Listen on Spotify

You know the kind of trainer an actor gets to prep them for the superhero role in a Marvel film? We're that, for math. If you want to understand what we’re doing but don’t have time to skim our 400+ page book, this episode sums it up in about an hour.
Here's a summary of what we covered:
[1:34] The big problem in math education is a lack of individualized instruction. In a classroom with one teacher teaching the same thing to all the students, it's way too easy for the top half of the class and way too hard for the bottom class. What we do is pinpoint the exact problem that each student should be working on right at this moment to make maximum progress in their math learning.
[4:46] So much difficulty in math learning can be traced back to missing prerequisite knowledge. That's why it's important to start each student off with a diagnostic that combs through many years of prerequisite knowledge that they need to know to succeed in their chosen course. If we find any knowledge gaps, we fill them in before asking the student to learn any more advanced material that depends on it.
[6:50] We get a very high-resolution picture of the student's knowledge profile by overlaying every question/answer event onto a structure called a "knowledge graph". The knowledge graph encodes all the dependency relationships between mathematical topics. We leverage it to squeeze a ton of information out of every single question that we ask the student -- not just figuring out what they know and don't know, but also figuring out exactly what learning tasks they should be working on to maximize their learning efficiency every step of the way.
[8:44] Elsewhere, lots of students struggle with calculus due to gaps in prerequisite knowledge. Good teachers know this, and try to fill those gaps, but there's a limit to how well the teacher can do this because all the students have knowledge gaps in different places and the teacher can only teach one thing at a time to all the students. But we can target these gaps precisely, backfill them, and move on based on what each individual student knows -- fully individualized instruction for all students in parallel, delivering exactly what they need to work on, focusing on those weaknesses and not wasting time on things they already know cold.
[12:05] If you have more talent/aptitude, then you're going to get more bang for your buck out of practice. You're going to require fewer reps before getting solid enough to move on, and you're going to generalize more naturally. However, of the students who get all the way up to calculus before struggle really sets in, the biggest roadblock is typically not talent/aptitude but rather gaps/weaknesses in prerequisite knowledge, an issue that can be resolved with fully individualized instruction.
[14:50] Math Academy origin story: Jason/Sandy coached their son's 4th grade math field day team, which turned into a pull-out class 3 days per week the following year, which -- after the superintendent came by and saw 5th graders doing trigonometry, advanced algebra, even a little bit of calculus -- turned into a pilot for the Math Academy school program in the Pasadena Unified School District, where students learned all of high school math (prealgebra through precalculus) during 6th/7th grade and took the AP Calculus BC exam in 8th.
Students were invited to the program by scoring in the top 7-8% of a middle school math placement that all students in the district took at the end of 5th grade -- and about two-thirds of students in the district were on free or reduced lunch. Also keep in mind that nearly half of Pasadena K-12 students are educated in private schools, compared to the California average of ~10%. Generally speaking, these were not smartest kids in California, and their parents were not Caltech professors.
Jason developed software to automate the process of assigning/grading homework, and together during the pandemic we upgraded it to figure out what each individual student should work on and teach it to them directly without any human intervention. We worked like maniacs to get it ready before school went fully remote the next year, and once we put the school program on it, educational outcomes (including AP Calculus BC scores) skyrocketed. Because of the software, our students experienced a massive learning GAIN, not a loss, during the pandemic. Naturally, it only made sense to keep the school program using the software even after class returned in-person.
[21:55] We have spent thousands and thousands of hours over the years building and fine-tuning our knowledge graph. It's not off-the-shelf, it's not automatically generated. It's the hard work from domain experts, primarily our director of content Alex Smith for the forwards graph (what are all the prerequisites you need to learn in order to unlock a topic) and myself for the backwards graph (when you practice a topic, what component sub-skills are implicitly getting practiced and to what extent).
[25:04] We analyze our knowledge graph by overlaying a big heatmap of where students are doing well or struggling at various parts in the graph. It's almost like traffic intersections in a city -- which ones are where most accidents happen? Let's go make those safer. We've been building and refining the knowledge graph for nearly a decade now with all these analytics.
[27:22] We have a wide variety of user segments. We can help anyone who seriously wants to learn math. Basically, anyone in any sort of educational situation, kids, adults, public school, private school, charter school, homeschool, grade school, high school, college, students who are accelerating, students who are just trying to keep up, "math team" people, people who don't yet think of themselves as "math people", adults changing careers to a math-ier field or pursuing a math-ier subfield within their current career, the list goes on and on.
[29:31] The best predictor of how long someone will use the system and how much math they'll learn is what kind of habit structure they have in place. Students who are consistent, as opposed to sporadic, go much further. It's that simple.
[34:13] The only math learner persona we can't help is the crammer -- the student who has an exam in a week, is nowhere near prepared, and wants a "quick fix". We are like a gym, and there's always people who walk in the gym and think they're going to work out really hard for a week and look like Thor by the weekend. There is no way to make that happen in a week, no matter how hard you work out. If you show up consistently, like 3-6 times per week for a 30-90 minute session, and then you keep that up for months, then you're going to come out looking like the mathematical equivalent of a Greek god. But if you are looking for some kind of easy, "how can I change my lief in one week," then I'm sorry, I don't know what to tell you.
[37:16] We alternate between minimum effective doses of text-based guided instruction followed by active problem-solving. It's the mathematical equivalent of a tennis instructor showing a quick demonstration of how to hit a ball, just for a minute, and then students practice hitting the ball with that technique until they're solid enough to move onto the next technique.
[40:16] Real-time reactions and hot takes: Jason on collaborating with school districts, my thoughts on the edtech industry, Jason founding a company with his wife, my experience interacting/growing on X, Jason's impression of Waymo, my impression of math textbooks, Jason's thoughts on the "move fast and break things" ethos, Justin's thoughts on people's screen time concerns.
[52:10] People say, "just give me the intuition." But intuition comes through repetition. That's how you get the automaticity, the natural feel, and that's what intuition is.
At the same time, it's important to be efficient. Don't work 100 problems of the same type in one day. Maybe do 10 to start, then 5 the next day, another 5 a week later, and so on, while you're filling the empty space with practice on a ton of other skills. You have to get your reps, but you also have to distribute them out over time. That's how you learn efficiently and build long-term retention.
When people want their math learning to be less skill-heavy and more concept-oriented, what they're often really saying is that they want a fast overview of a subject without going into the details, without really getting your reps on everything. A video that explains all of calculus in an hour, or how neural networks work in 20 minutes.
But what we're focused on is building up a true level of mastery. Not surface-level, not shallow. The optimization problem we're solving is NOT "how fast can we imbue you with a shallow level of understanding, enough that you can tell your friend something cool or that you think you have opinions about it." What we're focused on is how quickly we can get you to operating mathematically almost like a professional musician plays their instrument, or a professional athlete plays their sport.
[55:51] As a rule of thumb, if it wouldn't work in sports, it's not going to work in math.
[57:31] Students on our system typically learn about 3-4x as fast as a normal class. That's why, in our school program, the students could go from pre-algebra through AP Calculus BC in 3 years, from 6th-8th grade. When that first happened, and the Washington Post wrote articles about it, lots of people couldn't believe it. Which is why we had them take the AP Calculus BC exam so we actually have results.
[58:49] We hear all the time about students who are behind in their school class, and then use our system to catch up, and then start crushing their class, and then go well beyond their school class -- as well as the resulting change in the student's level of confidence. In just one year or less, just months, a student can go from thinking "I'm not a math person, I'll never be good at it" to "I'm crushing my school class, it's so easy." That change in the student's experience does wonders for their confidence.
[1:01:28] Is there an upper limit to how much math you can do per day and have it carry over into real learning results? Think about it like going to the gym. If you work out for 45 minutes, 5-6 days per week, you'll get in incredible shape. You can do more if you want, but there is a point where you hit diminishing returns. Whether it's Math Academy or the gym, it really comes down to how long you can sustain a productive full-intensity effort. It's hard to keep that up for multiple hours, though you might be able to get better mileage by splitting up a multi-hour session into a shorter morning session and evening session. But every person is kind of different in their breaking point, how much they can stay focused and work intensely on the system. In general, one hour per weekday is what we've found to be the upper end of a sustainable approach for most students.
[1:03:38] We make students do review problems indefinitely into the future, but with expanding intervals -- spaced repetition. It's the optimal way to keep your knowledge base fresh enough to keep building on it without constantly having to go back and re-learn things. But we make this review process as efficient as possible by tracking all the subskills that are implicitly reviewed when you do a review problem, and we're always trying to select tasks that kill many birds with one stone by exercising many subskills in need of review.
[1:08:41] Lots of people mistakenly think that students need a million different explanations of the same thing, and that one of those explanations is going to stick, and it's different for each student. But really, all you need is one really good explanation that's been battle-tested across a large number of students, and the students need to come into that explanation with all their prerequisite knowledge in place.
If you do that then you can get students learning the skills really well -- students pass our lessons over 95% of the time on the first attempt, and over 99% of the time within two attempts, without any additional remediation (because enough knowledge has consolidated into their brain from the first attempt that it makes it cognitively easier for them to get over the hump the second time around).
That's often surprising to people who think that every student needs a different explanation, but typically what they're seeing is a symptom of the student lacking prerequisite knowledge, and you're trying to come up with some explanation that allows them to grasp "enough" of the topic (not the whole thing) while at the same time not requiring too much in the way of prerequisite knowledge they're missing.
[1:11:08] What makes math hard is the same thing that makes climbing a mountain hard: the steepness of the gradient. What we do is break every steep section of math into smaller steps. If you break things into small enough steps, anyone can learn. And that's what we do with our analytics: where are the congestion points? Where are students struggling? It's always where we're trying to do too much at one time, so we break it up into more steps.
[1:12:25] It's so important to have a reliable source of truth about what a student really knows, and grades are no longer a good source of truth. You remove test scores from the admissions process, the last objective metric and the last Jenga block, and you get bad situations like at UCSD where 8% of students were not proficient in middle school math. So many issues in education stem from a student having a piece of paper that says they've learned something when they actually haven't.

