Moonshots #233: Joe Lamont Interview Transcript — How Alpha Schools Uses AI to Teach 200% Faster

Guest: Joe Lamont — CEO of ESW Capital/Trilogy, founder of Alpha Schools Host: Peter Diamandis — Founder of XPRIZE, Abundance360 Show: Moonshots with Peter Diamandis #233 Duration: 94 minutes Source: YouTube Analysis: Deep Analysis & Commentary

1. Opening & Guest Introductions

Time: 00:00:54 - 00:03:20

Peter: Ladies and gentlemen, everybody, welcome to Moonshots. Today we’re going to be speaking about the future of education and the role of AI in education. And this is not just for K through 12 university. It’s also for reskilling — something that this rapid pace of AI is going to require every major company to start doing soon enough.

This is a topic that we’ve been very outspoken about here on Moonshots. Specifically how secondary and post-secondary education is failing us. It’s failing our kids. It’s not preparing them for the near term, let alone the medium term horizon. We’re going to be diving into a new education model called Alpha Schools. I’m a fan — it’s gotten rave reviews. And we’ll discuss the pros and cons of AI-driven education.

Today we’re joined by the leadership of Alpha Schools — Mackenzie Price and Joe Lamont. Mackenzie is the co-founder and CEO of Alpha Schools. It’s a K through 12 program built around a two-hour learning model that she helped to pioneer. Prior to Alpha Schools, Price was for 17 years in the lending business and also working at Trilogy. A Stanford grad with a degree in psychology, she began questioning traditional schooling when she saw her own kid begin to lose enthusiasm for learning. I’m also proud that Mackenzie is an Abundance360 member.

Peter: And Joe Lamont is the founder and CEO of ESW Capital and Trilogy, a billionaire entrepreneur who is actively playing a central role in funding and scaling Alpha Schools and playing the role as the principal, which I find incredible. He’s built his fortune by buying and systematizing software companies, a playbook he’s now applying to education.

First off, welcome. And let me start by saying thank you for your service, because you don’t see a lot of highly successful people turning back to work on K through 12 education. There’s no real big money in it, and it’s purely a work of good in the world. So thank you so much for working on it.

Joe: Yeah, for sure.

2. The State of Education: Shocking Data

Time: 00:03:21 - 00:08:26

Peter: What I’d like to do is actually kick off by sharing a few slides that show the state of education in America today. And it’s not that great.

This is unfortunate and very shocking — US high schoolers at their historic lows in reading, math and science. Only 35% of seniors in high school, 12th graders, are at or above the reading proficiency level. This is down from 40% in 1992. Only 22% of seniors are math proficient, and only 31% are proficient in science. And this is going into a world where math and science and reading is so fundamentally critical for us.

On the flip side of secondary education going towards college, here are some other numbers that are seriously troubling. Americans today perceive college as far less important. Back in 2010, 75% of Americans thought that college was important for your kids. That has dropped down from 75% to 35%. And as a result of this, we’re beginning to see colleges going bankrupt. At the same time, tuition has grown 893% since 1980. As we’re demonetizing and democratizing everything else, the cost of going to college is going up. And then here’s the killer — today, college graduates are the group that are out of jobs the longest. They’re the longest being unemployed. And so for me, that just cannot — that’s not sustainable.

Joe: I don’t know where to even start, other than it is a very disappointing picture. And half of seniors in high school today graduate knowing as much math as a 99th percentile third grader. And that’s a pretty troubling statistic. But I think it really points to the fact that schools aren’t preparing people for the near future, and more importantly, what this world is going to look like with AI.

Mackenzie: And it’s even more critical that in addition to making sure that young people have academic knowledge in their brains to be successful and be great critical thinkers, we have to instill all of the other life skills that are going to be needed so that they can be adaptable in this new world and be able to keep up. And not just keep up, but stay on the frontier of what human knowledge is required.

Parents are coming to us with their little children saying, “What in the world should school look like in this day and age? What is it supposed to be delivering 10, 15 years from now?” And that’s the challenge that Joe and I are excited every day to take on.

Peter: The data shows that at Alpha School you’re able to deliver a full academic program with only two hours a day of academic education. You’re educating your students at two and a half to three times faster and still scoring in the top 1% to 2% of the country. On average SAT scores, Alpha School graduates are at 1410 compared to the national average at 1024, roughly 400 points lower. That puts Alpha School graduates in the top 94th percentile.

Mackenzie: Actually, Peter, I’ll correct you on that. 1410 is actually the average for our entire high school student body, freshmen through seniors. Our average SAT score for our seniors is 1535. And the reason I think that’s important to say isn’t just to brag on how awesome our students are, but really to show and highlight how easy it is to excel when students get a mastery-based tutoring experience, which is what has been made possible in this model of education. We believe any student is capable of being 99th percentile when it comes to reading and writing and math and science and being able to crush their APs.

And of course, the beautiful part of it is it doesn’t take over all day, chained to a desk, plus hours of homework at night in order to get those results. We’ve really figured out how to use technology to deliver these really incredible learning results and do it in a fraction of the time.

