking critical feedback? Do you actively go seek it out in order to improve?

The day that Alpha parents get really happy is when their kid comes home and is like, “Mom, Dad, I know I haven’t been a good member of the family. How can I be better?” And they’re like, “Okay.”

That whole concept of seeking out in order to improve is critical to just human development, getting better. We call it watching game film. Great athletes are like, “Yeah, I’m going to go watch the game. How did I do? And how am I going to get better?”

We do the same thing in academics. We have game film where you’re like, “Why did I screw up these questions on the test?” And they literally will sit there and say, “I missed this question because I didn’t know the material,” or, “This was a careless mistake.” So they can start to sit there and say, “What is my problem? I don’t know the material, or actually, it’s my test-taking skills.”

And you help them do that? You show them the game tape of themselves learning?

Exactly. That’s part of when we talk about our AI tutoring. We didn’t get into all the technology, but how we use it is completely different from what everybody else thinks.

I believe there’s this whole view that we’re going to deploy ChatGPT to everybody in America and things are going to get better because everybody needs to use AI. If you take that away from Alpha, you’re like, “Oh, we should have deployed ChatGPT to everybody.” Okay, wrong takeaway. It is literally the worst thing you could do. Just to be very clear, you are not a principal at a school if you think ChatGPT is good. Ninety percent of the use of ChatGPT is to cheat.

Now, there’s a 10% use case for people who are like, “I need a Socratic tutor. I’m super excited about it, and I need to dive in deeper.” That 5% to 10% of kids are going to crush it with it, and you should give it to those 5% to 10%. Ninety percent of the kids are like, “How do I use this to cheat?” Okay?

Because they want their time back.

Because they want their time back, because they’re just trying to get through it. We literally don’t have chat functionality activated in our AI. It is not a chatbot.

It’s terrible.

Yeah, and that’s not what people need. Back to where we use the AI: generating personalized lesson plans and personalized lessons themselves. That is a great use of AI.

A second great use of AI is actually to watch the screen and coach the kids, right? To give them the game film. We can go from the remedial: “Hey, kid, you’re literally scrolling through the article and looking at the questions and trying to search because you don’t want to read the whole article.” That’s not really going to work well. Or, when you miss a question: “Here’s the video that can help explain it,” the little explainer of what you got wrong, and you skip it every time. “If you really want to get out of here in 2 hours, read it. You need to do it.”

Our biggest thing is our vision model, which is actually part of the big, expensive part because you’re streaming full-time. It’ll come down in cost, but right now it’s a lot. We have a waste meter because, once again, if you’re going to give the kids the time back, they want the time back and they want to get better. They want to become more self-driven learners who can move through this faster.

So we literally say, “You’re wasting 50% of your time.” And the kids are like, “Whoa, coach me up. I don’t want to waste 50% of my time. I’m spending 3 hours when I could spend 2,” right? Giving them the coaching of how to learn better, how to learn faster and more effectively, that’s a great use of AI.

So, just so I understand, I almost want to imagine we have the engineers of the system here with us and we’re now in a technical explainer. The thing is built. I’m interacting with it. There are different things I can do and not do. Is the entire process—me looking at it, the touches, the scrolls, everything—being completely monitored? And that is being fed what, through what kind of system? That sounds complicated. How does that system know it’s the little thing at the end, the explainer, that I’m skipping?

Help me understand the engineering of the thing. Use Khan Academy, since everybody understands it, and imagine a dynamically generated Khan lesson. The video is dynamically generated. The video can basically be about half as long as a Khan video because it uses the concepts of learning science and is more active, right? Passive is bad.

So, more active engagement means you can teach kids in less time. It’s based on their interests, right? You’re doing that, and then the video stream itself is actually being fed to an AI model that’s sitting there looking for antipatterns.

What are the things that kids do that slow them down, that are learning antipatterns? It’s looking for those, and when it has them, it’s going to be like, “Ooh, look, that’s waste,” and then it’s going to update the waste meter or flag the kid and say, “Stop doing it,” depending on what the mode is.

When you think about the technology behind it, just think back to that data. What’s valuable data? It’s the same as having the data stream of all the Tesla cars running, right? You’re then saying, “Okay, what are we doing?” It’s the data stream of, “Okay, here’s what kids do when they’re learning, what they shouldn’t do, and what to avoid.”

Then you move to—the first phase is to go find all the antipatterns, and then the second one is to coach them up on it.

So maybe it’s a data stream, just a raw data stream that you have privileged generation of because the students are using the thing. Maybe talk about the team behind this. How many engineers are working on this? What kinds of people are they? Where did they come from? Are these the best engineers from your Trilogy days? Who are these people?

And then what is their process of turning the raw data stream into actionable lessons in the software?

There are a couple of different parts to it. Incept is the core generation engine, and the name actually matters. Incept sort of has 2 definitions. One is the movie, right? It’s important because you’re putting a mind virus in kids’ heads. That’s what you’re doing when you’re educating a kid: you’re putting these thoughts into kids’ heads.

There’s a second definition, though, from 1,000 years ago, the original one from Cambridge and Oxford, which was “qualified to teach.” There was a subtle underpinning that you trust the person, right? Teachers are trusted because you’re putting things into kids’ heads. It’s important to understand that it’s 2 parts: you’re putting stuff in a kid’s head, and you need to be trusted about what’s going in.

There’s going to be a lot of controversy, I believe. There is in AI in general, but especially in education, about how we make sure these are trusted generators that are putting things in. You have TikTok as an untrusted generator that currently dominates, right? We should talk later about our standard of TikTok as our competition, which is that we have to out-TikTok TikTok if we’re going to take over and get the right thoughts into kids’ heads.

So there’s the Incept team. You should think of it as using all the different LLMs, with our stuff on top of them, where we’re generating content. It’s going to be taking in streams of: What’s the curriculum we’re doing? What is the kid’s knowledge graph? What is their interest graph? What is their cognitive-load model? Then it will be able to generate the dynamic lesson.

You should think of this as your standard AI-type team embedded with a massive set of learning scientists.

Yeah.

That’s a critical aspect of it. That’s why, if you just have a normal LLM generating content—and why I believe all the LLMs that are out, including all the educational ones, aren’t embedded with learning science—they don’t follow these principles. They don’t embed mastery and all these other concepts. That’s why, generally, when you give them to kids, they just use them to cheat.

Then you have that side. The second side is the data stream, right? It’s basically a video model. It’s just a vision model, and you’re sitting there saying, “How do I optimize this vision model to find what I need?”

How it started was literally just humans. When you talk about human annotation—

Yeah.

—just watching videos. 3 years ago, we would sit there and have teams of people all night. The kids would finish, and then all night you would just sit and watch the videos and annotate them: “Okay, that’s what’s going wrong here. This is what the kid’s doing.”

All of this is reinforcement learning, and it’s just building out a video model—a reinforcement-learning video model—with human feedback. The human feedback is not from the kids; it’s from people watching it to say, “I need to train this up on how kids learn, what effective learning is, what the antipatterns are, and what we want to coach for.”

The first month, back 3 years ago in August, we would literally just take the output of watching the videos, and then the guys would tell the kids the next day.

Fundamentally, that’s what you’re doing. You need a vision model where you are sitting there training it on how to have kids learn better and how to be more effective—where the things are. Going back to the streams, you need this where you’re generating content. You need this to see whether they’re engaging with it, right? Then you need to be saying, “Am I really moving the standardized test scores? Are the kids really learning this material?”

You need that whole closed loop to be able to say, “Okay, yes, we came up with this concept, we generated this material, and, okay, look—yes, the kids are learning.” You can measure that.

Can we talk about the third one as the kid sees it? This is really helpful. I’m glad we’re going to a little bit more technical side of things. We’ve got Incept, we’ve got the vision model, and then we’ve got the feedback mechanism—or whatever—at the standardized testing.

How does the kid feel that? In the anonymous parents’ post about the school, they talked about hitting your minimums and your weekly scores, your daily scores—things like this. The kids seem to have this gamified sense of their own progress.

What have you learned about presenting that well and using that last closing part of the loop?

6. Gamification

Sure. The kids’ motivation: how do you drive kids’ motivation? At the end of it, you’re giving them a measure of how much they have to do a day and how much they want to do a day. You have a goal-setting part where, okay, kids, you can set your goals, right? This is what I want to achieve.

Then the system can tell you, “You need to do this much.” Actually, one of the changes we’re making for the coming year is that when you finish a lesson, one of the problems is that lessons are different lengths. This is a 25-minute Pomodoro, and I have 2 15-minute lessons.

