Learning Made Efficient: The Brainscape Approach with Andrew Cohen

Andrew Cohen is the founder & CEO of Brainscape, a web & mobile education platform that helps millions of people study more efficiently.  Brainscape allows students of all ages to create, share, and find great “smart flashcards” for any subject, and to study them using a fun, social experience that is scientifically proven to boost learning results.

Before starting Brainscape, Andrew worked as an international economist (U.S. Bureau of Labor Statistics), eLearning curriculum developer (Satori Consulting and United Nations), and government corruption fighter (World Bank).  He holds a Masters degree in Instructional Technology from Columbia University and is obsessed with optimizing human development using cognitive science.


Recommended Resources:
Brainscape website
Brainscape mobile app
Brainscape Academy
Brainscape YouTube

Get ready to explore the future of education! Join Edtech Insiders for a virtual conference featuring 30+ of the top voices shaping the future of Al + Education. A full day of keynote speakers, panel discussions, and networking!

Register now here: AI+EDU Virtual Conference

Transcript

Alexander Sarlin: 0:04

Welcome to Tech insiders where we speak with founders operators, investors and thought leaders in the education technology industry and report on cutting edge news in this fast evolving field from around the globe. From AI to xr to K 12 to l&d, you'll find everything you need here on edtech insiders. And if you liked the podcast, please give us a rating and a review so others can find it more easily. Andrew Cohen is the founder and CEO of Brainscape, a web and mobile education platform that helps millions of people study more efficiently. Brainscape allows students of all ages to create, share and find great smart flashcards for any subject, and to study them using a fun social experience that is scientifically proven to boost learning results. Before starting Brainscape. Andrew worked as an international economist at the US Bureau of Labor Statistics, an E learning curriculum developer for the United Nations and Satori consulting, and a government corruption fighter at the World Bank. He holds a Master's in instructional technology from Columbia University and was a classmate of mine there and is obsessed with optimizing human development using cognitive science. Andrew Cohen, Welcome to EdTech insiders.

Andrew Cohen: 1:30

Good to be here. Thanks for having me.

Alexander Sarlin: 1:32

Yeah, you and I have known each other a long time we went to graduate school together at Teachers College, Columbia, and I've seen you take this company to such great heights over the year Brainscape has had a journey for 13 years, can you give us a little overview of the initial concept of Brainscape?

Andrew Cohen: 1:51

Yeah, it's been a roller coaster. But it actually started as a personal project that I had back in my old life of International Development, I was working with the World Bank down in Panama. And I had to get really good at Spanish for this controversial project that I was working on. That was using Rosetta Stone, which was big at the time, I was kind of making a spreadsheet of all the new words and phrases that I was learning, and I just wasn't learning fast enough to really be able to engage in these cabinet level meetings that I was being involved in, in Panama. So I turned that spreadsheet I had into an automated study system for myself, where I would, you know, study a word, I would see the translation, or even a phrase, and then I would rate myself on a little scale for how well I thought that I knew it. And then it would use that system to then determine how frequently to repeat that for me, or how soon to show it to me again. And what I realized in the process is that I had sort of discovered the concepts of active recall, or retrieval practice, as they say, right, where instead of just recognizing, you know, multiple choice, are matching, like most, you know, edtech games for language seem to do, you're actually thinking about it from scratch. Metacognition, where you're assessing, you know, how well do I really know this and rating yourself and then spaced repetition, the system of, you know, determining how frequently something should show up based on how well you knew it. And what I kind of realized was that technology can automate those three principles much more efficiently than you could ever do with paper flashcards. So that motivated me to go actually back to grad school, where you and I met at Columbia Teachers College, get my master's in education, technology, and have a concentration in the cognitive science behind why these principles work so well for me, and actually, for some other friends that I had met in Panama as well. So I graduated from Columbia with a prototype of Brainscape, that we did a focus group, and had people study their national birds using paper flashcards. And then the other half of the group use the Brainscape prototype. And it turned out that they learned more than twice as fast performed more than twice as well in that 30 minute study. And so we realized that if that's the difference in just 30 minutes of studying, you know, imagine the compounded benefits of using this system over months or years of studying whether it's a foreign language, whether it's, you know, from medical school, or law or any really high stakes subjects. So I realized that I had a duty to bring this miracle of learning to the rest of the world via you know, the web and mobile technology which was kind of just becoming very accessible at the time. And I wanted to do it the right combination of cognitive science and use your experience talent and technical talent not just for individual students, but also for for teachers, professors, publishers, companies, associations or tutors or anyone who really values Efficient studying of huge amounts of info from pre K all the way to old age.