Want to get notified about new posts? Join the mailing list and follow on X/Twitter.

Listen on Spotify

You know the kind of trainer an actor gets to prep them for the superhero role in a Marvel film? We’re that, for math.

If you want to understand what we’re doing but don’t have time to skim our 400+ page book, this episode sums it up in about an hour.

Huge thanks to Nat, Elli, and AEI for having us on!

The transcript below is provided with the following caveats:

1. There may be occasional typos and light rephrasings. Typos can be introduced by process of converting audio to a raw word-for-word transcript, and light rephrasings can be introduced by the process of smoothing out natural speech patterns to be more readable via text.
2. The transcript has been filtered to include my and Jason's responses only. I do not wish to infringe on an external speaker's content or quote them with the possibility of occasional typos and light rephrasings.

Jason: Thanks for having us.

Justin: Glad to be here.

Jason: Math Academy is an online adaptive math learning platform. Think of it like this: if you wanted to learn a math subject — a math course like algebra or calculus — you would sign up, and the system would give you an adaptive diagnostic, figure out what you know or don’t know, and create a custom course for you.

Whether you’re an adult student trying to learn an advanced subject or get back up to speed, or you’re a parent with a homeschool student or a gifted student who needs some additional enrichment or an alternative approach to learning math — that’s essentially what it is.

Jason: I think for something like Khan Academy, which is fantastic for getting up to speed on some topics — you need to learn a topic for your physics class, or you’re out sick for two weeks with a bad flu and you got behind — you don’t understand trigonometry, and you watch some videos and do some lessons or topics. Fantastic for that.

But if you wanted to learn an entire year of math, you probably wouldn’t have the level of a systematic approach and enough practice and rigor to really make that successful.

Math Academy is thinking of it this way: you would go with a personal trainer and say, I’m auditioning for a Marvel movie. I have to lose 30 pounds, get muscular, and I’ve got six months. The trainer’s like, right, we’re going to give you a full breakdown assessment, come up with a plan, do this every day, and make it highly efficient. So when you go audition for that Marvel film, you’re ready.

Justin: The big problem in our eyes is a lack of individualized instruction.

Everything in our system is designed to pinpoint the exact problem that the student should be working on right at this moment to make maximum progress in their math learning.

Typically, in a classroom, you have one teacher teaching 20 or 30 kids. They’re all doing the same work, and the problems they’re doing are way too easy for the top half of the class and way too hard for the bottom half.

Our goal is to identify the exact types of exercises that the student should be working on and serve that to the student.

Justin: Yes, exactly. First we need to know what you haven’t mastered, and then we’ll master that on the road forward — and that road forward extends a long way.

Justin: Exactly.

Whenever a student signs up for one of our courses, in addition to doing mastery learning to learn all the course content, we also look back many years at prerequisite knowledge. If somebody signs up for calculus, we’re going to be looking back: do you know how to complete the square? Do you know your trigonometry? Do you know how to work with logarithms, exponent rules?

All these things that students often come into calculus courses missing — these foundational skills — lead to a lot of pain and inefficiency when they’re trying to learn calculus.

That’s the diagnostic.

Jason: It’s actually not too bad. It probably takes anywhere from 30 to 45 minutes, maybe an hour in the most extreme cases.

But think about it like this. If you were a teacher and you had a student transfer in halfway through the year and had to assess them — you’d be asking: okay, what does this student know? How much of the algebra? Can you factor this polynomial? Maybe not. Can you solve this linear equation?

Each one gives you a good sense. If they can’t do that, maybe just try a few things. You start poking around. It’s like walking around a dark room and feeling your way. Okay, there’s a wall there.

You don’t have to do that forever — you’re accumulating evidence of things they don’t know and things they do know. Except that if a human was doing it, it would be pretty rough, low-resolution.

Now we just do it at very high resolution. The student might do 30 to 40 questions, maybe even less, and we get a very high-resolution picture of what we call their knowledge profile.