What’s a Pomodoro?

A Pomodoro is a 25-minute segment that the kids do for each subject. They’re going to do math, science, language, and reading. Actually, they can do whatever order, but it’s 25 minutes.

You basically go into the app in the morning during the 2-hour session, and you’re like, “Okay, I’m going to spend 25 minutes in that,” right? Then the system is going to give you the lessons.

And you’re going to be like, “Okay, here are the lessons I’m going to do, and I’m going to do them to mastery. I’m going to do them to the highest standard possible.” I have rings that close, like the watch: “You did it.”

It’s literally just like the Apple Watch. It’s going to be like, “You did it at high accuracy.” If you’re doing it with low accuracy, you’re not learning. You’re just guessing and screwing around. You have to have 80% accuracy, right? You did it to mastery.

You didn’t just sort of mass-practice up front, where you locked it in. You spent the time. You did spend the 25 minutes, and then your ring closes. Boom, you get it.

Back to gamification, the kids have to do their 2-hour sessions to unlock their afternoons, right? All these privileges come around because you unlock things.

When you come into Alpha—actually, on your first day, if you’re transferring from standard school—you go into what we call standard school, where you literally sit at a desk. It’s like you’re in standard school, and it’s like, “Okay, your job is to become a 2-hour learner,” right? You learn the skills so that you can learn in 2 hours instead of 6 hours.

The kids are like, “I’m all into this. Let me do it.” They’re going to learn how to engage in the apps and how to use them correctly. If I sit and spin in my chair all day, I’m not going to be a 2-hour learner. I’m going to be stuck, right?

Then, as you achieve—going back to motivation models—as you show you can be a 2-hour learner, you can say, “Oh, I hit all my rings this week. I basically finished my lessons.” You graduate to be a 2-hour learner, which then unlocks all these different privileges.

Our guides have different ones at different levels, but when you see the Alpha pictures where the kids are in a beanbag with their computer, not sitting at a desk, they’ve unlocked those privileges. They’re like, “Yeah, you know how to be a self-driven learner.”

“You can do it however you want.”

The booths that we have are literally like the phone booths that you see in Silicon Valley. We have those in our schools, and they’re famous in our middle and high schools because kids voluntarily go into them to learn. Students learn, “Where do I learn best?”

My daughter is a talker, and she’s like, “I can’t sit at the communal table because I won’t get anything done. I’m just going to sit and talk to my friends.” So she’ll say, “I’m going to go into one of the booths so I can get it done.”

Kids learn this: “I totally know how to become a learner.” If they have the motivation—that is, you’re giving them their time back—they develop all those skills. When you talk about whether kids have agency, whether they own their learning, and whether they’re self-driven, that’s all part of the system.

That has to be where you create these two-hour learners, which is basically creating self-driven learners. When you think long term, that’s a critical life skill that you really want to have.

And then, how do you game it, or what is the gamification? There are all different ones. Guides will have individualized goals, team goals, and so on. At our Brownsville campus, the biggest motivator for the youngest kids is the petting zoo. If the whole level exceeds all their academic goals for a week and a session, they get a petting zoo. That drives motivation.

They don’t want money. They don’t want—what you get, right, is the bunny and the petting zoo. Those kids would go do 10 hours of homework.

Can you talk, for the kids that do want money, about Alpha?

Sure. We have a whole economy based on Alphas. Fundamentally, you do these academics, then you can earn Alphas. How can you spend those Alphas?

There are a lot of motivations that parents love. The petting zoo sounds really cool, or you get more privileges—you earn privileges. Money is obviously the controversial one, right? That causes everybody to just go, “Oh my God, that’s a terrible one.”

So, I’m going to give a couple of examples. Your kid can earn money for doing academics. Parents generally say, “Okay, I don’t like that.” Now, your kid can earn money by doing academics so they can fund their passion project, the afternoon workshop they’re trying to do. When she’s trying to do the Broadway musical and has to fly up, she’s going to earn money that’s going to allow her to do that. Parents say, “Okay, that seems okay. That seems like a really good idea.”

Or, for this audience, it’s going to fund your financial-literacy investment account. We’re going to take all the middle-school kids, and if they crush their academics, we’re going to fund an investment in Robinhood for kids or whatever investment fund to teach them investing.

The key, if you’re going to teach a kid to invest, is to make sure it’s real money, it’s theirs, and it’s something they earned, not something that was given to them, because of loss aversion and all that. Parents are like, “That’s a great idea. I love that.”

We’re going to have a 50% penalty for early withdrawal. Parents like that, too. They’re like, “We’re going to teach them to save.” There are all these behaviors around money that are good.

Those are all examples of what our system does, where you can earn the Alpha bucks and then spend them. This is what every parent should do, because this is super controversial. Everybody hates it, but if you care about academics, this is the single best thing you can do to fix academics.

The number-one issue in academics for kids—and let’s start at middle school—is that they’re struggling. If you look at the curves in learning in America, they all start to flatten out in middle school. The median person in high school in America actually doesn’t learn anything.

Scary.

It’s scary that, on a MAPS test, which is a 300-point scale, the median American goes up 1 point in 4 years. One point on a 300-point scale. They don’t learn anything.

It starts slowing down in middle school. Motivation and engagement are an issue. The other issue is the ramification of having a time-based education system instead of a mastery-based education system.

You just keep pushing people up when they’ve only learned 70% or 80% of the curriculum. You should think of it like Swiss cheese: you’re building a foundation with all these holes. Eventually, as you get high enough, it just gets to be too much and collapses, and you can’t learn anymore.

That’s what happens when kids hit algebra. The problem is not the algebra; it’s the prior knowledge. The number-one thing we do is say, “Okay, we need to go fill your holes.” So, let’s talk about marketing and motivation, which is probably half my job as principal.

If you go to a seventh grader and say, “Look, dude, it’s not your fault. However you got here, you don’t know this fifth-grade material, so I need you to go back,” they are not into it at all. The only people who are less into it than the student are their parents.

Parents have been taught, “My kid needs to be at grade level no matter what.”

Yeah, and that’s just the opposite of what the learning science says. It’s also the opposite of what a tutor does. When you hire a tutor for your school, all they’re doing is saying, “Dude, just give me a fifth-grade lesson.” They don’t call it a fifth-grade lesson or whatever, but that’s what you’re doing.

Fundamentals, correct?

Everybody understands when you apply this concept to sports. Every parent gets it. They all get it. They’re like, “Yeah, master the basics.” When it comes to academics, it’s like, “No, no, no. Do all the advanced alley-oops.”

That’s a fundamental way that parents think about the world that just isn’t right. So, how do you fix it? I’ve got to deliver it.

I went to the seventh graders and said, “Guys, all of you can get 100% on the Texas Star.” They’re like, “No way, Mr. Liemandt. No way.” I’m like—and back to money motivation—“I’ll give you a $100 bill for 100 on the Texas STAAR. $100 for 100.”

They’re like, “That’s still impossible. I can’t get 1.” I’m like, “No, no, no, there’s a catch: any grade level.”

They’re like, “I can take a kindergarten Texas Star and you’ll give me $100?” I’m like, “Well, the Texas Star starts in third grade.” They’re like, “Okay, I’ll take a third-grade test.”

Yeah.

And I’m like, “Great.” So they go take a third-grade Star test, and they get 100. You give them a $100 bill.

They’re like, “Can I do fourth grade?” I’m like, “Absolutely.” They do fourth grade and get a $100 bill. They’re like, “Okay, you’re really going to let me do fifth grade?” I’m like, “Absolutely.”

So they take it, and they get a 75 or 85. I’m like, “Do you want the AI tutor to generate the lessons so you can get 100?” They’re like, “Let me see how many.” Then they see it, and they’re like, “I can do that in a week.” They’re in. Then they get 100.

Then you do sixth grade. For literally $400, we can do 2 really important things. We can hole-fill, as we call it, which is going back and, wherever your kid is, filling all those holes and putting them on a foundation where they can keep learning.

But second, you actually change how they think: “I can get 100 on a Texas STAAR.” If you actually survey Alpha parents, fewer than 10% of them think kids can get 100 on a Texas Star. That’s because, in the time-based model, the GT kids—only the super-smart kids—got 100.

In a mastery-based model, everybody can. It’s just a matter of effort. It’s not about IQ; it’s just about effort. As soon as you break that link for a kid, it changes everything about education. They sit there and they’re like, “I can totally do this. Oh my God.”