Alexander Sarlin: 5:05

Yeah, you mentioned a lot of really important learning science principles. And it sounds like you kind of stumbled on them. But then realize you are on something that had a lot of research behind it. Retrieval practice is so important spaced repetition, for those who don't know about spaced repetition and the sort of forgetting curve, because there's a little bit in the weeds for instructional designers give people a little understanding of how the spaced repetition and the forgetting curve works and how Brainscape works to ensure that people really keep information top of mind.

Andrew Cohen: 5:36

Sure, so the forgetting curve usually attributed to him and Ebbinghaus in the early 1800s, is the idea that you know, in the first 30 seconds or five minutes, whatever, that you've been exposed to something, you very quickly, forget it, right. If you just learn the word for meatball and Spanish, it's a bone bigger. So I could I could tell you it in a story. I can evoke emotion, I can do all the cool mnemonics and you know, people say, right, you're gonna remember something better. But the fact is, if I don't ask you, again, within 30 seconds or a minute, how do you say meatball in Spanish, you're probably going to forget that it's on bone digga, because that's for whatever reason, you know, a tough word. So I need to ask you again pretty quickly. But then after you've retrieved it, again, you've made that mental effort, it was hard, maybe I corrected you one more time, then maybe I could ask you again, and you know, two or three minutes before it would totally slipped your mind. And then after that, maybe get in 10 minutes, maybe then an hour, then tomorrow than a week. And that interval tends to, you know, expand over time, you can go for longer and longer without needing to be reminded. And it's custom for every piece of information, every Supreme Court case, every you know, bone that you're labeling every cocktail recipe that you're trying to learn as a bartender. And so if you have a system that can customize the specific interval, before something is going to be repeated based on how well you know it, you can truly optimize the learning process, optimize your personal development stream. And so what we really need to be able to do is, you know, provide this type of learning for as much content as possible, not just for you to, you know, create flashcards for yourself, or for teachers and students to you know, collectively create stuff for our curriculum, but really aligned to most of the curricula that are out there, which, you know, for us now consists of a lot of standardized exams and professional certifications. But you know, more and more ways that we can integrate into the different types of curriculum.

Alexander Sarlin: 7:41

You mentioned before how the algorithmic approach that technology is so much better at that kind of spaced repetition, that kind of bringing things back up in relationship to your confidence than any individual teacher can be even a personal tutor. Because it can literally time, you know, okay, five minutes, it's been an hour, it's been two weeks, it's time for you to review. You know, I'm curious, can you tell us a little bit about some of the success stories that you've heard, I'm sure that you, you know, over the years, some people have come and said, This just totally changed my way of learning. Yeah, I

Andrew Cohen: 8:15

mean, you can go to the Brainscape app on Google Play or the App Store and just read troves of reviews from so many different types of people, whether they're, you know, professionals and you know, their medical student, for example, right, the most information that anyone really asked to learn, and you just need such a knowledge management system, all the way down to, you know, parents using it with their preschool students who needed to memorize sight words, or people who are using it to memorize come back, because they were being bullied on the playground, and they just want to, you know, come up with all the scenarios where your mom was so fat, or you know, whatever, they want to have something to say right to defend themselves. So the number of scenarios right, we've we've had, you know, ADHD, parents say that it totally changed their kid's life, the power of, it's not just the, you know, the making the studying of facts more efficient. I know, there's, you know, backlash against memorization that maybe is swinging the other direction more recently, but it's the act of having to rate your confidence. And in Brainscape, it's a it's a confidence on a scale of one through five. So you're assessing you see the answer you say, how well did I really know that? That is helping you to better understand your own knowledge, your own pace of learning, so that you can then calibrate over time, you know, how should I allocate my upcoming study time, you know, based on how comfortable I am with the material and that, that self awareness, it just it helps people grow as a learner as a student as a person in ways that again, you know, you could read about and all over the internet, it's really making a huge impact on millions of people's lives.