Justin: The diagnostic operates on top of this structure that we call a knowledge graph. The knowledge graph encodes all of these prerequisite relationships and encompassing relationships — what skills are used by what, and what skills use what sub-skills.

Using that information, we can actually get a lot of mileage out of every single question that we ask. If we ask, do you know how to factor a quadratic, and they don’t know how to do that, then you can reasonably infer there are a lot of things they don’t know — a lot of post-requisite skills that depend on that — that they don’t know.

Justin: Exactly.

In addition to reasoning about every single answer that a student submits on the system and providing remediation if necessary, we also have timed quizzes every 150 minutes or so of productive work on the system.

These are broad coverage quizzes over all the stuff the student has learned or should know up to that point. It really tests their ability — do you know this cold? Can you solve these problems quickly? Have you developed a high degree of automaticity?

If not, then that’s a perfect opportunity for us to deliver more practice to make them stronger in those areas.

Jason: My high school math teacher — he was a brilliant guy, actually a physicist — used to say, “The hardest thing about calculus is all the pre-calculus you forgot.”

To do calculus, you need to know your algebra, a lot about functions, a lot about trigonometry. If you’re taking more advanced calculus — parametric equations, polar equations — you need to know those too. Students typically forget or get rusty on this stuff.

A really good calculus teacher will say, “Anybody remember how to invert a function? Okay, we need to clean this up. We’re integrating, doing trig substitution — everybody remember cosine, sine?” Your brain is kind of fuzzy, right?

Really engaged, serious calculus teachers have to do a lot of cleanup because they know — but there’s only so much they can do. They have to teach calculus. They’ll say, “I know kids are weak in these things, so I try to supplement some additional stuff,” but they’ve got a lot of other material to cover. There’s only so much they can do.

Calculus is more cognitively taxing. It is more complex. A calculus problem may have many more steps than an algebra problem — instead of three or four steps, you might have ten. It’s more abstract. There’s more complexity to it — on top of the fact that a lot of students are building on an unstable foundation of missing prerequisite knowledge.

Jason: This is the logic, right? Yeah, 100%.

Let me just say one thing — maybe Justin will jump in.

Think about it this way: if you had two groups of students, and in one group, the students had a teacher for 30 students, and in the other, each student got their own private tutor — the parents of the students in the class would say, “Whoa, that’s not fair!” It’s not remotely fair, because everybody knows the private tutor is radically more effective. Any student falling behind would be told to hire a tutor — it’s so much more efficient than the class.

That’s exactly what Math Academy is like. Every student has a private tutor that delivers exactly what they need to work on, focusing on those weaknesses and not wasting time on things they already know.

That’s what makes it so powerful — every student has their own individual knowledge profile, and it looks different, and it’s constantly changing.

Jason: I’ll just say this — as things get cognitively more taxing, they’re going to require more effort. Some students are going to have a higher aptitude. It’s just going to come easier to them, just like playing soccer is going to come a little easier for some kids.

Everyone has their aptitude and their ability to learn something, and it may get to a point where the amount of effort it takes to succeed might be beyond what a student wants to do.

Can you take students who are in the bottom quartile of an algebra class and say, “We’re going to take you all the way through calculus”? You could do that. But it just may be beyond how much work those students want to do. They’re just like, “I don’t want to do this. I don’t care about calculus. I’m in the play — I care about Shakespeare,” or whatever.

It kind of comes down to how much effort students want to put in to learn. For other kids, it’s like, “Whatever, practice is not too hard, it’s fun.”

Justin: The way that we tend to think of aptitude or mathematical talent is as the flip side of how much effort a student is willing to put in.

If you have more talent, more aptitude, then you’re just going to get more bang for your buck out of all your practice. Even if you take two students and give them problems pinpointed for their individual levels, the higher aptitude student is going to require fewer reps. They’re going to be generalizing more naturally.

But the problem is that many students who enter calculus — or any high-level math course — struggle. I think the majority, really, who get to that level, who get to the point that this is actually a course they are taking, if they struggle, it’s typically due to not having foundational skills.

Of course, aptitude can play a role in that. If you have lower mathematical aptitude, you may need more reps on these foundational skills to really drill them in and make them automatic. But if you provide these kinds of students with that opportunity to fill in those skills, then yes, they can go much, much further.

Jason: My wife and I were sort of pulled into coaching a math team at the local middle school — fourth grade, actually. They were in fourth grade.

I tend to get a little carried away with things, and I maybe got a little more excited about doing this than was necessary. We were doing two or three days a week at lunch, and I had a lot of fun. The kids were having fun too.

It was kind of a combination of them wanting to know more and being surprised at things they didn’t know. They didn’t know what a negative number was. I was like, how do you not know what a negative number is? That shall not stand.

Then they’re like, what is pi? What? You don’t know pi? Oh, let’s go.

It was one of those things that was a lot of fun from both sides. Then I convinced the school to allow me to do a pull-up math enrichment course when they were in fifth grade. I took about 13 of them who were part of the original math team — three days a week — and that eventually became kind of a pilot for what became the Math Academy program in the Pasadena school district.

The superintendent had heard about our class. He came and saw it. He saw these fifth graders doing trigonometry, advanced algebra, even a little bit of what looked like calculus for fun.

He was blown away — could not believe what these kids were doing, how much fun they were having — and he said, would you help me create a pilot based on your model?

Over a couple years, working closely with the district and with a lot of support from the superintendent, we were able to create this pilot program within the school district.

Now, I’m a software developer, so to me every problem is ultimately a software problem. Instead of handing out homework and grading it, I was like, this is for the birds. I don’t want to do this every night.

I started writing software to make that a little more efficient, which over time evolved into the fully automated system that it is today.

Jason: I think it went pretty smoothly, actually.

The software had a lot of rough spots, and it took us a couple of years to really get it working the way we wanted it to, but the students were learning right away and making progress. The only problems were the bugs we had to fix in the system.

It wasn’t the approach — it was that the software was rough and early.

We followed a very standard, obvious approach. I want to teach you a topic: here’s a little introduction, here’s an example, let’s work some problems. You get a problem wrong — here’s an explanation, let’s do some more. Okay, you’ve got some of these correct — let’s move on to a slightly more difficult variation.

You kind of do that. Obviously, that kind of thing can work, as long as you’re not giving the students too much information to absorb, and you’re not making the problems too difficult.

That should work — and it did.

Justin: We actually found that the students did much better once they were working on the fully automated platform.

We introduced this mainly during the pandemic years when school was going fully remote. We were like, how are we going to keep these eighth graders taking and doing well on the AP Calculus BC exam? We’ve got to cover all of high school math from sixth grade to eighth grade, and they’ve got to do well on the exam. With remote instruction, that was going to be a very big challenge.

When we put the entire school program on the nascent platform we had built, instantly they started doing much better on our AP exams — in terms of just remembering stuff they were supposed to be reviewing but didn’t really get all their review reps in when it was just depending on human teachers.

I remember I used to do a lot of TA sessions for these students. Back in maybe 2019, there was a challenge that a lot of the students entering the high school program were not getting enough review on their foundational trig and calculus. They were learning linear algebra, multivariable calculus, abstract algebra — whatever — but they weren’t getting enough review.