If you take Alpha’s stats, more than 90% of our kids who have been at the school for a year are like, “I can totally get 100 on a Texas Star.” The reason is because, during the year, our building—one of our buildings, our K–6 building down in Austin—gets more 100s on the Texas Star than a school district of 100,000 kids.

It’s just a different mentality and a mastery-based thing: you don’t advance until you know it. You sit there and, back to motivation, how do you get it? If you make everything 100, everybody thinks it’s too hard. So you say you can pass at 90, but 100 for 100, and everybody’s like, “Okay, I’ll do the extra work. It’s only a couple of questions,” right?

Because when they get less than 100, the AI tutor is like, “Here are the ones you missed. Was it careless, or did you not know it? If you didn’t know it, do you want me to give you the lesson?” The kids can get to mastery.

You put in this very lightweight motivation. We spend $20,000 a kid in this country; Alpha’s $40,000-plus. But we spend plenty of money. For $100 or $400, we could literally change how everybody sees both their capability and how they see themselves.

That’s a part that, back to motivation, really matters: it’s a perfect example of using money in order to get people to change how they think about themselves. A lot of kids have a view that they have these limits. That’s high standards, high support. That’s what guides are doing: they’re breaking through it. In academics, kids have these hang-ups of, “Look, I can’t learn that much. I’m that kid.”

My daughter will probably kill me, but I’ll use her as an example. When she got to middle school, she got to the top 10%, and she was just like, “Dad, I’m not like my older sister.” Her older sister is the academic one of the two. She said, “I’m top 10%; that’s good enough. I’m not top 1%.”

I was like, “The one thing I know about you that you care about more than anything is that you love to shop. How about $1,000 bucks? I bet you $1,000 bucks that you can’t get to the top 1%.” She was in middle school, and she filled her Amazon cart with $1,000 worth of stuff. She would just look at it in the morning and then go grind and do her apps.

We had a lot of family discussions about this, because my wife and I were like, “Okay, so first of all, you’re paying for academics, and you’re setting this really high standard and all this stuff. Is this going to be good for her?” That’s common. You’re back to being a parent: you don’t want to set too high a standard, and you wonder, “Is this motivation bad?”

Anyway, she got to the top 1%, bought all her stuff, and then sat down with my wife and said, “Mom, I know you didn’t like us doing this, but you need to understand what it did to my view. I never thought I was as smart as my older sister. I now realize I can do anything.”

I’m sure that, listening to a story about your daughter—thank you for telling it—every single parent, myself included, is just going to picture their kids’ faces in their mind and agree instantly that they would do anything for this moment of unlock, or several of them.

That demands, I think, the really important and exciting question: how the hell are we going to scale this system across the country? Like so many things, like a great product, it tends to get built in a small space where it’s artisanal, and you’re trying things and figuring out the system. Then, hopefully, like software, you can spread it and scale it across the country. What are the biggest barriers to that being possible?

7. Scaling Challenges

Let’s talk on the technology side, because it’s not ready yet, and we have a whole team of people who wake up every day trying to make it ready to scale broadly. The engine—the TimeBack engine—we still have hallucinations. It’s AI, and there are still problems.

An Alpha environment is super safe. The kids love school; they’re all stoked. But with high school kids, it literally comes up with some whack-job thing, and the kids take a screenshot and send it in: “AI’s acted up today.” It’s fine. You release that into the wild, and AI robot tutors are terribleness.

We’ve got a burndown chart of how we’re going to get there. We think, and we do believe, that next year we’ll have betas; we hope to have them between August and December of this year. The platform will be out. The first part is that I’ve got to get the technology a little better.

Once that’s done, you get into your channels of distribution. Obviously, homeschools are the biggest thing: there are millions of parents who are like, “Okay, let me do this at home.” We have a test with 50-plus homeschoolers.

The problem with all of this is that technology is 10% of the solution. It’s only 10%. My 10% still doesn’t work yet. It’s got to. But then, how are you solving the 90% of motivation?

When you’re in an Alpha building or one of our schools, all of a sudden, we control it, we have the motivation, and we do it all.

You’re at a homeschool. Is there any motivational model?

Our test isn’t getting 2x in 2 hours with the homeschool kids. Literally, every time we test it, we’re building in motivational models and putting in other things so that we can get there. I believe next year we will have a homeschool version that somebody will be able to buy. The kid is going to be able to sit down and deliver 2x in 2 hours, for sure.

The second part is schools. We’re going to have a platform that, if a school wants to adopt this, then you can absolutely do it. Now, what are the hurdles? You’ve got to rebuild your whole school from the ground up. You can’t just tack it on. This is a full rebuild. Are schools really going to do that?

Public schools have a 10-times-harder job than me. If I were a principal, a teacher, or a superintendent of a public school, as a product guy I’d be like, “You have the most impossible job in the world.” My job is 10 times easier because I control everything. You have a really hard problem.

I believe, though, we need to start getting lots of data, proving out lots of things, and showing the efficacy—that this works—with the hope that public schools will adopt it a decade from now. My vision on this is that I’m going to spend the next 20 years of my life on this, and I’m going to need all 20 to get everything spread in public schools.

For sure, entrepreneurs need to build new schools, and this 2-hour platform will cover the 2 hours for you. You crush the afternoon.

Yeah.

Right. Let’s go start. Let’s have entrepreneurs build on top of it.

I’ll give you an example of one of the ones that—it’s not funded yet, but the business plan is super awesome—which is Texas just passed school choice: $1 billion. This group is basically building Texas Sports Academy.

Their goal is to go to every school district in Texas and build a micro school, which is going to be a micro middle school for D1 athletes. They’re going to be based on the TimeBack platform. Then, in the afternoon, it’s going to be one of the local coaches that everybody in town reveres, who is going to have a middle school for people who want to be D1 athletes.

They’re going to spend the afternoon doing whatever awesome athletics, whether it’s 7-on-7 football, basketball, or soccer. All these things allow you to create innovative educational models, so that, for a kid who’s like, “Oh my God, I will go crush my academics,” their commitment is that they don’t get to do the fun thing unless they do the TimeBack.

Yeah.

Unless they do the TimeBack. Because of that, and because of the learning science and the app, they’re going to make a commitment that—there are 1,000 school districts in Texas, 1,200 or something—they’re like, “Our 1,000 schools, our 1,000 micro schools—if you view us as an entire school district, we will be seen as the best school district in Texas academically. We will have by far the best academics.”

Tens of thousands of kids will be in these micro schools. Their hope is to launch a year from now, when the Texas ESA vouchers come due.

You can just see it for parents who are like, “Wow, that is an economic model I could get behind,” or, “That is an education model I could get behind,” where, “Oh my God, my kid’s going to love school, he’s going to crush the academics, and he’s going to freaking do a sport that he loves in the afternoon.”

That’s the middle school. There’s a whole set of parents who can then go build that out. As we prove this out at scale, there are actually a lot of famous professional athletes who are like, “I would love to start a school like that.” I believe that’s just a whole huge opportunity.

There’s another school that we’ve started, but it’s one unit, and we need entrepreneurs to scale it. It’s GT School, Gifted and Talented School. At Gifted and Talented School, we define it as loving academics for academics’ sake. When you’re motivated—

5% or 10% you’re talking about?

Yeah, it’s the 5% or 10%. My oldest daughter is actually like that, and she loves it. In high school, she read calculus books for fun. You’re like, there’s a set of people who just love that, and that’s where you’re just uncapping those kids. They aren’t learning 2x in 2 hours; they’re learning 3x, 4x, 5x. They’re just crushing it because they love it.

When we get all the guides together, we’re like, “Okay, how do we juice love of school? How do we make kids—how do we raise it?” The GT guides are like, “More academics. We need a third power hour.” All the other guides are like, “No, we don’t want more academics. More life skills.”

Anyway, with GT, you can just imagine: there should be 500 GT schools around the country built on the TimeBack platform, where kids who love that can just unleash it. The GT school is new, but it’s super fun.

I mean, those kids are like, “We’re going to be the last humans smarter than AI,” right? You literally just have this great group of kids who are like, “We love this stuff,” and who love academics for academics’ sake—Math Olympiad, robotics, writing novels, and all that.

So, back to when you really focus on kids having to love school, you’re like, okay, there need to be different school types for different kids, right? This TimeBack platform enables that because you can say, look, the academics are done. The coach doesn’t need to teach seventh-grade science, right? The platform does it. The app does it.