3. Learning Science: We’ve Known the Answer for 40 Years

Time: 00:08:26 - 00:11:03

Joe: Back to scores and academics. As the principal, I just did the meeting this week where they were taking some practice tests. The freshman’s goal this year is to have a minimum of 1410 as their SAT score as freshmen. So we have some who are already over 1500 as freshmen, which — think about how much time that opens up for the rest of what they want to do during high school.

But aside from that, how do we do it? We’ve known for 40 years how kids could learn two, five, or 10 times faster. There’s a whole field of learning science that very few people know of, but it’s been around for 40 years. The graduate schools of education of Stanford or Oxford or Harvard have been studying how kids learn and developing papers.

Back when I was in high school, Bloom’s Two Sigma is a seminal paper where the paper talked about how all kids could learn and get to the top 2% — which you’re seeing in Alpha’s data, 40 years later. The reason it hasn’t been implemented is if you read the papers — there’s probably 10,000 of them published over the last 40 years — they all say kids can learn this fast. Unfortunately, it doesn’t work in a teacher-in-front-of-a-classroom model.

One of the things I realized after I became principal three years ago was every parent wants their kid educated the way they were, and we don’t know any different. And when I sit in the room with some of these learning scientists, they will say, “We’ve known for 40 years — teacher in front of the classroom is the worst way to teach kids.” And I’m like, whoa, whoa, whoa — we like the people in this room, right? Every parent in America thinks a teacher in front of the classroom is the best way to teach kids.

And it’s that dynamic, fundamentally, that we have to educate parents on and show that this new way that is grounded in learning science — we’ve known how this works. We know everything from the best ways to teach to the chemistry of the mind to be able to do this. But for the parents, it’s what they knew. It’s how they grew up. So you get this cascading generational effect.

4. The Five Dimensions of a 10x Better School

Time: 00:11:04 - 00:18:45

Peter: Joe, I’ve heard you talk about what you call the five dimensions that make a school 10 times better. If you could recount that — because I found the idea of being able to make something 10 times better is something you typically don’t think about in terms of schools.

Joe: Back to when Mackenzie and I were working on this — if you sit down and you say, the school’s been the same for over 100 years, and if we’re going to recreate school that’s going to get our kids ready for the AI age going forward, first principles, what would you do? Just ground up — what are the core things you’d really want to do?

The first and most important, which is the most radical thing we say, is kids must love school. If we’re going to put our kids in it for a decade plus, kids should love school. And we actually measure it — over 90% of our kids love school. We measure every week. And over the last year, we actually upped it where we ask, “Would you rather go to school or go on vacation?” And 40 to 60% of our students say they’d rather go to school than go on vacation.

The highlight of my tenure as principal was this — before summer, six months ago, two thirds of the Alpha High students sent an email and said, “We don’t want to take summer break. Can you keep the school open?” You’d never catch me in high school asking if I could go to summer school.

When we all build businesses, you want to build an organization that people love to come to work for. Yet those same people, when they sit and look at their kids and say what school is supposed to be like — you’re not supposed to love school, it’s supposed to suck. School is spinach. We have this mental model that for everything else in our life, we’re like, “Yeah, you totally should build an organization that people love,” but with school, we just had this backward way of looking at it.

The second dimension is kids should learn 10 times faster. We sat down and built this engine over the last three years where we basically took all the learning science and we took AI tutors — not chatbots. If you take AI, generative AI and you take learning science and you generate personalized lesson plans for every kid at whatever level they are, those kids can learn 10 times faster.

Peter: Did you build or buy that tech?

Joe: We started by using off-the-shelf apps and then building around it — this AI wrapper around everything. We now have both the foundation, this platform called Time Back, and we use some third-party apps that are best in class. For any of you listening, Math Academy is the best app — you can just go download it. We put in over $100 million into this platform. What we have today is a learning engine that teaches kids basically K through 12 at 10 times faster than if they sat in class.

The third dimension — you come into school for two hours, and because we learn over 2x in 2 hours, what do you do the rest of the day? You teach them all these valuable life skills. We have our Alpha Life Skills curriculum: leadership and teamwork, storytelling and public speaking, entrepreneurship and financial literacy, socialization and relationship building, and grit and hard work. In the afternoon, you’re getting all these killer workshops — team-based, super fun workshops run by adults teaching them all of those life skills.

We have fifth graders running food trucks and running Airbnbs for their entrepreneur workshop. We have an Alpha High student who actually today submitted her research to Nature, which no high school student in history has actually been published in Nature.

The fourth dimension is we have transformed the role of the adult in the classroom. Instead of having to spend their time creating lesson plans and delivering lectures and grading papers and homework — which we don’t even have homework, by the way — our teachers, we call them guides, their sole job is focus on motivational and emotional support. Basically holding kids to really high standards, saying “you are capable, you can do big impossible things,” and providing the support and the mentorship and the coaching that these kids need.