Alexander Sarlin: 9:52

Yeah, I see 10s of 1000s of ratings on the App Store. And you know, it's I've used Brainscape myself, I actually used it for very specific purpose, you mentioned these certifications exams, I used it to pass my driver's test and which happened much later in life than many other people do. But they get their driver's license. But I used it to get all the rules of the road and to study and to do multiple iterations until I really, really got them into my head. And I tell you, I just firsthand that confidence rating system, it's really interesting because at first you think you're going to be overconfident. And then you quickly realize that there's sort of no benefit in being overconfident. If you really totally forget something, you should tell the system because then it's going to help you learn it. So you have this sort of intimate relationship, and you're going back and forth with a brand scape app. And it's really helping you bring everything up, I understand you've launched a smart study feature recently that sort of begins to touch on personalized learning, maybe even interleaving, which is yet another important learning design concept. Well, how does the smart study feature work?

Andrew Cohen: 10:55

I'm really glad you asked about that, that's probably our biggest innovation on the learning process that we've made in the past couple of years. So interleaving practice typically is what they call switching between different disciplines, different subjects. So maybe you study math for 10 or 20 minutes, and then you you weave in a little bit of history, then you weave in a little bit of science or other types of topics. And the idea for that is that the mind works better when it's presented with fresh information, right, you sort of get fatigued, thinking in just one subject realm after a short amount of time. And so you should mix it up, it also helps you be more engaged, right, it can get kind of bored if you just been studying one subject for too long. And it's also you know, more realistic like with how the real world might treat you because you don't, you know, just go to history class in the real world. But a history fact might pop up in a conversation in a meeting about product development, or, or wherever. And so, you know, you need to be constantly on your toes. And so we took that principle of interleaving practice that they're always trying to recommend students use. But that's just so impractical in real life because of switching costs and getting out the notebooks or textbooks for different subjects constantly knowing what to study. And we put it into Brainscape as the primary way that you would actually be encouraged to study in Brainscape. So you open up the Brainscape app. And you know, there's a list of all your classes as there always has been right so biology 101, Spanish and wherever you're studying. But at the very top, there is just one giant master study button that has a little configuration icon next to it, and you hit the config icon. And you see sliders for the weights that you want the smart study mix to include your classes and so let's say you know Spanish is very important to me right now I'm about to go on a trip to Costa Rica next month, I might turn up the weight of Spanish to 80% I want 80% of my flashcards to be Spanish but you know, I actually want 5% of my flashcards to be no jokes. I'm trying to memorize jokes and other 5% to be personal development flashcards, right, I want you to occasionally remind me, you know, are you sitting up straight right now? Are you breathing from the abdomen, and another 10% to be just kind of, you know, English vocab, or, you know, history facts, or whatever I'm trying to learn. So you can customize exactly what's important to you, how do I want to develop myself in the pattern that works for you, and then you just push one study button, and you're gonna get that stream of Brainscape flashcards with the same, you know, confidence based repetition algorithm that you always get, but from across all the different subjects that you have in your Brainscape account. And so we even we preload every new users account with three great classes worth of content one called Knowledge rehab, basically everything you should have learned in K 12 school, but maybe forgot street smarts, a whole bunch of other thing, you know, the dog breeds and flowers and trees, you should identify, right? What are the parts of a car, you know, all bunch bunch stuff that everybody should really know. And vocab builder. And so even if you came for Brainscape, for just one class, you know, your biology class, whatever, we are giving you the experience of having a smart study tool, because you're preloaded with things that you can move the sliders for. So we're really, really proud of it. We've had so many people just both publicly and privately and on our Discord servers, talking about how that's just changed the way that they have been thinking about their learning going forward. And we see it as a foundation of a lot of other things that brands gigs can be building in the coming years.