What the coach does is motivate the kids and then do the afternoon workshops. In this case, they’ll be athletic. So, if I think about scaling, there are lots of paths, and they all need to be tried. There are micro-schools, entrepreneurs starting schools, and refitting, which probably happens later in the adoption curve. Maybe public schools are the very last in the adoption curve—that classic adoption curve.

But the answer is, we just need to go hard at all of these solutions, where TimeBack and the motivation—the creativity around sources of motivation—are thoughtfully paired together. This is one of the tasks on my list to do. The first part is, okay, I’ve got to build this product that makes kids learn 10 times faster.

Then the second is, okay, we have to have these schools that are the exemplars that prove the case, because no one would believe it. Alpha is obviously the flagship high-end private-school model. But there’s a third part that we have to show in order to really unleash this stuff, and that’s a business model, right?

The business of education, because capitalism enters the chat.

It’s a trillion-dollar market with no good products.

Capitalism enters the chat, right? We need to bring capitalism in here, right? That’s obviously a very controversial view, and my view is, with TimeBack, we have to show: here’s a platform, but you can build great businesses on this, right? You have to be able to build great businesses in order to attract great talent.

The economics historically of education having to work completely changes. I believe we’re going to show—I mean, we’re in the cash-burn phase, and we’re burning hundreds of millions of dollars building all this stuff out—but it’s just like anything else: it’s a huge market, and if you have an awesome product—

The one element that I guess I would highlight, if I were talking to an entrepreneur, is that everybody’s like, “Don’t go into education. You can’t make any money. It’s terrible. It’s failed.” All those things. Great products, right? Parents will absolutely pay for great products. Societies pay for great education.

The amount that we spend is a trillion dollars for a relatively bad product right now, right? As an entrepreneur, if you look at this and build a great product, you can actually build a great customer base. It doesn’t have to be a money-losing nonprofit that’s supported 100% by donations, which is the world today.

You can’t sustainably grow. Part of the problem with schooling today is everything has to be nonprofit. The organization is nonprofit, and so you’re caught out scaling right there. There are good school systems and charter schools out there where their gating factor is, “Well, I’m waiting for someone to give me $40 million so I can build a campus. I have a 3,000-person waitlist.”

But, back to my economics, I need someone to donate $40 million versus, okay, well, let’s build a really good business model. You can build a business model out of it so you can open up 500 of them, because capitalism will provide capital.

Let’s adopt the ultra-skeptic mindset for a second. As the principal, I know you get to deal with this because you’re dealing not only with parents who are at the school, but parents who are considering joining it. If you had to name a thing or 2 that you struggle with the most personally—not skeptics that you can answer easily, but skeptics that are hard to answer—what are the hardest skeptical points of argument that you come across most often?

Well, your best example right now is just that we haven’t done this at scale.

Yeah. Easy for you to say. You’ve got all this money.

You have infinite money, a small selection effect size, and all those kinds of things. Absolutely. If the viewpoint is, well, there are seedlings of data here, trying to figure out, wow, look at some of these data points of what we can do and have been able to do, and then figure out how you would scale it.

But fundamentally, the best example is, okay, you’re this really high-end private school, and you have all these adults and all this money around that. The bigger issue is that education is hard because you have to define the product, and your customer bases have very different needs.

This is why, when I talk about public school being impossible, parents do not agree on what the purpose of an education is, right? What a parent wants in education is a big bundle. It’s a huge bundle, and they’re diametrically opposed.

We’re having that in Austin on the public-school side, where we have 3 schools that have failed the standardized tests, and so they’re trying to be shut down. The kids would move to schools where the superintendent is like, “Okay, we’re going to shut them down and move the kids to schools where they do better academically.”

There are parents who are like, “I don’t care about that. I don’t care that the school’s academics are failing, because my oldest daughter went to that school, and the seventh-grade teacher there is so great. I know her kids don’t really learn math, but she transformed their lives.” They’re going to pick it and say, “I don’t want academics.”

Actually, stepping back, one of the things that I did not realize going in is how few parents actually care about academics.

Back to product-market fit and pivoting and all that stuff from the product side: 3 years ago, I would sit down in the parent meetings and be like, “Look, if your kid’s going to engage in the app, they’re going to learn twice as much. 2× learning, right?” I’d sit there with the parent and say, “Look, Johnny’s not watching the video. If you just watch the video after he misses a question, he’ll learn twice as much.”

The parents are like, “Why are you pressuring him to learn 2×?” They just don’t want it. It took me hundreds of parent meetings to realize that I was slow on the listening-to-the-customer side.

Then we switched it to 2-hour learning. You’re in the same meeting with the same parents and the same Johnny, and you’re like, “Look, I’m trying to get Johnny out of here in 2 hours, but he won’t watch the video, and he can’t do the workshops until he does.” The parents are like, “Johnny, why don’t you watch the video? Don’t you want to do the cool things?”

So, back to product-market fit, figuring out positioning matters.

Positioning matters.

8. The 200-Year Education Problem

Positioning matters. Every parent wants stuff differently. When Mackenzie puts her GT families and the Sports Academy families together and has lunch with them, and the parents are talking about it, parents want different things for their kids.

The Sports Academy mom is just like, “My kids come home from Sports Academy, and they’re sweating so much. I literally just throw them in the shower fully clothed.” The GT parents are like, “What is wrong with you?”

They’re like, “Well, GT school was having a chess tournament at the end of the day, and when they came home, we just continued the tournament.” Parents want different things. There’s not a one-size-fits-all, and we need entrepreneurs who are domain experts in the field.

Do you want to do entrepreneurship? Build an investing high school and all of that stuff. I believe, if you look at our New York campus, it will have more financially oriented workshops in the afternoon than our San Francisco one, which is going to have more tech-oriented workshops.

Anyway, I believe we need to bring entrepreneurs in, we need to bring capitalism in here, and we need to create all the different parts of school choice so that parents can pick what they love.

There is this enabling technology. When you talk about whether you have something that’s 10 times better that can transform an industry, kids learning 10 times faster so that they can do it in 2 hours a day is that breakthrough product that can do it.

Back to your point, what’s the skepticism? It’s like, okay, well, we have to prove this. The product’s not done, so we’ve got to get it done. We’ve got to scale. Whether it’s us or anybody else, the tech behind this is totally doable.

Then the second question for parents is, okay, now that I can think about it, what do I want my kids to do in the afternoon? I think the broader question on all this stuff is—and I believe it’s at a societal level—when I was talking to the education ministers, I was like, once there’s this tablet and it’s 2 hours a day, what is your education system going to be? Because it’s coming.

How do we want our kids to spend their day, right? And what does society want to say? This is where kids should spend their day, because there's no point in having them in a classroom for 6 hours a day. That's gone, right?

Back to my 3 things that I want every parent: we need to make it where kids love school, right? Because they only need 2 hours. They can love school and then be in an environment of high standards and high support, do awesome things, and learn life skills. That's the vision that I want, and it is hard to execute and do all of those things, but that's the next 20 years.

That's the next 20 years.

No, that's my next 20 years: to do it. And I guess one last thing to anybody in the audience who's thinking, “Maybe I want to do this.” I can tell you, in the last 3 years, career-wise—obviously, I had a great career, and we'll talk about that in a minute, made plenty of money and all that stuff—but there's nothing that's more fun, exciting, and rewarding than working on education.

The last 3 years, in the whole scheme of things, have been 10 times better. It's awesome. In recruiting entrepreneurs to the vision, mission, and excitement, totally come aboard. We need to go build great businesses, but there's nothing that's more fun and rewarding. Back to the bigger picture, there's nothing more important for society.

Can you tell us a little bit about Prodigy and the idea of using games to deliver this?

9. Video Games for Education

Prodigy is a game out there. People have talked about gamified learning forever, and we very strongly believe in this. I'm investing a lot of money in this space because when you talk about scale, we have our physical buildings, right? I've got guides and an expensive footprint—an expensive cost to serve—to drive motivation. This engine, if I get kids to go through it, is going to provide the learning.

How do I get engagement? The people who know how to engage kids more than anybody are video game developers. They know how to engage kids: TikTok and video games, right? Back to the things, that is absolutely a scalable way.

When I talk about this tablet, I absolutely believe that there will be video games on that tablet that are totally engaging to kids, that they will go through, and that will teach them everything they need to know. When Timeback launches, there'll be a section where we have a team of people using AI to build video games on top of the Timeback engine that kids will sit and play.

The easiest way to think about this is, let's take math fluency. Let's take one example: schools have stopped teaching multiplication and division tables, as well as vocabulary lists. Memorization has gone out of style, which is really, really bad. It's really important that kids learn this, but people don't like the grind. The engagement is the hard part.