The fifth pillar is character and culture and classmates. Who are these kids being raised to be and who are they surrounded by?

5. Inside Alpha: What a Day Looks Like

Time: 00:18:45 - 00:24:39

Peter: Mackenzie, can you dive a layer deeper for moms and dads and kids watching this — you arrive at the school and what’s the experience?

Mackenzie: It’s such a great question because people just don’t understand what a school looks like where kids actually cry at the end of the week saying, “Why do we have to have breaks on the weekend? Why can’t we go to school seven days a week?”

I think people often think, “Okay, some AI school — it’s like robo terminator, dystopian, anti-socialized children.” And nothing could be further from the truth. In fact, the model of the day is one where kids get so much more time to interact with each other and to interact with the adults in the school building.

You walk into school — first of all, it feels more like a WeWork space than a traditional classroom.

A day starts with our kids coming together for what we call a Limitless Launch. Think Tony Robbins for kids — this is a chance to do some sort of challenging, often physical experience together. It’s when we start incorporating growth mindset strategies and practices that they take into the rest of the day.

Then they go into their two-hour core learning block. This is the academic period of the day, and this is when kids are on computers. The key difference here — kids are getting this one-to-one interaction on the computer that’s meeting them at the level and pace where they’re challenged, in what we call the zone of proximal development, where lessons aren’t too hard that they turn off and not so easy that they’re bored and disengaged. They take breaks every 30 minutes, get time to stretch and play.

And by lunchtime, they’re finished with their academics and the rest of the school day is these super fun workshops. That’s everything from five-year-olds learning how to climb a 40-foot rock wall or swim in the deep end of the pool, to high school students that are out building real businesses and doing big things.

6. Reimagining the Physical Space of School

Time: 00:24:39 - 00:30:38

Dave: I want to visualize even better this idea of a WeWork-looking school, because it just jumps out at me immediately. You see all the little chairs lined up with the little folding desk and the lecture in the front and rows and rows of these rooms — all wasted space. Now you’re saying, no, no, no, get rid of all that. But it’s not like our office. What does it look like?

Mackenzie: Picture a really comfortable environment. I was in our kindergarten and first grade classroom at one of our schools this morning. You’ve got some kids who are at a standup desk because they realize they like being able to just move their feet as they’re doing their work. You’ve got another group of kids sitting at a table with a guide right next to them, working together.

And my favorite — I always see this when I walk into one of those classrooms — you’ll see some kid who’s laying on a bean bag with his feet up against the wall and he’s got his computer on his lap. You might look at it and go, “What’s that kid doing?” And I’ll tell you — he knows that he can get his work done and he can hit his goal. He’s kind of earned that privilege of being able to relax.

Motivation is such a huge part of our culture at Alpha. We’ve built in all these different motivation models — everything from earning the privilege of getting to do your core work in a little tent fort or a cubby thing with stuffed animals surrounding you. It’s just kind of fun.

Joe: Our schools — we’ve opened schools all over the country and we’ve started them small. We’ll have schools with 25, 50 kids. You get a playground — kids need to be outside. In fact, our students get an hour and a half of just non-structured outside playtime every day.

And the same is true with the way we do our workshops. I was at our middle school this week and they’re doing an incredible band music class, but they’re creating a rock band. As part of that, they’re also learning about marketing and branding and mixing music in addition to playing instruments. You’re not really going to walk into one of our schools and see kids all playing the recorder as part of their traditional music class.

Salim: When they reach college — I teach at MIT and a little bit at Harvard and at Stanford — the kids that go to school in England arrive a full year ahead of the US-educated kids. They can use that to take basically either a year of tuition back in college, or they can start a company while taking a semester or two off. If you start talking about 10xing the rate of learning, and 1400 SAT scores of freshmen, you’ve got to figure you’re in that same range — maybe even a year ahead of that by the time you get to college.

Joe: Two points. About your England comment — we have what we consider the world’s best team of learning scientists who work for us. Most of them come from England rather than the US. In England, the academics have been going up versus the US going down. A lot of it’s these people who have been for 15 years making sure those concepts were put into the curriculum. We’re now putting it into our platform so everybody can get it.

In our high school, they spend four years working on a passion project. We have one of our seniors who’s putting on a Broadway musical — it’s going to be the first all-teen, all-produced musical. She spent everything from “I have to source the singers off TikTok” to music rights, to how do you fill a whole venue, and all the sales and marketing. She took her love of arts and theater and has had years to actually work on it.

7. “Guides” Not “Teachers”: How to Find the Right People

Time: 00:30:39 - 00:36:04

Peter: I’m curious how you source your guides. Super excited that Alpha is going to be opening up a school in my hometown in Santa Monica. How do you go about sourcing your guides? What’s their background? They’re just not traditional teachers? Were they the top teacher someplace else?

Joe: The biggest issue, back to scaling schools — schools don’t scale because you’re limited by teachers. That’s always the critical aspect. The five things required for a great teacher in a traditional school: you have to be a domain expert, you have to be good actually at teaching kids, you have to connect and motivate students, you have to be good with parents, and fifth, administration.