Alexander Sarlin: 14:31

Yeah, no, it makes a lot of sense. And it's interesting because it really, as you say, it sort of changes the nature of sitting down to study or sitting down to sort of think about what you want to remember because instead of saying, Oh, I've got to study my Spanish, I've got to study my history. I've got to study my biology. You just say, I need to study every day, every few days. I need to go in and study and now I can refresh everything I need to know in proportion to how much I want to know it. You know, you mentioned that testing, you know that certification tests or school tests are a major use case for Brainscape. And I can totally imagine that people prepare, you know, as you say, you might turn up your 80%. Spanish, if you have a Spanish test as well as that, yeah, or a medical exam them cats or things like that. Spaced Repetition is also dedicated to, you know, very specific intervals of time. So I'd love to hear how the app sort of works around time does it sync with calendars for upcoming events? How can users make sure that the timing aligns to what they need?

Andrew Cohen: 15:33

Yeah, it's not so much synchronization with calendars as it is the built in time estimates at every level of content organization that we have. So what I mean by that is, you know, one of the biggest studying mistakes that most students make is is overconfidence, and thinking that you can just sort of wait until the last minute and then cram it all in, which I guess for a small quiz, you know, maybe it's fine, you got 30 or 50 questions, you could probably memorize those in a short amount of time. But if you've got, you know, big midterms, or finals, or certifications, or medicine or foreign language learning, you're really going to have to pace yourself. And so when when you start a Brainscape, study mix, let's say for an entire class or an entire exam, we tell you a constant update of how much time we estimate that you have left to get to 100% mastery at your current rate of learning, let's cool. So you hit a checkpoint after you know, 10 cards, 50 cards was a hey, at this rate, you know, you have four hours and 20 minutes left, before you're going to be at a confidence level that you're comfortable with to take this test. And so you're able to take that as a learner and say, Oh, how am I going to then allocate my time, over the next couple of days or weeks, or however it is in conjunction with spaced repetition in order to reach that goal. And it turns out that being able to estimate, you know, the pace of learning, and you know, how much time is left is a bigger technical challenge than one might think, particularly thinking all the time spent across web and mobile, and what if you were offline and making sure to, you know, keep things updated. And that's why kind of no other flashcard platform or study platform is really able to do this. But we have increasingly been, you know, thinking about this as a bigger and bigger part of our brand and our own sort of habit building philosophy, you really you have to have a time component built in and made transparent to the user.

Alexander Sarlin: 17:27

That's fascinating. I have never heard of that approach. And it's a really, it's an exciting way to look at it. I mean, so many things you're saying are sort of sparking my instructional design, memory. And this is really about metacognition, I mean, the ability to understand your own pace of learning is so unexpected. I don't think a lot of people know that very well, no matter how many tests they've studied for how many things they've learned to say, Oh, I'm gonna need, you know, three weeks to bone up and really know this. It's such a unique value proposition to be able to tell people at this pace, this is what you would do, here's how many hours you need. It's really, really interesting. As I hear you talk about Brainscape, it's so clear that you've evolved the platform that you've added all of these features, and you've been very responsive to both the research and to what you've been hearing from learners. And you know, you've been in this space for 13 years, your experience as an entrepreneur has been pretty unique. And I think you know, other founders might be really interested in how you grew this company, especially because you decided early on to follow a non venture funded path. So can you tell us about that decision and how your experiences have helped you learn about what entrepreneurship can look like in the tech space.