That was another major thing I noticed. When we put everybody on the system, kids actually had these prerequisites really down cold.

Jason: The way students would be selected for the program is they would take a middle school math placement test at the end of fifth grade.

Students who scored around the top 7 or 8 percent were typically invited. The district would review their performance on the test and say, okay, as long as there are no major holes and they seem generally strong, there would be an invitation: “Do you want to do Math?”

We had three different middle schools. If you got a slot, you would start with pre-algebra, picking up right at the end of fifth grade math, and you would learn everything up through AP Calculus BC by the end of eighth grade.

These were on the stronger side — we’re talking 7 percent. We’re not talking 0.01 percent, like the smartest kids in the state of California. Two-thirds of the students, when we started the program, were on free or reduced lunch. This wasn’t like, “They’re all kids of Caltech professors.”

It really was about efficient, effective instruction and taking things very seriously.

We’ve all seen this in other areas — whether it’s music education, where you see kids who are 12, 13, 14 and unbelievably good at piano or violin — or in sports, where you see soccer or basketball teams with kids playing at a professional level.

If you take kids who have reasonably high aptitude and take it very seriously, it’s night and day compared to normal development.

Jason: I’ll start — Justin can jump in.

Initially, we were just thinking about how we have these courses — calculus, integrated math one, whatever — and we’re building these courses out. We’re breaking each individual lesson into what we would call a topic. A lesson would be something that might take ten minutes to learn — sort of a fairly granular level.

We were just building out these individual courses, and the courses would connect as prerequisite — one course to another. Algebra would go to geometry, then algebra two, and so on.

We were creating the lessons, creating the topics, and figuring out, okay, these topics connect to those topics, that kind of thing. It wasn’t like we sat there for three or four years in a vacuum with this abstract platonic ideal of what this is. It was like, okay, we need lessons that work.

What do they need to know that goes into this? Oh, kids are struggling with this lesson — what’s happening? You know what, we need to split this into two topics.

It’s really about how things work in reality.

Now, Justin — working with our Director of Content — spent a lot of time, thousands and thousands of hours over the years, fine-tuning and creating this knowledge graph. Justin, you might have more to say.

Justin: The knowledge graph is built entirely by hand. It’s not off the shelf, and it’s not automatically generated. It’s really just hard work from domain experts — primarily our Director of Content, Alex Smith, and myself.

I also work on encompassing relationships in the knowledge graph. We have a forwards and backwards knowledge graph. There’s the forward graph — what do you need to learn before learning this new topic — and the backward graph — when you practice this topic, how much are the various sub-skills getting practiced?

Are the component skills getting practiced 100% with any problem you do? Or is it maybe only half of the sub-skill that gets practiced? That sort of stuff.

This has been a very challenging and time-consuming process, but what we’ve found is that you really have to craft it well. You can ask an LLM to generate a knowledge graph for a course, but it’s not going to be crafted in a way that students can really learn from. It takes a lot of refinement.

In addition to creating the knowledge graph, we also analyze it. We overlay a big heat map of where students are doing well or struggling at various parts in the knowledge graph.

It’s almost like traffic intersections in a city. You can see where traffic is piling up — and we need to make that intersection flow better.

We’ve been doing this for nearly a decade now — just refining and refining and refining the knowledge graph with all these analytics.

Justin: We’re continually expanding the knowledge graph.

Right now we go from fourth grade up through university math. I think the most advanced courses we have right now are Methods of Proof, Probability and Statistics, and Discrete Math. We’ll have Differential Equations out soon — in about a month — which is part of your core engineering math.

We’ll be expanding further into Abstract Algebra and Real Analysis, as well as expanding the lower end of the graph into third grade, second grade, multiplication tables, that sort of stuff.

Whenever we put out a new course, we always take what we’ve learned in previous courses — how to scaffold things well, best practices, heuristics we’ve developed over the years — to make sure students can go through smoothly. But there’s always another level of refinement that happens as we get actual data from students going through the course.

For content we’ve had a long time, like Pre-Algebra and Algebra 1, the knowledge graph is quite stable. For new courses, we’re constantly refining it. We get it pretty well on the first version, but we’re always looking for ways to improve.

Jason: We have a variety of very distinct user segments.

You have a homeschool segment, which would seem sort of obvious. You want to homeschool your kid — you’re like, “Oh, I can’t wait to teach them history and chemistry,” and then, “Oh gosh, I’ve got to teach them algebra.” Math Academy is great for the homeschool situation.

You have parents of gifted students. These might be fourth or fifth graders, but a lot of times it’s more like pre-algebra or middle school when parents start realizing that whatever’s happening at school is nowhere near good enough. Their student is ready to do a lot more.

You also have parents coming in saying, “My student has struggled — they changed schools, the pandemic, not a very good teacher, whatever — and now we’ve got a disaster brewing. We’ve got to fix this.”

Then you have a lot of adults. We have a very substantial adult contingent. Some of them are going back to grad school. Some are in undergrad and feel like, “My teacher isn’t doing a great job — I need a better approach.”

You also have a lot of people learning on their own, especially in technology — people who want to get into fields like machine learning and artificial intelligence, which require a lot of math. They’re like, “I’m trying to read these papers on AI, but it’s been five, ten, fifteen, twenty years — I don’t remember this stuff, and some of it I never learned in the first place.”

Plus, you have charter and private schools that are dipping their toe in and experimenting.

This is happening all over the world — it’s not just in the U.S.

Justin: What we found is that stickiness goes hand in hand with how much structure the student has built up around using the platform.

Homeschool groups are using the platform as their primary math learning resource, and they’re doing this every day. The habit structure is very strong there.

Students who are using Math Academy outside of class are not quite as sticky as the students who are actually doing it in class. One thing we often recommend to students is that if they start using the system outside of class, they’re going to quickly get to a point where they are above grade level. The goal is for them to start using the system in school as their primary resource. Once you lock in that structure — this is what we’re doing every day — the student is much more likely to stick with it.

Adults, of course, are often more pressed for time and have a harder time setting up these habit structures. But we’ve found that adults who do manage to carve out a specific time of day — “I do Math Academy before work at 7:30 a.m. each morning and I do this amount” — they go much further.

People who use it sporadically, almost like a crossword puzzle or just whenever they feel like it, are not going to get as far.

Jason: It’s not going to take them as far.

Jason: It’s $50 a month. But we also have a volume discount structure for schools that have hundreds or thousands of students.

Jason: You come in and there will be a list of learning tasks. We have up to five learning tasks you can do. Typically, there will be fewer. Sometimes it’s a quiz, sometimes a review.

Each task takes anywhere from six or seven minutes to twelve or fifteen minutes on the longer side.

You’ll see a topic — say, solving a system of equations, or quadratic formula — and you pick one: “I’ll do this one,” or “I’ll do that one.” You work through it step by step.

You’re awarded what we call XP, or experience points, which most people familiar with video games will recognize. One XP is equivalent to one minute of focused effort.