Outside of the business world, you ask an HR guy 101, he’s going to say, “That’s a complicated spec. Those are different skills. It’s hard to have all five skills in one person.” In education, the way we solve that problem is to underpay them. And then we’re like, “Oh, my God, no one applies for this job.” It’s the hardest spec in the world, and we’re going to underpay them.

If you’re going to rebuild this, any HR guy is like, “Well, simplify the spec or pay more or both.” And so at Alpha, that’s what we’ve done. Those first two items — subject-matter expertise and how to teach it — we don’t need because the AI tutor and our Time Back software does that. The third dimension — you have to connect and motivate kids — we push that to the max.

Mackenzie: Just about half of our guides come from a traditional teaching background. These are generally people who were phenomenal, loved working with kids, but realized that the traditional system is built to fail. We have a guide who was the Principal of the Year in the state of Florida in 2020. After 14 years working as a principal, she was attracted to come to Alpha because her desire was to get back in the classroom and impact and interact with kids.

The majority of our guides come from different backgrounds — they wouldn’t have been interested in a traditional teaching world, but they were experts in motivation. They come from professional and college-level coaching and athletics. They come from corporate backgrounds, executive backgrounds. What they all have in common is this idea of wanting to positively impact and motivate and connect with kids and serve as a mentor.

And to Joe’s point, we’ve taken the opposite strategy as traditional education. We believe that anybody who is going to put their blood, sweat and tears into working with this next generation, they deserve to be paid well. So our guides start at six-figure salaries. And that’s part of what has enabled us to really attract incredibly talented people.

8. 80,000 Applicants: Diverse Backgrounds & the Transition from Traditional Teaching

Time: 00:36:04 - 00:46:13

Joe: Just to put some numbers behind it — we had 80,000 people apply for our teaching roles. At the kindergarten level, we actually have 5 to 6-to-1 ratio of student-to-teacher at Alpha. The reason for that is that’s the one place where our apps aren’t fully developed — bootstrapping kids into reading. Once you can read, I can teach you math, I can teach you all the other subjects. But the detection of a four-year-old — the AIs aren’t good enough yet. We think we need another 18 months before AIs can do that.

Mackenzie: Let me add a mom’s perspective. In addition to reading, think about kindergarten — we want our kids to feel loved and welcomed and we want kids who come into school curious and excited to maintain that. Our kindergartners learn how to ride a bike and do a five-mile bike race as a team together.

I was just talking to one of our guides who was talking about a kid who was really feeling challenged around riding a bike. But this guide knows that he loves animals, so she has a stuffed kangaroo and she runs alongside of him and says, “Can you go faster than the kangaroo?”

Dave: When you were talking about specialists — the other fundamental skill we have to teach young people is the ability of learning how to learn something. When you talk about kids building businesses in fourth grade, doing jewelry-making businesses, or building mountain bike parks or Broadway plays or mental health apps — how in the world would you find a specialist who’s an expert in that as a teacher? Instead, our guides have the skill of coming alongside each other and figuring out how to become an expert in that field, how to connect you with experts in the field.

Joe: As you get older, we increase the ratio. By the time you’re adolescent — “Whoa, too many adults around here, give me some space.” On the business model side, we take that money out of the guides and we put it into these workshops. We have ex-athletes — I have ex-NBA assistant coaches. Jermaine O’Neal is teaching a middle-high school program. We have ex-Olympians — sets of coaches and athletes who are able to connect with these kids and provide that motivational emotional support. And that’s a lot of the magic that makes this work — making sure you hire those, which is why we pay more, why we built our business model around doing it and source 80,000 of them to find the right ones.

Dave: That gives you a real head start on what’s coming too. The whole college curriculum at MIT moved online years ago, and it got off to a fast start, but then it completely fizzled. The reason it fizzled is because it’s so insanely boring to learn on your own through a browser. But what’s coming next is the interactive AI version of that. It’s going to have exactly your favorite personality on the other side. The AI version can be any voice, any personality — it could be Shaquille O’Neal reincarnated as a math teacher. It’s going to be so incredibly engaging, and it’s not quite ready yet — it’ll be there within a year.

Joe: We have one where our middle schoolers are allowed to pick their avatars. In AP World History, they’re all like, “Okay, Genghis Khan is going to be my avatar.” And it yells at them if they’re not studying. The kids love it.

Peter: When you hire someone from the traditional side, there’s obviously a transition period. What’s the average length of time?

Mackenzie: The last step in our hiring process — when they’ve gone through interviews and tests, everything from cognitive aptitude tests to case studies — the last thing they do is spend time in one of our schools shadowing and working. What’s really interesting is that the most common reason for falling out and not getting hired is actually traditional teachers. The reason is they are not willing to jump in and engage with the kids, get at their level. Traditional teachers are so used to standing forward or standing back and sort of pointing, “Do this, do that.”