Andrew Cohen: 18:44

Not sure if your audience can see me smiling right now. But I think it's very kind of you to say I decided to take a non VC funding. I was very much on that path for five years and raising a bunch of, you know, seed angel money, trying to kiss every VCs, but to raise that series, and did not succeed. So that's why it took a lot more time than it needed a result of a couple of mistakes. The first big one was coming out of grad school, I had this prototype, it had proven, you know, in a lab setting to work so well. I had been obsessed with it for so long, that I decided I just have to build this thing. I'm not an engineer, and I took a couple coding classes in our grad program. But I decided I was going to take all my life savings, beg borrow and steal from every family member and friend, a family that I could to hire software engineers overseas, right? I thought I was saving money by you know, hiring cheap talent that was not really a co founder right on preserving my equity and ended up you know, because we weren't iterating like a couple of co founders would on a minimum viable product and testing you know, with users right there in the room, but instead I was shipping PowerPoints overseas and waiting three weeks and then they say Tada, here's your your crappy product. I was constantly on this like flywheel of the product wasn't good enough, it was too complicated, right? I didn't, you know, sit there with good product and tech people to cut down on the number of features that I was comfortable releasing with. And we constantly had to just re Oh, just a couple $100,000 More a couple of $100,000 more with a product that was just always, you know, right around the corner from getting good enough. And so what happened was, you know, I thought I was about to raise like a five or $10 million Series A, I was in later stage conversations with some venture capital firms. And a couple bad things was all happened at once our CTO got poached, actually, by Twitter, another kind of, you know, automated stream of information that, you know, saves you from having to go read the blog itself, right, you get all the all the blogs in one place, kind of like we're doing with flashcards, so they loved our CTO, couple things happened with like the Google and Apple algorithms. So suddenly, our search volume, and discoverability, plummeted, so our user numbers were going down. And so these VCs pulled their money, I hit the wall, we went from 12 people and in New York with all the hype and keg parties, and ping pong, tables, whatever. They're basically just me, my original co founder, Andy lotze, was our head of product. And he had come from Princeton Review previously. And then Jonathan Thomas, who really saved the company, he became our new CTO, he was an engineer we had recently hired he had a background from for Oracle, and Intuit, and he had built a couple other startups, including an education before and just really decided, you know, he's gonna buckle down with us, he took a little bit of a pay cut, I was kind of the only person we were paying at the time, I was consulting on the side to make some money myself. So it was Andy. And for just the sweat, the equity and the passion of saving something that we all really, really believed in, we focus and the two biggest things that we did is we sort of improved the revenue model went from a bi flashcards, one off to sort of a subscription model, all you can eat plus a couple features that, you know, is healthier for the business. And we also finally completed the user generated content marketplace so that anybody can make flashcards that are public, for the rest of the world to see and collaborate with. And both of those things, you know, they took about a year and a half to really kind of get going. But once we did by like 2017 18, we're starting to get out of the woods, enough to start hiring other engineers, and other team members on sort of the operation side, our customer service assistant, Courtney, who was down in the south, who was part time, you know, she became kind of our director of marketing, and we just were working with the scrappy team that we had. And slowly but surely, we've grown back to about 25 people all over the world, this time around, we're 100% distributed, we've got an old cap table of all those angel investors and seed investors that need to be taken care of. But we've got a very good profitable, healthy company, and a renewed sense of purpose, and runway ahead of us that we really all collectively think that we can be building the foundations of a whole new type of education platform, from the inside out from the bottom up with just better focus than we ever had. And that would be my advice to any other entrepreneurs, whether you think you know, your VC fundable, or just selling to schools or bootstrapping, or consulting, just focus, make a list of everything you'll ever have to build or want to build, and then choose the 2% of it, that you're going to be the best in the world that and just nail that first and spend all the rest of your time on trying to get people to care about that 2% Before you try to build everything else, because nobody's gonna steal your idea. They're gonna want to work with you if you're really the best at it.