I just saw this last week: think of a game like Candy Crush.

Yeah.

Right, but it's teaching math fact fluency, and it's super fun, engaging, and high quality. Kids are basically just going to play this game and learn all their fast math facts, right? That fluency will do more to move up their math capability than anything you can imagine. It's obviously highly scalable, and you can get that out there through broad distribution.

Now, if you take that example and then just go large, you can imagine a great triple-A team—the best video game designers. We've invested over $100 million in getting some of the best video game designers in the world. They're down in Austin, and they're building a video game on top of the 2-hour learning engine that they believe will rival any video game on the planet.

The difference is that parents will be happy to let their kids play it and get their kids to the top 10%.

Yeah.

And I absolutely believe that. That's back to motivational models.

Yeah.

Absolutely. When we talk about the software side, right now it's all the physical-building side. Back to the scale, the Timeback team is like, “Okay, I've got to get the engine working.” 2026 is: How do I take parts of the motivational model that's only been done physically, in person, and get it into the software?

All of a sudden, you're like, okay, if the kids just voluntarily play a video game and they're going through it, you can imagine what that makes possible. Let me give you an example. I watched a test case where you ask a kid, “If I had a training center that would give you extra health, so it would increase your chances, how long would you be willing to go to the training center?”

The training center is just going to be math problems or some academic activity. These kids are like, “It's going to give me extra health, so I can increase the win. The more I do in the training center, the more health I get.” The kid's like, “I'd do 10 hours in that training session.”

I'm not a video game person, so maybe I'm going to get the wording wrong. I don't know how long these video games are, and they're like, “Oh, 10 or 20 minutes.” This kid literally just said, “I would do 10 hours of academics.” Now remember, it's only 20 to 30 hours to do an entire grade level.

There's all this motivational stuff. Back to it, I absolutely believe the next set of big video games are going to be built on an academic engine that's going to teach kids. The old ones aren't based on learning science and mastery standards, and when they're not engaging enough, the kids won't grind through them.

So the kid won't grind through it. Yeah.

Right. And you have to have that standard. You have to have both. You have to have a learning engine plus the world's best video game. Yeah, there's going to be a ton of video game developers.

Back to getting entrepreneurs: the old video games are too competitive. Come do educational ones, because let me tell you, financially, parents are willing to pay way more money. The ARPU on a video game that's getting a kid to the top 10% academically—are you kidding me?

Are you kidding me?

There's 10 million moms in America who would pay $100 a month for their kid to love academics and drive it up.

Yeah.

Back to product-market fit and deciding, one of the schools that was opened was a video game school. It was a middle school only to catch kids up, so they're ready for high school, and then you do video games in the afternoon.

Back to product marketing, you talk to them and they're like, “Wow, it really works, but there are parents who are like, ‘I'm not sending my kid to video game school.’” That's a hard, hard push. But if you just pivoted a little bit and said, “What if your kid came to you and said, ‘Mom, Dad, I want after-school tutoring?’” parents are like, “I'd love after-school tutoring.”

You can actually create—and there's one that's going to be created—tutoring and tournaments. After school, you're going to have tournament games, right? Video game tournaments. You do your 1 hour in the app, and then you get 1 hour of tournament gaming time.

We're actually working with some of the big streamers and video game players who are going to host these big events. Compare that with the problem today, where parents have to go push their kid for Kumon and do all that, the kids hate it, and it's all this point of confrontation, versus kids coming—and back to kids must love school—you could totally build an environment where they love it.

They're like, “Yes, I'm going to crank the app, especially because it's going to give me extra health in the video game that I'm going to go play. I'm going to play in this huge stream with all these other kids.” You can build those solutions.

It's so exciting for me. We've talked before about this, but this conversation really made it click for me how much Timeback is an enabling platform technology. If you get that 3-part engine right in the software, which you're working on and which is a tractable problem, a thousand flowers can bloom on top of it.

That makes me wonder—we spent all this time talking about all of that. I would love to just hear a little bit about you and your background. There's not a lot about you in the world. Most of the press that exists about you is from the '90s. At one point, you were the youngest self-made Forbes 400 member or whatever, and then you went into the wilderness. I don't know; you disappeared.

10. Joe's Background & Trilogy

Tell us your life story. I would just love to hear your personal background, and then we'll dip into a couple of topic areas.

I grew up in a business-oriented environment. My dad worked for GE. I was learning what gross margin was when I was 12 years old, and things like that.

Back to my schooling, I didn't like school. I'd get in trouble. I almost got suspended and kicked out of high school. I didn't graduate college because I'm like, “What am I? I can't sit. I just can't sit through these classes.”

Now, one of the things that's really important is academics is one piece of a very complex bundle.

And Stanford is an awesome school even if you don’t like the lectures.

Yeah. Right, because everything else around it—when I was at Stanford, Steve Jobs is talking. One of my best friends, Julie Zalinsky, was like, “Oh, I’m going to go work at NeXT with Steve Jobs and learn all this stuff.” Bill Gates comes down, or Scott McNealy, and you sit and learn from it all, whether it’s your classmates or just the environment in general, all this awesomeness. It’s important in education to understand it’s a complex bundle, not just who does the best academically, even at a school like Stanford.

My daughter’s there now as a freshman, and she’s like, “I am going to start skipping class.” She hasn’t had a teacher in 10 years, right? She hasn’t had a teacher in 10 years, and she’s the one who loves academics. She always looked down on me: “You would skip classes, Dad?” She’s very much that rule follower.

Now that she’s there, she’s like, “Oh, wow. Math Academy is a way better math app than my math lecture, and I’ll just use that.” But she loves Stanford and everything else around it. So school’s a bundle. I dropped out and started Trilogy. In the ’90s, we were the rocket ship in the ’90s.

Trilogy started in ’89.

So we started officially in ’89. I would have been class of ’90 if I had stayed and graduated. We were the rocket ship in the ’90s.

What was Trilogy at its inception?

The core product we built was an AI product—an expert system. One of the core products of an expert system back then was a configurator, which is putting things together. If you go to a Dell website or a car website and do a configuration, now imagine it for a phone switch, a Boeing airplane, or medical equipment—things that are super complicated, where you have millions and millions of combinatorics.

And if you sell it wrong—

—you only find it out on the manufacturing floor, and then it costs you millions of dollars to fix. Basically, what our system would do—it was called SalesBuilder—was take all the knowledge that was in engineering and manufacturing and put it on a laptop. Laptops were new back then, in 1990. It would put it on a laptop so you could sell things that would work and be eligible, and it would save companies millions of dollars a day.

And so they’d give us millions and millions of dollars for all that stuff. That was our first product. We then built around—we had sales stuff, so we built other things around that suite. After 2000, we started buying companies, not just builds. The ’90s were all organic builds. Then 2000 was, “Let’s acquire software.”

And that’s Chapter 2?

ESW, and that’s continued right through it. It’s a private company down in Austin, generates lots of cash flow. It’s why you can take $1 billion and say, “I’m going to go spend it on education.”

3 years ago, I went to my team, who’d been with me for 25 years, and I was like, “You’ve been telling me for 25 years that you could run this company better than me. Let’s see. Right. Good luck. I’m going to go be a principal.” And I said, “By the way, I’m taking $1 billion so I can figure out and get that going.” Now I’ve turned from boss to shareholder.

Yeah.

Literally, I’m the worst shareholder in the world now. I feel like if I was—I feel so sorry for her, because I’m like, “Guys, I need all this capital. I need more cash.” No, this expansion that Mackenzie wants to do is going to take $300 million, so she has to raise money or we’re going to have to give it to her. Generate more cash flow, guys.

And all these software companies you’ve got, you’ve got to figure out how to be more capital-light. To their credit—and I’ll give them a callout—I think on Bill Gurley’s last one—Bill Gurley, by the way, lives right next to our high school—he said, “I’d really love to see a private equity firm that is taking advantage of the 1,000 dead unicorns.”

That very much is what Trilogy and ESW—the team—is doing. There are all these SaaS companies that had loaded on a ton of debt in a growth play, and they’re gone. Basically, their equity is zero, and they’re being handed back to the private credit guys. The private credit guys are like, “What do you want me to do with this?”

When I put the pressure on them—“I’m going to suck a ton of capital and cash flow out of you guys”—they’re like, “How do we become capital-light?” They’ve signed a few of these deals already, and we’ll see if it really takes off, but they’re going to all the private credit guys and saying, “We’ll handle this offer. We’ll—hey, we’ll take this whole company. Give us this company for $1.”