But for those traditional teachers who do make it through, we’ve got great training. The big thing is helping them remember that their job is no longer to teach academics. You’re not going to see one of our guides sitting down and showing a kid how to carry the one. Instead, they’re going to be helping kids develop that skill of learning how to learn.

Today in that kindergarten classroom, I was watching a couple of kids — when they would hit one of their academic goals, they would give a secret signal — touch your nose. And that would signal “I hit one of my academic goals.” And then that starts — get ready for it — a silent dance party. A five-second dance party where the guides would dance and the kid would do it. And that was enough to get that kid psyched up and ready to say, “Okay, let me go hit my next goal.”

You can walk in a classroom and have two 7-year-old boys sitting right next to each other, but each of them is working on something that’s exactly at their pace. One 7-year-old doing algebra while the kid right next to him is working on his multiplication tables. That’s the beauty of personalized learning — it’s not isolating, it’s not making that kid march up two grades to sit with older kids.

9. Mastery-Based vs. Time-Based Learning

Time: 00:46:13 - 00:51:39

Dave: Everything I see in schools — you either care about your grades or you don’t. And there’s no room in between. If you’re near the top of the class, you get into this group of people that are really competitive about their GPA, and those are the ones that end up going to MIT and Harvard. And then you’re like, “I can’t compete on that, so I’m completely competing on the basketball team.” It’s really sad because most of these kids would be great at both. But the peer pressure just buckets everybody in very irrational ways. It must be completely different in your school.

Joe: On the academic side, this is the difference between a time-based system and a mastery-based system. All our kids master their grade-level material — they get hundreds on their standardized tests for each level.

If you ask Alpha parents, “Can your kid get 100 on the Texas STAAR test?” — fewer than 10% of the parents are like, “Yeah, of course.” If you ask the Alpha students, 95% say you can get 100 — because they have. In a mastery-based system, you don’t move up to the next grade level until you master what you’re on. The tutor is going to sit there — the system is going to give you the personalized lessons so you can.

In a time-based system, the two things that determine academic success are IQ-coded and Big Five conscientiousness-coded — are you a grinder? Those are the people who do naturally well. That’s 10% of the market. 90% of kids don’t. In a mastery-based system, everybody can.

If you ask your average Alpha student — it’s not about IQ, it’s about effort. And that’s why we call our software Time Back, where it literally tells the kid, “You’re 17 hours away. You’re five hours away.” So it moves everything from “I’m not smart enough” or “I’m not capable” — which is so debilitating — to “Okay, I’m five hours away from mastering it and getting 100.” And that’s just a huge unlock around self-confidence and resilience.

Peter: I want to say this for everybody listening. The traditional school models — you start school at 100% score and every time you get something wrong, your score goes down. In the traditional video game models, you start at zero score and every time you learn something, your score goes up. It’s a completely different motivational system.

Joe: Think about this engine. Everybody’s like, “Oh, these AI tutors and they’re terrible.” But imagine generating personal lessons for every kid that’s based on: What are you trying to teach them? What is their knowledge graph — what do they know and not know? What is their interest graph — Taylor Swift, Genghis Khan? And then you can actually take into account cognitive load theory — the chemistry of your brain, how many working memory slots you have, how many reps you require to store the long-term memory.

The AI can generate a personal lesson that makes sure it’s at 80 to 85% accuracy — not too easy, not too hard. At 99%, you’re not learning — you already know it. And every game designer will tell you if you drop below 66%, the kids disengage because it’s too hard. So you can imagine this unending stream of personalized content for every kid right at their zone. This is why you can learn 10 times faster.

10. The Good and Bad of Screen Time

Time: 00:51:40 - 00:56:35

Peter: Let me ask some of the difficult questions. You’re asking parents to trust more screens in the classroom at the same time that there’s a backlash going on for screens with kids. What do you tell parents who don’t want their kids on computers all the time? And are you able to enforce a distraction-free use of their laptop?

Joe: Those are super valid concerns. Screen time and bad screen time, even AI and bad AI — these are all super legit concerns. Here’s how we look at it: We believe there’s good screen time and bad screen time. If you can learn 10 times faster, and it frees up the rest of the day to do all these great workshops and life skills, it’s worth that trade-off — even if you don’t like screens.

There are parents at Alpha who say, “I’m a no-screen-time household.” And they’re like, “But if my kid can actually learn the academics in two hours instead of six and get those afternoons to do all these great things, that trade-off is worth it.”

Dimension B — if you give kids ChatGPT in a school, 90% of them use it to cheat. Chatbots are cheat-bots, and so unmanaged AI is bad. In our mornings, we don’t have chat functionality enabled. Even in our world we’ve tried — even our best kids, they all use it to cheat.

But what we tell the kids is: if you’re using chat in the morning, you’re probably cheating. But if you’re not using it in the afternoon on these life skill workshops, you’re probably failing. The student who’s trying to do a musical needs help. So it’s that dual sense.