Alexander Sarlin: 23:41

That's a really fascinating story. And I'm sure many entrepreneurs and founders and aspiring founders and VCs, for that matter, are probably listening to this and saying, Oh, really, that's a really interesting trajectory. And we're in a pretty strange space. In terms of venture funding. Right now, there's not as much money flowing, certainly many fewer big rounds, than they have been in quite a while. And I think that kind of scrappiness and focus and you know, making the most of what you have and thinking about profitability, maybe earlier than you would have otherwise, are very powerful, very powerful ideas in this particular moment in time. Are there any other lessons that you would pass on? I know, you just gave a couple of great pearls of wisdom. But if I were to push you even further and say, in your experience, you know, going into that VC world, talking to angel investors, and then sort of flipping over and saying, Well, now we're going to do this in a more bootstrapped way with the money we have with you having to actually earn money. What is the biggest difference between how the company worked in the first phase and the second,

Andrew Cohen: 24:44

I think, getting a better understanding of whether you are VC scale, right, venture fundable? You know, nowhere to say like can you really see a viable path to you know, 100 million a billion dollars in revenue in seven to 10 years, you know, with successive rounds of funding, but is your market that big is the problem that you're trying to solve that big do you have the network effects to really, you know, achieve that, because if you don't maybe it's just it's a cute niche, you know, math fractions product, or you know, something that kind of does a really great thing and a great way, but just isn't really venture scale, then, you know, trying to play that game and get the angel investors and the seed funds or whoever probably is going to end up, you know, not being the most efficient use of your time. And so yeah, first things first, you know, understand, like, what's the potential scale of your product, and then adapt your strategy to that. And if you are venture scale, or you think that you're going to be, then yeah, you know, it's okay, for example, to have a product as long as you have sort of the financial cushion and ability to do it, that maybe you know, you're not monetizing yet. Maybe it's free, and you're just proving that you can get traction. And if you can be growing virally, you know, with hardly any marketing, because everybody's telling, you know, five people about it, but you could have retention, you know, probably even more important, right? If the average person who uses it is still using it every day, you know, three months later, that's a really, really strong sign to a smart venture capitalists like, hey, look, if the product has that much retention, they'll figure out how to scale and monetize it, because retention is gold. So that's if you're venture scale, if you're not venture scale, then yeah, stop trying to play the selling game, to the VCs. Instead, play the selling game to your customers, with actual, like, show me the money, show me the dollars. And I completely get that, you know, for a lot of products, there's probably a lot of entrepreneurs out there that are hearing. Well, yeah, that's great for you to say, but it's gonna cost me you know, half a million dollars in software engineers to build the thing that I want to go out there and sell. And there's the chicken and the egg problem, right, but the capital. And so if I were in that position, if I had something that costs a bunch of money to build, but wasn't quite venture scale, but I really just know that it needs to exist. And maybe there's still a five or $10 million opportunity there. I would maybe try to find a way if you if you have to bootstrap, right, you can't raise the money. Find a way that like, you could maybe start it as like a service business. Right? So let's say you've got, I don't know like the math fractions, perfect. You know, calculate whatever it is very, very small thing, become a tutor become like a math tutor yourself, right? So you've got 30 customers right there who could possibly use it. And then either yourself right learn to code have a coding buddy, you know, something, build like the such a minimum viable product that there's no way it could stand on its own right? It couldn't go viral, it couldn't have all the you know, website mechanics or whatever. But it's enough for you to sit there in a room with one of your clients and have them at a command line, you know, practice factoring the, the fractions, or whatever it is. And if you can prove that, you know, you're able to maybe charge a couple more dollars to your tutoring students, whatever, you're, you're continuing to fund the development of it through a services business, it's gonna take longer, but in the process, you're building a business, maybe you're hiring a couple other math tutors, and your dog fooding the product or I use that word, Google right to eat your own dog food, right to use it yourself, which is just the fastest and best possible product feedback loop that you can think of, right? You're using it both as yourself as the customer and you know, sitting in the room with the customer. So whether Yeah, it's tutoring, whether it's, you know, you're selling to schools, and instead, okay, you've got a consulting business for schools, but this is sort of your own tech enabled thing. And maybe it starts as a Google spreadsheet and evolves from there. Yes, sort of figure out a way to build what you want in a way that that funds itself with your time, if the money isn't going to be as venture available.