Yeah.

ESW and Trilogy are known as being the best operators in the space, right? They can generate more cash than anybody else. Some of the big names have sold them a few companies, basically for $1. They’re going to split the cash flows and see how they go. The point is, there are $20 billion of those companies sort of out there on the books, and I’m super—good luck, guys. Go freaking crush that for me, because I’ll take the cash.

If you think about the lessons, especially from ESW—I’ll talk about Trilogy too—I’m always interested in software: how incredibly hard it is to build software that anyone cares about, pays for, and uses, and then how surprisingly efficiently it can be run once you’ve achieved product-market fit.

If you take the Elon example, every time I talk to a software CEO in private, they’re like, “Can you believe this guy cut 70 or 80% of the jobs?” And things seem totally fine. There is just a massive efficiency gain in software systems that are at maturity or past product-market fit, and ESW fundamentally was about, “How do I take something that’s working but wildly inefficiently run and make it massively more efficient and therefore more cash-flow generative?” Is that a fair summary?

Yeah, growth is expensive.

Yeah.

One way to do it is, if you’re trying to be a hypergrowth company, your cost structure and your culture—it’s just very expensive. When you say, “Okay, I don’t need that kind of growth,” that’s why everything becomes that much cheaper.

I guess there are 2 dimensions, because the Elon example brings it up: first, just growth. Back to these deals, you’re the private credit guy. The management team of that failed unicorn has been spending hundreds of millions of dollars to be worth billions and billions. That’s their culture and attitude. If you go in and say, “Dude, you’re wasting some money here,” they’re like, “What are you talking about?”

“We have built a culture around all of this. If I change it, I’ll lose all the employees.”

When you go in, you have to be able to say, “We don’t care if they go. Look at Elon Musk. Elon’s firing them.” But you fundamentally have to be like, “My team can run this company.”

Yeah.

Right. And that’s back to Trilogy and ESW: their capability is, they’re like, “Yeah, we can run this.”

That old team that wants growth, they should go on. Companies have life cycles, and you’re like, “You guys are the growth team. You should go get your next growth.” There’s an AI startup that you could go raise a ton of money for and go play—do your growth playbook. But this particular company, wherever it is, it’s not a growth play anymore, and so you should run it not like a growth play. Yes, Trilogy and ESW are very good at that, and you can generate tons of cash flow on it.

If you had to name the 3 things that are most responsible for this—okay, you get a new business at ESW, either for $1 or for more than $1. $1 sounds like a good deal. And then, a year later, however long it takes, it’s now way more profitable than it was before. What are typically the 1, 2, or 3 things that you’re doing to make that possible?

The number 1 is growth versus installed base. That’s the way to think about it. In a growth company, you care about the new customers, and that’s all you care about. In an installed-base company, you care about your existing customers.

When our team buys a company, the first thing they do is call all the customers, every single one. You’ll be shocked at these multimillion-dollar customers who are like, “Wow, it’s the first time I’ve gotten a phone call from this company,” because management of a growth company—it’s all on the prospects.

“Hey, customer success,” right? The mantra at Trilogy and ESW is customer success: “Hey, what do we have to do to make you successful? Why don’t we set up quarterly calls where you tell us if we’re making you successful and what the joint metrics are?” And they’re like, “Wow, this is the first time anybody said that.”

When you do that, all of a sudden, you get retention. So retention of the installed base—that’s actually how you grow the installed base—means they’re going to sit there and tell you what they want and how to do it. Sometimes you can do it; sometimes there are upsell opportunities and all of that as well, where you’re like, “Great, I’m solving the customer’s problem,” and that is actually leading to more.

But you have to sit there and look at your installed base and be like, “This is it. You don’t get more.” I don’t get more. So I really need to treat these guys really well, and I really need to understand what they want. I need to figure out, hey, if I do add AI to this product, is there an upsell there? Is there something that’s really valuable?

Or, oh my God, the prior guys have ignored you, and it’s a total nightmare. You’re trying to cancel, and it’s hard to get off enterprise software. So you’re 2 years into your 3-year transition plan. Can I reverse that? Can I provide something where, all of a sudden, I care about you, you’re going to stop migrating off and get back on? So that’s 100% of what the playbook is: just saying—

Is the idea—the objective function—actual cash flows that could be pulled out of the business versus enterprise value? Is that the same as the switch from growth to installed base? Like, look, this thing is a melting ice cube—that would always be the term of art in this space. So it’s like, keep it cold for longer, basically, and take the cash flows. And if the terminal value is zero, fine. I don’t care.

Correct. You’re looking at what your DCF is. You’re pulling the cash out, and if it goes to zero—

Right. They’re built to last.

Yeah. There’s a cycle in a company where it’s built to die.

Yeah.

Right. Right. You’re just like, “This product is done.”

Yeah.

Right. And what you’re just doing is managing—you’re pulling the cash out to do it. There is investment you make because that extends it. Software still is one where, if you focus on the customer—I mean, just basics—you can extend the area under the curve.

Absolutely.

Then you get some nice things like an AI uplift and things like that that allow you to do it.

What was the transition like for you, going from this extremely young, extremely successful person who was on the cover of magazines and stuff like that in the ’90s to choosing a much more private, quieter existence?

That was all done basically because I got married. So in the ’90s, crazy Trilogy, super awesome, right? We’re on all the covers. We’re growing 100%. It’s totally awesome. Actually, my last podcast was 20 years ago. Coincidentally, 20 years ago I got married, and so I just decided I wanted a different path for the family and the kids. That was a different stage.

Now, why I’m here is, well, my youngest is now going to be a senior, right? And so she’s like, “Okay, back to trying to give your kids as normal a world as you can.” I just felt that being under the radar and more private was better.

Now, as I look forward to my future, I’m like, I’ve got 20 years, and I need to be out there because we have to educate parents. The phase that I’m in very much is—we talk about 2 things. We talk about how our schools educate kids, and at the same time, we educate parents. Our marketing department slogan is, “We educate parents,” right? Because parents’ view of education is 200 years of all this old stuff, right? And they just think things that aren’t true. It’s not true anymore. So educating parents about what is possible—that’s sort of my new role. I have to go out and do that.

What lessons did you learn from Trilogy University? I think it’s so interesting that that used to exist back in the Trilogy days, and now here we are again. Tell us what that was and what you learned from it.

Sure. We started in Silicon Valley, dropped out of Stanford, literally living in the garage. We couldn’t raise venture capital. It was different back then. We thought our ideas were terrible. One of our co-founders was like, “If I’m going to be poor and have no money and live in a garage, I’m going to do it in Austin, Texas, because it’s way cheaper than Palo Alto.”

And so he moved to Austin, and we all followed him. He was our best engineer, developer guy, the visionary behind the engine itself. The problem with Austin, Texas, in 1992, when we moved there, is that there were very few software engineers. So what we had to do was build the software group down there.

What we did was go to all the best schools in the country and hire 2,000 Ivy League grads—Stanford, MIT, Princeton, wherever—and bring them to Austin. And back to the culture, we brought them in, and it’s an interesting thing about high standards. These are the best kids in the country, coming from the best schools, and we’re like, “If you come down to Austin, we’re going to put you in Trilogy University, and this will be the hardest 100 days you’ve ever done. You are going to work 100-hour weeks. It’s going to be 8:00 a.m. to midnight, and we are going to build the most awesome products.”

It’s going to be the hardest thing you did, but it’s going to be great, right? And that was why we were able to recruit everybody: because of the high standards. It was actually interesting. Microsoft at the time—our claim to fame was that we got all the best students. At the time, I think Microsoft was sort of at the peak of their power. In ’97 and ’98, 90% of students would go to Microsoft versus any competing company. Against Trilogy, it was like Microsoft only got 30%, and we won 70%.

Part of the reason was that we made it harder. Kids want to do awesome stuff, and graduates want to do awesome stuff. You don’t want to give them a cushy job or all this stuff. You want to put them on the hardest problems, where they can work like crazy and do it. And so that was Trilogy University. That ethos is something we continue today.

There’s a program in Austin that’s running right now called Gauntlet AI because they’re trying to figure out AI training. They’re bringing down 100 people into Austin. They basically scour the world, looking for the top 1% or 2% of kids. These are 20-somethings, and they’re bringing them down to Austin and doing 100 days of 100-hour weeks to learn AI, build apps with AI, and retool their whole careers to be like, “Okay, what am I going to be in the future?”