For putting facts, ideas and concepts in kids’ brains — ChatGPT steals that away from the kids and you want to fill kids with ideas. Our AI use has two dimensions. One is generate the personalized lesson — the generative part of AI generates a dynamic lesson. The second part is the vision models. It is watching the screen, coaching you.

We basically stream the screen to a model. We spend like 10 grand a kid right now in AI tokens, because the AI is watching the screen saying, “Look kid, you’re scrolling, you’re guessing the answers” or “When you miss the question, you’re not listening to the explanation and you’re just trying to drive the next question.” It’s coaching them into good self-driven learning habits.

This is why EdTech isn’t the solution. We talk about EdTech being like 10% of the answer. Everybody says to educate a kid you need a motivated student and you need to put them in lessons of the correct difficulty. EdTech does the second really well. But how do you do the motivated student part? For us it’s give the kids their time back. If you’re not using the apps correctly — skipping or guessing — you’re not going to finish in the two hours. We literally have a waste meter in the corner — “Dude, you’re wasting 50% of your time.”

Dave: That’s mind blowing. Those vision models only came into existence since maybe a year ago and you’ve already got them deployed. 10 grand a kid is a lot of token usage though — that’ll come down a factor of 10.

11. AI Monitoring & the Closed-Loop Data Cycle

Time: 00:56:35 - 01:01:40

Joe: We’re going to get it down to on-device. But to be honest, three years ago when we started, we literally had humans reviewing the video at night, annotating — “This is good learning. This is bad learning.” You have to train a model — it’s just like Tesla autopilot. You have to train this model of: “Okay, this is good learning, this is bad learning. What are these kids doing? Oh wait, they’re switching to another screen to ask ChatGPT. Oh wait, they’re playing Minecraft.”

Our unit of measure is XP. One XP is one minute of focused learning. The kids know the AI is deciding — “Was this a minute of focused learning? Okay, you get credit for it.”

Dave: Just to make an analogy — if you’ve ever coached soccer or football or tennis, imagine you could look at every single kid and study every footstep they take and say, “That step to the right should have been a step to the left.” But you’re in their AirPods telling them one-on-one coaching of every single move — that team will just take off. And the vision technology to actually do that through AI, that’s brand new in the world. It would work in coaching just as well as it works in screen monitoring for teaching.

I can just completely see how every single action in the learning experience is now a measurable piece of data that can be personally analyzed. It’s brilliant that you actually had human beings trying to do that at night — that must have been the most laborious thing in the history of the world. But that’s how you get a head start into knowing how you’re going to deploy it with AI.

Joe: Back to the data — what we have now, and why our learning keeps increasing, is we’re the only ones in the world with a closed-loop data cycle. It’s reinforcement learning for kids. Our learning science team comes up with an idea, they put it in, generate the lessons, the kids do the lessons, you measure the learning — did this work better? And then they can see the scores based on standardized test results and adapt.

A concrete example: in August, I stood in front of all the parents and said, “We got a new math curriculum, K through 12. Kids are going to learn more in 20% less time. We’ve made improvements.” Eight weeks later, at the end of the first session: “Good news and bad news — 4th through 12th, we’re crushing it. We’re above our thresholds, beating the 20% metric. K through 3, we’re not. The kids are getting lost in the path — we gave them too much freedom. So we’re changing that for the next eight weeks.” And now, we’re back on track.

Dave: Peter and I went to 1X Robotics to interview Bert Bornick, and he had the exact same thing for robots in the home. They gather every single motion, all the data comes back, gets transmitted into a nightly learning process that retrains the model. You’ve got the exact same thing in education.

Mackenzie: Think about how much time is spent studying game film in sports to see how you can improve. That’s never been done when it comes to learning in academics. It was just — hopefully you’re good at school or you’re not. We actually give kids coaching where they learn that the best way to go forward is actually to slow down, read the explanation, watch the video, and then take your time to look at the question before you go answer it.

That’s part of what really puts learning in the kids’ hands and gives them a sense of ownership over it. They start seeing, “I can get better. I have the agency and the ability to improve if I put in the work and the good practices.” That completely changes the way a kid sees themselves. That’s really the magic.

12. Resistance to Disruption & Micro-School Expansion

Time: 01:08:43 - 01:12:46

Salim: One of the things we see a lot is when you try anything disruptive in a legacy environment, the immune system attacks you. So are you getting attacked by the legacy system?

Joe: This is one of the things I’ve been trying to figure out — can you convert an existing school, or do you have to be a new school like Alpha? The answer is — if you hear this pitch and you like it, you come to the Alpha info session, which is why there’s a thousand people at New York’s Alpha info session. All the people who don’t like this message stay at your school.

There are two types of people — there’s a set of people who say, “I want it, but somebody else has to be first.” Everybody wants to come to an Alpha. But as we open up all these new cities — once Alpha has a hundred people, then everybody wants to come. Nobody wants to be the first 20. That’s the key.

Peter: Let’s talk about how you get your schools open.