Alexander Sarlin: 28:53

Yeah, it's terrific and very thorough advice from somebody who's sort of been on both sides of that world. And you mentioned sort of starting a services business from the b2c side, like a tutoring business with customers. I've also heard some entrepreneurs talking about, you know, starting a services business from a b2b side, that could be consultants, professional development training, there are ways to use your sort of budding product in a broader context. And I think what you're saying really resonates a lot, you could be using your fractions product to teach teachers how to use fractions and getting pretty good money for that, which can then be used to go back into the product and make it

Andrew Cohen: 29:30

pay exactly and it forces you to be building like the thing that you need the most just the very next feature that's going to serve you the best rather than that you think is going to maybe help your market the best and before you know it, and you know, a couple orders are yours, you are going to still have that great product, but it was it's 100% solving the users problem and not just you know, some big bet anymore.

Alexander Sarlin: 29:51

Absolutely. You mentioned something in passing that's a little bit close to my heart and it's an interesting one which is that you know for quite a while the concept have sort of learning facts and memorizing just really looked down on in a lot of educational circles, it was considered old fashioned, it was considered almost like a thorat Aryan. It's just this feeling of like, you know, Hey, kids should be playing and learning through construction and constructing knowledge doing all this stuff. But you know, why would you have to know the facts in a world where with Google and World of Internet, and I think it's really been a pendulum swing back. And specifically, I want to ask you something. I'm curious about your thoughts on this. So one of the thoughts that a lot of educators have been having in this new world of AI enhanced learning where students can access, you know, incredible amounts of knowledge beyond even a web browser, people are thinking that one of the future types of assessments might be a little bit more of a sort of live assessment, a oral defense, an in person evaluation, a performance tasks, something that you can't, you know, quote unquote, cheat on with an AI. And that actually takes us much further into the world of you have to know your stuff, more than you have in quite a while. I'm curious as somebody who has been sort of advocating and championing knowing your stuff, knowing the facts, knowing the declarative knowledge, being able to build on a corpus of understanding for over a decade, I'd love to hear you talk about you know, why you've been championing that and what it means right now, in this strange moment.

Andrew Cohen: 31:26

Now, I love that you brought that up. I mean, I remember several classes back at Columbia, where I was the lone person that was defending the value of memorization, right? People are saying, Oh, you all you want to drill and practice, you want to turn all these kids into, you know, robots, that all they do is just regurgitate facts and memorize dates, when they should really, you know, be developing real skills and projects. And I think you can have both, as they say, in improv, yes, and the facts are important. You can't really make mental associations between concepts, unless you actually have the concepts, you know, sort of available in your brain. I think the pendulum, you know, what, what people really were against, and maybe they didn't, you know, articulate it as well as they would have liked. But it's having memorization in the classroom. Because if you have a teacher who's sitting out in front of their classes, okay, who knows the answer to number seven, and you raise your hand, then yeah, that's a waste of everybody's time. Because half the class, they already know that so well. So there's just they're bored, right? Of course, I know that. And then the other class, like, maybe they're so far behind, that it's just out of their zone of proximal development, and what they really shouldn't need it is like, remedial. So yeah, take the memorization part, take the knowledge acquisition part out of the classroom, because that's what technology is perfect for. It's perfect for personalization, it's perfect for you know, it's on your own mobile device. It's on, it's on the web. And it's you know, with your own logged in account, it's your personal tutor, it's solving that, you know, two sigma problem that Benjamin blooms posted, and you know, the 1980s where, you know, tutors are able to do so much better than just, you know, a student on their own. But technology can actually can actually solve that problem better than anything, and then use the class time for the collaborative activities for the project based learning for you know, maybe some, you know, teacher super remedial effort as they can. So, yeah, I love that the pendulum is now kind of starting to switch, you know, further back to memorization, and I hate that word, because it sounds like it trivializes knowledge acquisition, important, and then that AI is pushing in that direction. Because like you said, you know, if I can very easily use chat GPT to write my paper for me based on my prompt, it's a lot harder for teachers and professors to trust that, you know, you genuinely wrote that paper. But if you have to write that paper in class, or if you have to, you know, take an assessment in class, whether it's oral or written, the only way to do it is to actually get that knowledge and the skills and you know, obviously, deeper understanding into your brain. And the only way to get it into your brain in an efficient and permanent way is through the practices of active recall, metacognition and spaced repetition. There's just there's no way around that