They have it set up so that it’s actually free for people to go through. If you make it through—and there’s a 50%-plus washout rate—they then get a $200,000 job because there are all these companies lining up, being like, “Oh my God, these people just worked 100 hours a week. They’re leading-edge AI. I can’t wait to hire them.”

It’s a unique thing because, when it was started, people were like, “No one’s going to go sign up to do that.”

Yeah.

But if you actually talk to the Gauntlet grads, they’re like, “This transformed my life. This is the best 10 weeks—100-hour weeks—I’ve ever done.”

If I was to watch a day in your life, from the time of your podcast 20 years ago through today, with the focus on Trilogy and ESW, what was your life like? Were you just a normal CEO running a software holding company, basically? What was it like? You strike me as a person who’s relentlessly pushing to make things better. Where would I find you pushing most in those days? What were the parts that mattered back then?

You’re trying to figure out—it’s always back to capital—the trade-off between how much do I have to pay for this company?

Mhm.

Right. And the more I pay, the better-caliber asset I get.

Yeah.

Right, versus, okay, what’s the cash flow I can generate out of it?

And then what are the things that generate the cash flow? So first, you’re like, okay, what are the innovations we can make in customer interaction and delivering customer success, and how do we innovate there?

The second part is, okay, how do we take this old codebase and make the product better? And how do we make sure that we’re able to take these and revitalize them? How do I go get people who care about this? And how do I find them?

So it’s just looking at what are the parts of the company that drive your returns. Fundamentally, back to systems, you’re like, you have to have these scalable systems. We’ve bought over 100 companies. Anything we’re doing, you have to be able to say, “Okay, can I apply this to 100 companies?”

If you had a fresh billion dollars, with a new young version of yourself in a new LLC wrapper or something, or corporate wrapper, and you wanted to invest it in software, how much worse do you think the returns you would earn in the next 20 years would be from applying a similar playbook to what you’ve applied over the last 20 years? How much worse is the future?

Well, it’s 2 parts.

When we talk about SaaS versus AI-driven companies, the SaaS stuff is going to be way worse. Just pure transactional SaaS, I believe, is going to be dramatically worse than the big winners—the big AI-native ones. It depends on whether you call AI part of SaaS, AI-native companies, and call them software companies or whatever. AI-native companies are going to be worth more than any SaaS company ever has been, because they can just do so much more for the customer.

Yeah.

It’s just the breadth of what you get to do. You could fundamentally change how the business works. Take education: you’re literally like, “Okay, kids could learn 10 times faster.” It’s a multitrillion-dollar industry globally, and no other software that’s out there—no edtech, nothing that’s ever been built—can have kids learn 10 times faster. You’re like, “Okay, I have to build a really good product, and now kids could learn 10 times faster.” The value that it will drive to the economy is just going to be huge.

Can you tell me what you’ve learned from Jim Abel?

11. Lessons from Mentors

He was my head of HR. He came from the old school. He was from Bristol-Myers Squibb. I think I hired him in the early 2000s. Jim Abel was one of the most influential guys in my career. He was awesome, and he still is awesome on advice. It’s ironic: if I had to say what the best business advice I got was, it came from Jim.

First of all, software businesses are HR businesses, right? They are HR businesses, and it’s your biggest cost structure. It’s your biggest leverage for getting a return. It’s all that matters, right? He was an awesome HR guy, and I was a terrible HR person. I still am. He would just sit there and be like, “Let me explain why you’re the world’s worst boss or manager, or whatever category you wanted.” He did more to help educate me and grow me than anybody. We could sit and I could give you 20 stories.

Give me 1 or 2.

I’ll give you 1. Here’s the first one I remember. He comes in, and we’re doing our annual meeting. I used to write a strategy document, and it would be 20 pages: “Here’s our company’s strategy. Here we go.” There were 1,500 people in a big room, and I was going to get up and talk about this strategy.

Jim was like, “What are you doing? This is the dumbest thing ever. Here’s what you’ve got, Joe: You’re allowed 3 lines, 3 words, and you need to distill your strategy down to that,” because 20-page strategy documents are useless. I remember sitting there thinking, “God, does Jim understand that we hire all these guys from the Ivy League? Everybody can read a 20-page strategy document. This is slogans, right? I’d be embarrassed to get up and slogan my way through.”

He was like, “Joe, smart-person disease. You have smart-person disease. You need to craft your message and make it simple and clear.”

Yeah.

I was like, “Whatever,” and I ignored him. Then he went and interviewed every single person in the company. It took him basically months. He came back, pulled me into his office, and showed me on the whiteboard. He was like, “This is what people actually think your strategy is, because I just asked them all.”

Yeah, because you said something.

I was like, “Wow, that’s literally the opposite of our strategy.” He was like, “Right.” And he was like, “Yes. In a 20-page strategy document, there’s always 1 line that someone can attach to and say, ‘See, I’m aligned.’”

Yeah.

He was like, “You need to go create your strategy where there’s no escape, where your message is so simple that everybody can sit there and read it. If they feel discomfort because they’re not aligned, that’s what you’re looking for.”

Right?

That lesson—you can feel it in the Alpha School presentation.

Kids must love school, right? The 3 commitments.

Yeah, kids must love school. He’s always about the edge. He’s like, “And how are you going to measure that, Joe?” And he’s like, “Kids instead of vacation.” Jim Abel would definitely pass that. There’s no escape from that, right? If you’re a guide, if you’re going to be hired to be in charge, you sit there as a teacher and you’re like, “I need to make the kid love school. He has to crush the academics and do the life skills.” There’s no escape from that. If you don’t think that’s possible, you won’t take the job.

Last week, a friend and I went looking for new companies that we might be interested in investing in. We went through 500 company websites blind; we didn’t look at who their investors were. They were just companies that had gotten some funding.

We went through 500 websites, and we didn’t find a single one where it was totally clear and obvious what the damn product did for the customer. I think if everyone in the world did the 3-line, 3-words-each exercise, the world would be a far better place.

It absolutely would. He’s doing it from corporate strategy, but then once you do it, you’re like, “Okay, we’ll just communicate everywhere.” It’s just true. It’s true everywhere. If you communicate that out, that’s good marketing, right?

He always talked—he’s an HR guy, but his marketing insights came through Pepsi and stuff. He was a great consumer marketer as well. He’d just sit there and be like, “It needs to be a magnet.”

Good marketing needs to attract and repel.

You need to go find who your target audience is. If you’re giving platitudes, then that’s not repelling anybody; that’s terrible marketing. He was like, “That’s terrible vision for a company.” He’d look at most of them and say, “What are the 5 things?” He was like, “If someone won’t say the opposite, it shouldn’t be on the list,” because then it’s a platitude.

A controversial one when we were back then was integrity, because everybody has integrity on their list. Of course you want that. He was like, “It’s a waste word, because no one’s going to put on their corporate core values, ‘I want a company with no integrity.’” Nobody says that. So, yes, you have to have it.

I love that—the opposite. If it’s not the opposite and you have to have it, it’s a magnet: attract and repel. If you won’t say the opposite—if there’s not somebody else who will say the opposite—then it’s wasted, right? You’re not getting the clear message across.

Is there anyone in addition to Jim that taught you something as valuable that you remember from your business career?

Jack Welch was a mentor of mine. My dad worked for him, and then my dad passed when I was 25. Jack called me and was just like, “I’m here—anytime, any coaching or help I can give.”

This was during the ’90s. I would fly up and spend the day with him at GE or wherever. He was awesome. His ability to motivate you—and, back to limitless, his ability to motivate you—was phenomenal.

How did he do it? What did you take from him?

I’ll give you an example. I was flying up there, and I’m the youngest member of the Forbes 400. I’m on the cover of the magazine. I’m freaking 20-something. Life is awesome. I’m awesome, right? That whole attitude.

I go up there and I’m like, “Yeah, look at this.” We talk, and literally at the end of the lunch he’s like, “Okay, so Joe, you’re like a good product manager? Is that how I should think about you? Like a level 7 or whatever GE had?”

First of all, he was just brutally honest with the feedback: “Let’s talk about your strengths and weaknesses here, Joe. You sort of suck at this. Not just sort of suck—you’re terrible.” But then he would take that and make sure you understood where you were, and motivate you.

I remember I flew back from that trip and thought, “I have so much to do. I have all these things I need to do. I have to get better on this dimension, and my company has to get better here, and I have to build a better product. I need to learn more.” He would both inspire you to just incredible greatness with the highest standards possible around it.