Mackenzie: In general, we’re calling these micro-schools — about 25 students is the start. One of the advantages if you’re a founding family is you can actually help create the opening class. You can get your kids’ best friends in there. I think half of education is who you go through the educational process with — are they people holding you up or pulling you down?

Joe: This is the Mackenzie story. When she was co-founding Alpha, she woke up and said, “I want to go find the 20 people who are going to surround my kids.” And I was actually a laggard — it took Mackenzie two years to convince me to come to Alpha. I went to Catholic school, didn’t really like it, almost got expelled because I skipped school all the time, dropped out of college. But my kids were going to Catholic school.

But in every city, you need the same thing. The core of growth is going to a city and finding those first families who are like, “Yes, I’m in.” And then they go recruit the other families. I always say the two things parents need to move a kid from their school: they need a reference from a parent they trust in their city, and they need to see a kid at the school do something their kid can’t do. Whether it’s love of school or academics or whatever they care about — they have to see those two things. And when they have the two things, then they’ll move.

13. Business Model & the Challenges of Scaling

Time: 01:12:47 - 01:20:37

Salim: Do you use a franchise model or do you own all your own schools?

Joe: For Alpha, it’s 100% us. We run it. There’s no franchise. We’re 100% in charge of quality control. But Alpha’s the high-end private school — we designed this when we said, “What are we doing? Price is no object.” It’s expensive. It’s the high-end private school.

With this Time Back platform we’re building, we’re building a whole set of other schools based on it — different models. We launched Texas Sports Academy — the afternoon is all sports. We’ve launched a Gifted & Talented school for the kids who love academics — when you ask what would make them love more school, they say “more academics.” Those kids are learning at 5x — they’ll be 1500 by 8th grade. We have a wilderness school, a Montessori school. You can imagine all these different models — as soon as you have this magic of two-hour learning, you can start to say, “What should we do in the afternoon?” Next year in 2026, you’re going to see all these other models beyond just Alpha that roll out based on this Time Back software.

Peter: What is the reaction of the authorities and government departments of education?

Joe: I’m going to be doing this for 20 years. Mackenzie’s in for 20 years. The public school side I think is going to be the second decade, not the first decade. It’s all private right now.

The Texas Sports Academy is using the Texas vouchers — the billion dollars they just allocated — where it’s $300 a month parent pay. So bringing the price way down from an Alpha.

Mackenzie: I applied for 10 charters in 10 states for in-person charters. Zero — rejected 10 times. We got Arizona — one virtual, not a physical school. The school boards literally say things like, “I was put on this earth to stop people like you.” That’s what I talk about when I say the immune system.

Joe: The private school market’s a $50 billion market. It’s a big market. For those parents who want to opt in, they can. We’d love to move into public as soon as they want to invite us, but on charters, we just said — we know we’re not wanted here, so whenever you want us, come call us. But otherwise we’re done applying for charters.

Mackenzie: The Secretary of Education came and spent the day at Alpha in Texas in September. We get people who are inspired, but it’s really the idea of — can you really change the whole model of the day? We are probably the most talked-about school in the world because we’re serving as inspirational ideas for what’s possible.

Peter: Is everything perfect? What do you guys worry about? What keeps you up at night?

Joe: First — your average parent hearing this thinks this is totally wacky and must not be true. Educating parents that this new model is not fake and false and a hoax is a huge lift. Our marketing department is — our schools educate kids, we have to go educate parents.

Second — how do I scale while keeping quality? Every time we open a new location, every new guide — how are we making sure you’re getting that same level of love of school, academic performance, life skills? We spend a lot of time building this for scale.

Third — our cost. I’m burning 10 grand on the AI. I’ve got to drive that down. If somebody flipped a switch and every parent wanted it tomorrow, we couldn’t fulfill it. We’ve got to drive that curve down.

And then the biggest one — I’ve got 20 years. We’re going to build 10,000 schools. But how do I get to a billion kids? If motivation is the number one issue, the engine part’s easy in the scheme of things. But how do you motivate kids when you don’t own the school day? The biggest motivator of a kid is giving their time back. For the billion kids where I don’t control their school day, and they’re not going to restructure their school, I still have to get to those kids and motivate them.

So we’re partnering with some of the best motivators in the world. We have some of the world’s best video game designers — that’s going to be released in 2026. We work with some of the biggest influencers, packaging it all up — motivation plus the learning. That’s a big lift.

14. Pharmaceutical-Grade Trials: Convincing Governments

Time: 01:20:37 - 01:22:23

Salim: Have you talked to other governments? There’s certainly countries around the world that are much more open to the future of education in a modern way.

Joe: Yes, we get demand everywhere. On governments — it’s the same thing almost everywhere. If countries have a private model, you absolutely can get take rate on the private side. The issue is ministers of education or departments of education are pretty much globally going to wait and see until there’s dramatically more proof. And I’m not against that.

Peter Attia — his daughter went to the school. We were comparing medical-grade testing versus education. In education, we roll stuff out to 55 million kids and there’s no randomized control trial. There’s no million-user test of “Does this really work?”