If a task is worth 10 XP, it should take you 10 minutes. You might finish in seven, you might finish in fifteen, depending on your focus and how well you’re doing. Some people work really quickly; some are more careful.

You earn XP based on how well you do. If you do really well and don’t miss any questions, you get more XP. If you struggle or you’re not taking it seriously, you won’t be awarded as much.

You accumulate XP. You’ll have a daily XP goal. An adult might say, “I want to get 40 XP a day” or “20 XP a day.” A parent might set it up for their student — maybe 25 XP a day for fifth-grade math, or 50 XP a day for calculus.

That’s 50 minutes of focused effort.

You just work through a series of tasks every day. You hit your 50 XP or 51 XP, and you’re done for the day.

It’s similar to walking into one of those fitness places that has a list of workouts — that’s the workout. Do the workout, and you’re done. Or you walk in and the trainer says, “This is what we’re doing today. Pick one of these. Just show up every day, give it full effort. You’re in and out in 45 minutes or an hour.”

Over a period of months, you’ll make massive progress.

Justin: Absolutely.

This is not a platform that you use for cramming for an exam that you have in a week because you haven’t learned any of the material — you’ve been slacking off for months.

If you come to us and you’re like, “I need to fill in my gaps in mathematical knowledge and get really far in math, and I’m looking at many months’ time horizon,” then yes, we can help you with that.

It’s very much like a gym. You think about how there are always people who walk into the gym and think they’re going to just work out really hard for a week and somehow get six-pack abs and look like Thor. The trainer has to say, “Listen, this is going to take longer. You can’t make this happen in a week. It’s not going to happen in a day.”

Even if you come in and put in a very serious workout — you’re in the gym for 60 to 90 minutes doing full-intensity workouts — it’s still going to take a longer period of time to get you looking like that.

Our retention characteristics and how far people go with using the system pretty much follow everything you’d expect with the gym. If you show up consistently — three to six times a week for your 30- to 90-minute session — and you just keep that up for months, you’re going to come out looking like the mathematical equivalent of a Greek god. It’s going to be insane.

But if you’re trying to look for an easy fix — “How can I change my life in one month or one week?” — sorry, I don’t know what to tell you.

Jason: We use the gym analogy because it’s accurate. It’s really like a personal trainer. Khan Academy would be almost like going to the gym and there’s just stuff to do. You have stuff that might help you track things. Maybe you use this machine, maybe you do that.

We’ve all known people — we’ve done it ourselves — you go and you kind of mess around. It’s better than nothing, but it’s not really doing anything.

Compare that to someone who says, “I hired this really amazing trainer and I’ve been working with them for a year,” and you look at them and you’re like, “Oh my god — you look like you’re training for the Olympics.” That’s because they’re systematic and serious about it.

People ask, “How much should I use Math Academy?” I say, well, how much do you work out to stay in shape?

Once a week? I don’t think that’s going to work. I’ve tried — it doesn’t work.

Twice a week? Maybe.

Three times a week for 45 minutes — now you’re talking. That’s solid.

Four to five times a week for 45 minutes? Okay — now you’re making really great progress.

That’s really the mental model you need. It’s about consistency. That’s what you have to do over a lifetime.

Anyone who stays in shape works out a few times a week. That’s how they maintain most of what they’ve built. Same with math.

Jason: Initially, we did create videos.

Two things: yes, they are expensive to create — it’s time-consuming — and then anytime, like we were talking about before, how we’re iterating on the knowledge graph and splitting topics, you’d have to recut and redo everything all over again. It’s really painful, especially early in the program, when you’re evolving the content continuously.

But what we actually found more important than that is that the students weren’t really watching the videos. They would just skip through it. If you sat with a student — they want videos because it’s easy. You just sit there and do nothing. It’s passive.

But passive sitting and watching somebody do something isn’t really learning. That’s just familiarizing yourself with the concept.

If you hire a tennis coach to teach you tennis, watching that guy demonstrate the backhand is not making you better at tennis. You need to watch it for a minute to understand what you’re supposed to do. Then you need to start swinging the racket and getting feedback on what you’re doing correctly or incorrectly.

What we try to do with text is say, okay, let’s show them the least amount of information we can to get them doing math. Math is a participation sport. You need to be doing math.

If you think you’re learning math by watching videos and not doing a lot of problems, you’re just fooling yourself. I can’t tell you how many videos I’ve watched on astrophysics and quantum mechanics — I still know next to nothing. I find it interesting, it’s fun.

Justin: My knowledge is paper thin.

We’re all about minimum effective doses. What we’ve found is that you can keep the platform very engaging even with a text-based format if you’re doing these minimum effective doses.

A lot of times when people think about learning from a text-based resource, they think about a big dusty textbook with dense text — you have to read a chapter or ten pages before you get to start working out problems. That’s not what we do.

Our text slides are very short. They’re just the minimum effective amount of information you need to understand what you’re doing in the next exercise.

It’s really the equivalent of a tennis instructor showing, for just a minute or two, how to hit a ball — and then the student starts hitting the ball with that technique.

We’re constantly alternating between these minimum effective doses — some guidance on a text slide, a worked example, a new technique that we’re practicing — and then jumping right into solving a handful…

Jason: Oh geez, that’s a tough one.

I give it a C. You can end up having some great people to work with, but the nature of bureaucracies is that you just have a lot of things working against you — a lot of friction.

Natural friction in bureaucracy. Doing something new requires more work. It means changing things — changing how people are doing things. They have more stuff to do. You’re breaking the schedule, you’re breaking logistics.

The system will fight against you. It’s like your body rejecting a transplanted organ — it doesn’t want it.

You can have a superintendent or some really amazing people you’re working with who are championing this and pushing it, but it takes — it’s almost like an act of God — to make something like that work in a district.

Justin: The problem with a lot of ed tech today is that it’s about passive learning — a lot of just videos without solving problems. Or if you do solve problems, it often lets you fail forward into new material.

It’s like, “Well, you only got one out of three problems correct. Ready to move on? All right, let’s go.”

There’s really a lack of educational technology systems that leverage all the decades — a century, really — of research that exists in the cognitive science and education psychology literature. It’s long overdue that more of these serious education technology platforms start popping up.

Jason: Well, for me, I’d give it an A. For my wife, she’d probably give it a B-minus.

I love working with my wife because we’re just a great team. We complement each other. I get to play more of the mad scientist; she’s more like, “Okay, this is the stuff that has to get done, and this is when it has to get done.”

The same things that make our marriage work are what make Math Academy work — we complement each other, we make up for each other’s weaknesses, and we respect each other’s strengths.

I love it. It’s sort of a dream come true for me. I’m not quite sure she’d say that.

Justin: In 2025.

Personally, I’d give it an A for our experience with it. X has been a platform where we’ve met lots of people who are interested in serious education. I’m sure there are lots of different niche circles in it, but it’s really been nice and interesting to meet so many people who are serious about education like that.

It’s been very helpful in getting the word out about our product — about Math Academy.

A year and a half ago, I had 19 followers on X. We were not on social media at all. There were a couple of people using Math Academy on X. People were interested in it, so we kind of came on and thought, “Okay, I guess we should probably interact with them.”