I’ll give you one more related example. GE Medical had bought 1 of our products in 1997, and I was up talking to him. He was like, “GE Medical doesn’t like your product.”

I was like, “I don’t know if you know this, but it just won E-commerce Product of the Year—in the year of e-commerce, in 1997. It freaking won E-commerce Product of the Year.”

Jack was like, “Joe, I don’t care if it comes from heaven above. If General Electric doesn’t get a return on its investment, your product sucks. You suck. Your company sucks.”

You go from, “Oh my God, I won Product of the Year,” to, “He’s making a really good point about return.” If you want to sell to the big companies, you have to show return. That helped us create our customer success and figure out how we could show return, so we wouldn’t just be a product company but could show value to the customers. That was what he definitely did.

I love that. I’m curious about your motivation.

We talk so much about other people's motivation. What is the throughline of all your hard work? One of the things that comes up a lot when you ask people who know you decently well, I think, is that they'll say some version of this: Joe's 9/10 just a really nice, high-energy, normal guy, and then he's 1/10 business-killer psycho. I'm curious about that 10% of you, and maybe for you to describe it—if that resonates with you or not, and if so, where it comes from.

I'm just a normal guy with this extreme. You should ask my daughters. They probably have a good view on that. But there is an element of pushing the limit, right? That's where you get all the alpha, right? That's where you get all the benefit. It's at that extreme where the real truths come out, right? I absolutely believe I'm always trying to figure out how do you take this to its limit and turn it to 11.

Yeah.

Right. You're like, “Let's work 100 hours a week. Let's figure out how we're going to get all these kids right.” We're going to build the craziest boot camp children's university.

Back to what's hard about educational change, parents want it the same way. On my first day—I literally sat in front of the parents. It was actually the first day of school, which was 2 weeks after I started—and I was like, “We're not going to have any adults doing academic teaching in this school.” There were hundreds of parents who were just like, “This dude's a whack job,” right? I appreciate all of those who stayed with us. Some did rage-quit, obviously, and totally understandably. To me, I'm like, “Yeah, well, this is how we're going to make it work, and it's going to be fine. Of course there aren't going to be teachers doing academic teaching, because the kids can do it better.”

That's where you get the—okay, that seems like that 10% of just pushing to it. Today, our 2-hour learning engine—2X in 2 hours—you're just going to go, “Well, there's no reason why not to, because the kids can get 100 facts and ideas an hour in their head with the TikTok interface. Whatever this is, well, let's just do more.”

12. Life Skills & Afternoon Workshops

Generally, I think, on the scale, when you say I'm the 10%, Elon Musk is always a thing. You go to dinner with him, right, and he's like 100X. I'm like, “Oh, I really suck. I'm terrible, right?” He's like, “I don't understand why you don't have 1,000 schools already. I don't know what's going on.” I'm like, “Well, what about all these issues?” And so it's all a continuum, right, on where you sit.

But yes, that part of it is designing the system and designing the ecosystem. Whether it's a product or whatever, you're like, “What are those things? What are the levers that, if you really push them, that's what keeps going?” So learning sort of happens at the limit, would be the cap.

It's the Spinal Tap approach to product building: turn it to 11. When was the first time you experienced that? I'm so curious about its roots in you. Were you like that as a kid? What was the first time you remember pushing something to 11 to see what you could learn at the limit?

In high school. This was a dysfunctional 11, I guess. I was always thinking, “What's the minimum academic score I can get and still hit my goals?”

So I would sit there and be like, “Well, if I get an 89.5, that'll round to 90, which gives me an A, which is the same as getting 99 or 100.” I literally would just do the work to get an 89.5, right at the limit. There was a time when I literally turned in a test without finishing all the questions because I'm like, “Oh, it's enough now. I hit my 89.5,” and turned it in. The teacher's just like, “God, I hate you, Joe. Really? You can't just finish it?”

So it was the limit side. That would be the I-didn't-like-school side, and I was trying to figure out how to game it. How many days could I skip school and not get expelled? When I really turned it on was when I dropped out and started Trilogy.

My parents were very unhappy that I dropped out of school. My dad was definitely more of the rule-follower type, and I was too much of the renegade. My mom was very upset. There's that Forbes article where they're like, “You're a Moron.” It's a famous title. This was a little over a year after I'd started Trilogy.

I was with my dad, and my dad was like, “I know I didn't like school. I'm so excited that you started Trilogy because I was so worried that you were a lazy little shit. I was so worried about it, and it was just because you were in academics and you didn't care, and so you were always trying to shirk and do the minimum.” He's like, “You know, we're working 100 hours a week. We're sleeping under the desk. We're just doing everything right.”

He saw that I found what I loved, right? That was that unlock—you want to say, that feeling of limitlessness—and I would just—we would run through walls to do that. For him, that was the point. And so that would be the first time when I really demonstrated that, at least to him.

I agree that every kid has a thirst for learning. It's just not always academic, right? You're just trying to tap into that, and for me it became starting a business and learning everything about it. I was one of the world's experts in configuration. I can't tell you why, but it was a means to an end.

I wouldn't have just done it. I'd sit in the Stanford computer science library, read—back then, you had to read papers—and all that stuff. It was finding that unlock, and then once I did it, you're like, it brings the energy and excitement to it.

What was the switchover from Trilogy to ESW? What was the first time that that new unlock happened?

Yeah, there really wasn't. To me, it's all Trilogy. There was a name issue. There's a corporate legal entity, but it's all the same Trilogy. And so there really wasn't. We bought a couple of companies, and we're like, “Oh, we really are good operators.”

Yeah.

Right. We're like, “Oh, wow, look at this. We bought a company and we made, I don't know, 10 times our investment on it.” You're like, “Oh, wow. This is cool.”

Yeah, this is really cool. It's so interesting, the limitless thing. Maybe it's a nice place to start to wrap up.

It's the same person I mentioned earlier, my friend whose name is Graham Duncan. He had this idea from a teenager selling sweatshirts who was so motivated by it—an amazing kid. He has this notion of trying to get people into positive feedback loops. And it strikes me that that is what you are doing now with your time, your money, and your attention.

I haven't quoted numbers, but it's an enormous amount of your own personal money that you're putting behind this. It's pretty amazing for me as a dad—and I'm sure for everyone listening with kids—to think about your time, your energy, and your money. Basically, what you're doing is putting kids into positive feedback loops where they can feel that feeling of limitlessness, which, for anyone who's felt that, you don't have to talk about it. They know that is the biggest unlock of their lives.

I'm very thankful that you're doing what you're doing. I think it's one of the coolest projects happening in America that people are just starting to learn about. Hopefully now a lot more will learn about it.

13. The Kindest Thing

My traditional closing question: What is the kindest thing that anyone's ever done for you?

I've had so many. My wife runs a kindness campaign around all this stuff, and the kind world that we live in is just—when you look at the kids at the school, and you just sit there and say, “When I wake up every day, the kids who are in the schools, right, we are forging new ground.” This is a startup school, and there are all these kids who buy in and are like, “We are part of it.”

To a large extent, they're putting their faith in the adults—and that includes, obviously, me—to be like, “This seems really crazy, right? This seems very different, but I'm all in and I buy in.” That's the part that, for me, I'm just like—you know, at the end of the day, it all comes back to the kids and their buy-in around this stuff and their willingness to be part of it.

That is the thing when I think of gratitude and kindness. That's the core that makes this all happen.

And that's where my appreciation goes—immense amounts to all of them. There are a few of them watching this who are going to be like, “It’s time to receive feedback, Mr. Liemandt. Here is my critical yet constructive feedback.” Alex or Tatum or Ella is going to do that.

That’s the part of this whole thing when you think about it. And to your point, there’s nothing more important for a society than to raise kids.

We have an opportunity, because of things that have been invented and where we are in the world—

To totally make it awesome, right, for the kids. And there is nothing—not just more valuable—there is nothing more rewarding. Right. And so, to the point of, if I can assemble the dads and the moms who wake up every day and are like, “I want this change for the kids,” we have the opportunity. We need to get all together and make this available to 1 billion kids, billion-plus kids, around the world.

I’m in, right? I’m in for the next decade, 2 decades, 20 years. I’m putting in the first billion, right? Transforming this is going to require more than that. It’s a trillion dollars a year, right? It’s going to require more than that. But there’s nothing more important than all of us working on it.

Joe, this is unbelievably fun for me. Thank you so much for breaking your 20-year streak with us here. Can’t wait for everyone to hear. Thanks so much for your time.

Thank you.