One of the things we’d love to do is — as we get all this data out there with Time Back at scale — do a million-student randomized control trial. Pharmaceutical-grade trial. So everybody can be like, “This isn’t a hoax. This actually works.” We need some of that proof because a lot of things in education — there’s a lot of bad learning science we need to undo. I believe proof and pharmaceutical trials might be the way around that.

15. Facing the Future: Technological Change & Optimism

Time: 01:22:24 - 01:33:42

Dave: I would love to know how you deal with the rate of change of technology and what your process is. If you think about — you start with a batch of 6, 7, 8-year-olds and you’re planning that to end as 18-year-olds, but in the intervening 12 years, half the kids will have Neuralinks and the other half won’t. They’ll be doing something in the back of their mind you can’t even see. How are you going to deal with this constant rate of change?

Joe: We’re humble enough to know we don’t have all the answers. Mackenzie and I talk to thousands of kindergarten parents — the kindergarten parents are the ones who want the biggest change because they’re looking out at 12 years from now and they all know this current school system is not going to work. They’re motivated to change — versus if you have a senior in high school, you’re like, “Let me just get them to college.”

All we say at Alpha is we wake up every day and we feel we’re the leading school that is saying, “How do we adapt to what the technology changes?”

Our high school kids have a concept called a Brain Lift — the data structure for human knowledge. It literally divides: this is the stuff LLMs are going to be good at or are good at, and this is the part that humans are good at. And we’re teaching them — as of today and with what we can project in the future, this is as good as we can get, kids.

And maybe Grok 7 comes along and does everything the Brain Lift thought, and we’ve got to move on. But we believe there’s a — and I think this is an important point about our culture — Mackenzie’s best line is: “This is the best time in history to be a five-year-old.” AI is going to impact adults very differently. In Austin, we have Waymo robo-taxis — not a good time to be an Uber driver. But AI gives kids superpowers. It’s going to transform their childhood. It’s going to let them do things that our generation never even dreamed of.

I have a dozen high school kids in my guide crew group — Sparta group, shout out. They wake up every day and when you talk to them, they’re like, “Yeah, you adults — it’s wraps for you guys. AI is going to give us superpowers and we’re going to take over the world.”

All our kids read Rational Optimist by Matt Ridley. He actually came to the school and he was sort of an AI doomer when it came to education. I couldn’t have been more proud as our students sat there and used his arguments from his books and examples from what their day is like, on why he needs to see the world differently.

Mackenzie: We had a big foundation — a very large education philanthropy foundation — come and visit our schools. At the end, the head of the foundation said to my guides and me, “You guys are so happy.” And they said, “We go and visit schools and talk to educators, and they’re all just so encumbered and depressed and it’s so hard.” And I was so happy to see my guide say, “What do you mean? This is the greatest time in the world to be an educator! Because we finally are unleashed to do what we got into this world to do, which is positively impact kids.”

If you come to one of our schools at the beginning of the school year, we’re not going to just dust off the same old presentation. We’re more like the Apple iPhone — we update every six weeks. We’re constantly iterating, growing and learning, seeing what works and what doesn’t.

Salim: If you go back to your founding, what would you have done differently that would have accelerated things?

Mackenzie: To partner with Joe at the very beginning!

Joe: Three years ago, why I got involved — I put my kids in 10 years ago, but I said to Mackenzie, “Just like all education, this isn’t scalable. This is great for Austin, great for our families.” It wasn’t until generative AI came out when I said, “This is the technology that can let us scale to a billion kids.” So it really was that moment. And if you just rolled me back three years, what would I have done differently? Understanding and educating parents. Three years I didn’t do a podcast. Now I realize educating parents — Mackenzie’s been out there — we have to get out there. Because the single biggest impediment to making education great for our kids is what our parents believe.

Peter: I’ve got to say, I came into this podcast saying, “Well, I always learn something, but how big a deal can it be?” And I’m coming out of it saying — this is so inevitable and so huge. This has been great. Critical for America, critical for the world. The speed of change is so extraordinarily fast that if we are still using the old methodologies, we’re hitting brick walls. We need not just Alpha School — hopefully many different mechanisms that are going to enable us to have our kids, our college kids, and even reskilling — using this technology to reskill our employees is going to be fundamentally critical.

Thank you, Joe. Thank you, Mackenzie.

Mackenzie: You can go to alphaschool.com and learn more about our model. We’ve got Alpha Schools launching coast to coast — 13 new schools this year. And “Future of Education” on social media — I’ve got a podcast called Future of Education. Joe’s X handle is @JayLeemont, where you’ll see life as a principal and views around the business of education.

Families can absolutely come visit. Students can do a shadow day to see if the school is a fit. But we warn parents — don’t send your kid to a shadow day unless you are convinced that you want to move, or your kids will rebel. They’ll be like, “I’m not going back to my old school. I want to go to this new school.”

Transcribed by AssemblyAI, organized and edited by Claude Code video-to-article skill