Next thing you know, we have a fairly big online presence there. I’m excited to see what the next year holds.

Jason: I vote no.

I’m one of those people who was excited about self-driving cars. We’ve talked about them for 10, 15 years — one of those things that’s always just a few years away.

I haven’t been following it that much. My understanding is that Waymo is running in certain cities — Phoenix or wherever — where it’s easy. The fact that it works at all is pretty good, but the fact that it’s still not usable by most people in most places?

I’ll give it a C-plus.

Justin: D.

The problem with math textbooks is, a lot of people open one up and struggle because they don’t have the prerequisite knowledge. The textbook doesn’t do anything to fill that in.

Sometimes you get a textbook with one introductory chapter at the beginning, but that’s not enough. The amount of knowledge gaps most people have going into math courses — they need much more.

Additionally, to really learn from a math textbook takes a lot of diligence and knowing what effective learning looks like on your own.

I’ve learned from math textbooks on my own. I managed to make it work. Was it nearly as efficient as if I had an adaptive learning system serving me up ideal problems? No, not at all. I could’ve gone much further with the amount of time I put in.

That said, there are some useful math textbooks that have been written. If you chop up the content and serve the right piece at the right time as a student is working problems, they can go a long way.

But in general, if you compare how far a typical student will go with just a math textbook versus being on an adaptive learning system?

Yeah — it’s a D.

Jason: I think I give that a B-plus.

In most cases, it’s better to just try and move quickly, get things done. You get into these situations with bigger companies where everybody’s trying to cover their butt. Everyone’s talking about consensus and discussion — and nothing happens.

These bigger companies move at a glacial pace, whereas smaller companies, with a small number of people who are highly autonomous, highly motivated, and have a high level of trust, can just go and make things happen.

The vast majority of innovation happens because of small companies operating under that principle. Even with Apple — if you go on Wikipedia and look at all the companies they bought — you’ll find that the vast majority of major innovations, whether it’s iTunes or iPhone technology, came from small companies they acquired. Apple didn’t build that internally — they just refined it and put it together.

Now, there are cases — medical science, medical technology — where you have to be a little more thoughtful, more careful. I don’t think the principle is generally applicable to everything.

But what we’ve found is that the vast majority of the best learning happens when you’re living in reality and getting real results. If you’re moving fast and you’re willing to accept stubbing your toe — or hurting someone’s feelings on occasion — you can make vastly more progress by getting as much experimental information as you can.

Justin: I would give that a B-minus.

A lot of the concern about screen time traces back to things like TikTok, Instagram — social media that’s not centered on learning. These are addictive platforms where you can just curl up on the couch in a state of brain rot, and suddenly four hours have passed while you’ve been scrolling through Reels.

I get it — that’s a real concern. I’ve seen this stuff. I taught in schools for several years and saw a lot of the impact that kind of thing has.

At the same time, I would say there are some people who say, “No screens at all.” They see this issue happening and don’t want their kids to use any screens, which I think is a little overreactive.

There are a lot of ways that technology can produce great outcomes for students. With adaptive learning platforms, for example, the computer making decisions to serve you the right problems can help you go much further in your education than you would otherwise.

It’s a double-edged sword, but I think it takes a more nuanced decision — what screen time is good, and what screen time is not so good.

Jason: We have three standard formats.

You have the multiple-choice format, which is a very low-friction way for students to enter information. If you have five choices to choose from, guessing typically doesn’t work. We’ve designed it so that you really cannot guess your way through a lesson — you’re going to fail — so you have to take it seriously.

But there’s also something to actually having to type in the answer. If you don’t have options like, “Oh, it’s got to be a square root or something because I see a lot of square roots here,” you’re not being primed. That’s why we also have free response — a text box where students type in an equation, expression, or number.

We also have a “select” approach. There’s a math-modal argument or a set of statements about something, and you have to complete the sentence. It’s kind of like ad libs. You use a dropdown to select one of five or seven options at three or four places to make a correct mathematical argument or complete the reasoning.

Sometimes in school they’ll say, “Show your work” or “Explain how you got this.” That does have pedagogical value. Why is this true? Why are these things congruent?

We have that as well — a variation that goes in steps, and you complete these long math-modal arguments through a series of these select statements.

Jason: Let me jump in real quick and then let Justin follow up.

I think the reality is, people say, “Well, I don’t want to do the work — just give me the intuition.” Unfortunately, intuition comes through repetition.

I can sit here and explain quantum mechanics to you, give some cute little analogies — but you don’t really know it. You have to do it.

Why did Michael Jordan and Kobe Bryant spend hours and hours doing free throws? Repetition. All their different skills — they had to drill those skills. You need a certain amount of repetition to get the automaticity, the natural feel. You just know how things work.

Whether you’re practicing chess or violin or anything else, it works that way. You don’t just have someone explain and transfer their abstract notion to your head with a cute analogy. It doesn’t work.

I’m here to tell you — you’ve got to do the work.

However, what Justin said earlier — we’re focused on minimum effective doses — is critical. You don’t have to do a hundred factoring problems. Maybe we do seven or eight or ten to start. Then the next day, we do five, or three, or four. A week later, we do some more.

We train with limits so you don’t have to do excessive amounts, because that is highly inefficient. And efficiency is the name of the game.

The more efficient you make the system, the more students can learn, and the quicker they can get through what they need to. Whether you’re trying to get ahead and move through quicker, or you’re just trying to get it done because you don’t really care about math and you’re more interested in English — you want to get done with it.

Efficiency is the key.

Justin: One thing that I’ve noticed — and a lot of people who want their math to be less skill-heavy and more concept-heavy — is that what they’re really often looking for is a light-speed overview of a subject without going into the nitty-gritty, without really getting their reps with everything.

You might think of a YouTube video that a lot of people find really interesting because it explains all of calculus in an hour, or all of linear algebra in 30 minutes, or how neural networks function in 20 minutes. While it can be exciting to get a surface-level understanding — to go from not knowing anything about it to knowing some of the surface-level pieces — there’s really a lack of knowing the details of what you’re doing.

Being able to code it up, or implement it, or just reason about the details — that’s what’s missing.

We’re really focused on building up a true level of mastery. Not just surface level. Not just shallow.

The optimization problem is not: how fast can we imbue you with a shallow understanding of something — enough that you can tell your friend something cool about it, or feel like you have an opinion about it?

What we’re focused on is: how quickly can we get you operating mathematically, almost as a professional musician plays their instrument, or a professional athlete performs in their sport?

We’re all about building up everything — all the skills — completely, like an academy for math.

Jason: A good analogy we use is this: ask yourself, if it wouldn’t work in sports, it’s not going to work in math.

You don’t look at any successful basketball team, football team, soccer team, hockey team — what do they do? They work on individual skills. They work on conditioning. They build up component skills with two-person drills, three-person drills. Then at the end, you might have a controlled scrimmage with a lot of feedback. The coach stops and says, “No, I want to see this. I need to see that. That’s not correct.”

Then maybe you do some pre-formed practice. That’s how you need to think about it.

You cannot get good at a sport by just listening to a coach talk, or watching videos, or having someone say, “Let me explain the theory of good defense.” A little discussion is fine, but then we’ve got to practice how to stay in front of a player, move our feet, and do all that.

If it’s not going to work in a sport, it’s not going to work in math. That’s a really good mental shortcut. If somebody’s telling you something, and you think, “Well, I used to play this sport, and that wouldn’t work at all” — that’s your answer.

We talk about project-based learning. Project-based learning is kind of like scrimmage. It’s fun to scrimmage at the end of class, at the end of practice. You need to do some scrimmage. If you’re learning to play basketball, you’ve got to play the game.

But if you just scrimmage the whole time, none of the skills are going to develop. Any skills.

Jason: It’s funny — I just spoke to a school yesterday, and we had a really nice conversation.

I’m really excited, but I say, don’t listen to me — just try it. Try it, but take it seriously. Let’s get a pilot going with a small number of students, get them on it, make sure you’re paying close attention to what they’re doing, and you will see incredible results.

Typically, students learn about three to four times as fast as in a normal class. That’s why, in our school program, students could go from pre-algebra through AP Calculus BC in three years — from sixth to eighth grade.

I remember when that first happened. There were Washington Post articles about it. People couldn’t believe it. They were like, “This is not true. This is smoke and mirrors.” It’s 100 percent true. We know because they took the AP Calculus exam — we actually have results.

I would say, try it. We have a 30-day free trial — you can cancel and get your money back. Try it, but take it seriously. Don’t say, “I signed my son up, and he did it a couple days ago.” Make sure they’re doing it every day — 20, 30, 40 minutes, whatever is appropriate for their age.

Keep a close eye on it. Are they taking it seriously? Are they using paper and pencil? If they do it, you will see incredible results.

Jason: I would say it’s, in a lot of ways, more useful.

Sometimes students get so far behind that their mathematical journey is basically over. They’re just trying to keep their head above water for a couple of years until they can drop out of math.

We can go back and fix these problems. By giving them an adaptive diagnostic, we can find all their missing prerequisites. They could be missing a whole course or more. We can go down to that level and start building up those foundations.

One of my favorite sayings is, “Nothing succeeds like success.” Once you meet the student where they are and start building from a stable foundation, step by step, and they’re being successful, it changes everything.

We’ve had tons of emails from parents saying, “My son or daughter was a year or a year and a half behind, had math phobia, it was a disaster — and now they’re being successful. They went from getting a D to getting an A in class.”

Clean up the missing pieces. Build a positive mindset from real personal success — not because a parent is saying, “Have the right mindset,” but because the student experiences success. They go, “I can succeed at this because I’ve been succeeding at it. I’ll just keep succeeding.”

Justin: That’s pretty much how you want to think about it.

Something we hear all the time from students who were behind in their school class and used Math Academy to catch up — and then start crushing their class and going well beyond it — is the change in their level of confidence.

In one year or even just months, a student can go from thinking they’re not a math person — that math is way beyond them, that they’ll never be good at it — to thinking, “I can do the homework. I’m crushing my class. This class is easy. I’m actually learning above grade level.”

That change in the student’s experience can do wonders for their confidence.

Jason: Okay, here’s what I always say.

Sometimes you run into a parent who gets a little carried away, and you’re like, look — an hour a day is great, unless you’re in a time crunch. Say, your student was in an AP Calculus class and the teacher left, and the new teacher is a disaster, and there are only three months left before the exam — okay, now you might have to put in two or three one-hour sessions a day.

But you want to split them up. Three hours in a row is just too much.

I always tell parents, 50 to 60 XP a day is kind of the upper limit for normal — more for high school, maybe upper middle school. Maybe 90 minutes if the student is really serious.

Anything more than that, I’m like, you’re getting a little excessive.

It’s more about quality. I always say it’s like going to the gym. Don’t go to the gym for two hours a day — just put in a solid effort for 45 minutes, five or six days a week, and you’ll be in incredible shape.

It’s really kind of the same thing.

Justin: It really comes down to how long you can sustain productive, full-intensity effort.

An hour-long workout — whether it’s on Math Academy or in the gym — that’s pretty taxing. It’s hard to keep that up for multiple hours.

Now, if you want to do multiple hours, you may be able to split it up — one-hour session in the morning, one hour in the evening. But really, every person is different in how much they can stay focused and work intensely on the system.

That general one hour a day is what we’ve found to be the upper end of a sustainable approach for most students.

Justin: Absolutely.

In addition to mastery learning, which we’ve talked about, there’s also spaced repetition. Everything that a student learns on the Math Academy system, they’re going to review indefinitely into the future — but at expanding intervals.

Maybe you do a lesson on a topic today and see a review again tomorrow or the next day. Then you pass that review and see it again a week later. You pass that, and you see it again two or three weeks later. That’s the optimal way to keep your knowledge base fresh.

Everything you’ve learned is going to come up later as you try to learn new topics. New material always depends on previously learned material. If you’re not keeping that material fresh, you’re going to struggle to learn new things. It’ll take longer, you’ll have a harder time, and the learning process will be less efficient because you’ll end up having to re-learn things.

If you let something sit without review long enough, you can get to a point where you have to relearn it from scratch — and that’s bad. We make sure that doesn’t happen.

At the same time, we try to make this as efficient as possible. If you learn one-step linear equations — like ax = b — and then tomorrow you learn two-step linear equations — ax + b = c — learning the more advanced topic fully practices the earlier topic as a subskill. That counts as your review.

In our knowledge graph, we’re tracking all of this. When you learn something new, we ask: what subskills are being implicitly practiced? We update the spaced repetition schedules of those subskills so we can knock out reviews implicitly by having you learn new material that encompasses them.

We always try to select tasks for students that knock out as much review as possible. They get the review without having to do multiple tasks. They learn something new, and by learning it, they’re implicitly practicing something they’ve already learned.

There are a number of other cognitive science principles involved — interleaving, also known as varied practice — and we do a little gamification with XP, leaderboards, and testing effects.

We include retrieval practice through quizzes. Roughly every two hours on the system, students get a quiz.

Jason: I would say it’s by design.

It was like, let’s get some gamification up. We needed some way of accounting for progress and effort. The XP model was a great way to do that. What have you done that’s productive? How much progress did you make? If I’m a parent or teacher, am I happy or sad with what I’m seeing?

We focused more on things related to the actual learning process, rather than side quests, achievements, badges, and that kind of stuff. I don’t have anything personally against that. We’ve talked a lot about doing more of it, and we think it would make the system more fun — probably for adults, but also for kids. They really get a kick out of that kind of thing.

We could layer more of it in. You just have to be careful it doesn’t get out of control, where it becomes distracting from the learning itself.

The point isn’t to make it all about the gamification — it still needs to be about the learning.

But I don’t think that’s a big risk. I think we’ll probably be layering in more gamification over time.

Justin: Yeah. Well, I think probably the thing that is most surprising to people — actually, some people don’t even believe it, it’s so surprising — is that a lot of people think students need a million different explanations of something, and that one of those explanations is going to stick. And it’s different for each student which explanation they need.

But really, all you need is one really good explanation that is battle-tested across a large number of students. And the students need to come into that explanation with all their prerequisite knowledge in place.

If you do that, then you can get students learning the skills really well. Actually, 95 percent of our lessons are passed on the first attempt — above 95 percent. And above 99 percent of lessons are passed within two attempts, without any additional remediation.

If a student fails a lesson — on the rare occasion that they do — then they try again maybe the next day or two days later. Just enough of it has consolidated in the brain from the first attempt that they are able to get over the hump the second time around.

That’s always very surprising to people who, maybe from their tutoring experience, think every student needs a different explanation. But typically, what’s happening when somebody thinks that is really a symptom that the student is majorly lacking prerequisite knowledge. You’re trying to come up with some explanation that allows them to grasp just enough of the topic — not the full thing, but good enough — without requiring too much of the prerequisite knowledge they’re missing.

If you ensure that students are coming into new instruction with all their prerequisites in place, then really every topic can be explained the same way for each student.

Jason: Can I just jump in and add one thing to that?

One thing we always talk about is: if you lower the learning gradient, make it small — anyone could climb Mt. Everest if it were just a slight incline. What’s hard is when you go through sections that are extremely steep — too many difficult steps at once.

What we’ve done is try to break every steep section into smaller steps. If you break things into small enough steps, anybody can learn anything. It might take a little while, but you just go in small steps.

That’s what we do with our analytics — we ask, where are the congestion points? Where are people struggling? Where are we trying to do too much at one time?

Let’s break it into two or three steps.

That’s the biggest thing. A lot of times, when something’s not working — assuming the student knows the prerequisites, because that’s often the problem — if that’s not the issue, it’s because you’re trying to do too much at one time.

You need to break it into smaller steps. Let them absorb that first thing. Let it lock in over a day or two.

Justin: Exactly. Layer the next skill on top of it.

It’s not that you need to explain it in a different way. You just need to take your existing explanation and smooth it out. Break it up over multiple steps in this learning staircase, so that each step isn’t too big for the student to take.

Jason: I’d say the first thing is the lack of accountability.

Most of what you find out — like at the University of California, San Diego — is that 8% of the students were at an elementary math level. Which is an unbelievable reality. But this is widespread. It’s not just there — it’s everywhere.

The reason that is, is because we’ve kind of given up accountability. It’s lost in the bureaucratic structure. Who’s at fault? Is it the teacher? The principal? The teacher from the year before? The school board? The curriculum designers? Nobody’s at fault.

If you don’t have accountability, and you also don’t have transparency — we get rid of standardized tests, or the tests aren’t really shared, or we eliminate the SAT — then nobody knows anything. We’re just hiding the ball.

And the reason we’re doing that is because we’re hiding from accountability.

You don’t have to go to the NFL and say, “You have to run this offense.” If you don’t win, you’re out of a job. Your coaching staff is fired mid-season if you’re losing games. So everyone starts looking at what’s working — because they’re held accountable every Sunday.

You have to have transparency and you have to have accountability. If you do that, the education system will converge on things that work. You don’t need us saying, “You have to do active learning” or “You have to do mastery-based learning.” You don’t need to run a West Coast offense — do what you want, but if you lose, you’re out of a job.

Now, it doesn’t have to be as cutthroat as professional sports, but you do need accountability and transparency if you want a highly functional system with high-quality outcomes.

Justin: I would say something similar, along the same vein.

There used to be a reliable source of truth on what a student actually knows. We’ve got a problem now that grades have become very untethered from what a student actually knows.

A student can have an A in the class, yet somehow not actually be able to demonstrate most of the skills that are supposed to be learned in that class.

There are many reasons for this — problems with assessment, failing students forward, parents lobbying for their kids to get higher grades — it’s a mix of issues.

But in my experience, so many of these issues in education — of students essentially just not learning the material — come down to, at some point, a piece of paper says they’ve learned it, when they actually haven’t.

Jason: You almost always need an adult in the room.

Kids are like puppy dogs. If you don’t watch puppy dogs, they’re going to run around and chew up your shoes. You’ve got to be on top of them. You’ve got to keep them focused. You need to keep them motivated.

In the way we use the Math Academy platform in the school program, we have instructors. A couple of them are PhDs in math — so it’s great having them there.

What they do is spend time at the beginning of class saying, “Let’s talk about this math idea,” or “This interesting challenge problem,” or “Let’s talk about the history of this problem,” or “We’ve got this ongoing project we’re working on.”

Then they’ll spend maybe 30 minutes or so on the system making progress — so they don’t have to do a lot of homework.

It’s a mix. Some of it is group interaction, instructor-led. We’re doing this thing for 10, 15, 20 minutes, then it’s, “Okay, everybody, let’s get to work. Let’s make some good progress on this.”

That way, you go home and have 15 minutes of homework, not an hour and a half.

Jason: Oh geez — what do we not need to improve?

I look at it and I’m just like—any time I get an email from a customer saying, “Oh, I wish it did this, this, and this,” I totally agree. I have the roadmap a mile long.

I close my eyes and think, what would be the ultimate online learning system? That’s what I’m always thinking. Then I ask, are we doing that? No? Okay — we need to do more of that.

The thing that Justin and I always do when we’re figuring out what to do or how to do something is say: “If you were tutoring a student, what would you do in this situation? Would you do this?” No? Then let’s not do that.

We’re always trying to replicate the nuanced behavior of an expert, caring adult. What would this adult do to keep the student in a positive frame of mind, making progress?

One of the things we’re going to be rolling out soon is what we call in-task coaching. When a student is working through a lesson, if they skip through the tutorial — and it should take them maybe a minute to read — but they skip through in 20 seconds, you’d be like, as an adult, “Whoa, whoa, slow down. Come on. You’ve got to read this.”

Now we’ll have an in-task coach — kind of a little robot avatar — that’ll say, “Slow down.”

Or let’s say it looks like they guessed on a question that should have taken at least 30 seconds, and they did it in five and got it right — okay, well, on paper, that looks off.

It’s that kind of in-task coaching we’re building. We have this little avatar to make it more fun, a little softer. I was working on this messaging and thinking: what would I say if I were sitting next to a 13-year-old kid who’s rushing or not being careful?

How do I keep them motivated and feeling good, but also hold them accountable? Like: “Look, you need to actually read the explanations. Don’t just guess.”

Jason: Thanks so much for having us.

Justin: Thanks for having us.

Prompt

The following prompt was used to generate this transcript.

You are a grammar cleaner. All you do is clean grammar, remove single filler words such as “yeah” and “like” and “so”, remove any phrases that are repeated consecutively verbatim, and make short paragraphs separated by empty lines. Do not change any word choice, or leave any information out. Do not summarize or change phrasing. Please clean the attached text. It should be almost exactly verbatim. Keep all the original phrasing. Do not censor.

I manually ran this on each segment of a couple thousand characters of text from the original transcript.

Want to get notified about new posts? Join the mailing list and follow on X/Twitter.