Alexander Sarlin: 34:09

very, very well said, it strikes me as I listened to your talk about that. This might be one of the reasons why language apps are so dominant in the you know, mobile space for ad tech, because the language is one of the only one of the only subjects besides medicine, which you mentioned, in which memorization and knowledge acquisition and knowing the sort of pieces is very, very highly valued. People know, they need to know the vocabulary and 1000s and 1000s of words in a new language to be able to use it. So I think language has gotten a little bit of a slide in terms of that perception. And I think it's led to actually some incredible improvements in apps like Brainscape and other apps, you know, flashcard apps as well. As you know, the Duolingo is in babbles of the world, people are just much more comfortable learning language on their phone than they are many other things and I feel like that might py? I don't know. Do you agree?

Andrew Cohen: 35:01

I do agree. Yeah. And there's, there's all kinds of personalization that you could also do with language. And I really like what does it memorize. So I believe you had, you know, the founders on, they're doing a lot with kind of like real life scenarios and personalizing it for you know, here's, here's what this, like, this user needs to get a haircut, or has long hair, and they have to ask, you know, these types of questions. Whereas if, you know, if I'm in a classroom, and like, you know, 70% of the kids in the class, like have buzz cuts, like, what a waste of their time, like, they're not going to be caring, they're not gonna be engaged. So you could do so much personalization. And then yeah, if you could layer the spaced repetition, on top of that, and right, you know, personalize the content, you're gonna be learning and doing it in a way that is going to invoke those the self assessment that metacognition space repetition that showing you the stats for you know how you personally know each concept and you know, what you might personally need help on the mobile really take set to a higher degree. And I think the, what we've been learning from language apps is probably going to continue bleeding into other subject areas, as the educational establishment becomes a little bit more accepting of knowledge acquisition being a key prerequisite to fuller understanding. Yeah,

Alexander Sarlin: 36:13

I've heard a number of companies say they're the Duolingo of blank fill in the blank subject are, and you definitely hear that I'm looking forward to a world in which Personalized learning is based on the length of of students hair, I think that's just so interesting. Vision. But you know, personalized learning, like memorization has been a term that's been thrown around, beaten up used to being a lot of different things, adaptive learning, too. And I really agree with you that I hope we as a community can sort of do a hard reset on the concept of personalized learning. Because in this world of AI, and in this world of, you know, having a lot of data and a lot of personal relationships to apps, and you know, personal accounts, personalized learning really has a major place, I'd love to hear you have a very specific kind of data in your app that I think is probably unique, which is this confidence data. You know, as you mentioned, in Brainscape, people are studying particular facts and pieces of knowledge and skills, and you know, you small pieces of things, but for every single one, they have to keep telling the system how well they know it. And that means you have a literally, you know, over time, these incredible graphs of oh, they didn't know it, they didn't know it, then they started to know it, then it went down, then they really knew it. And then it stuck with I mean, that's data, I just don't think people have that sort of self reported, you know, confidence data. How might you use that in the future to develop some really interesting features for Brainscape, especially potentially with an AI model?

Andrew Cohen: 37:48

Sure. Well, first, you know, the concept of scaffolding, which typically means starting with easier content and getting progressively harder, you know, once you kind of reach that that learners zone of proximal development, you'd be able to start any flashcard deck or class in Brainscape. With the easier flashcards well, who the hell are we to decide if we're just creating the flashcards from scratch, which are the easiest ones, right? What's the easiest national capital? What's the easiest bone for people to label? Maybe the femur, maybe the tibia? I don't know, it sounds hard. But we have the data to actually, you know, the owner has the highest amount of views that are required for a user to get from a confidence level of one to a five. And so because we know that that's hard, we're probably not going to make the you know, the first flashcard, we're going to ask the phone by labeling back, right? Because then the user would get discouraged, you're starting to heart start with something that is, you know, very easy, okay, the spine, right, that's number one, right, and then get progressively harder from there. So that's kind of one way that we can use user confidence ratings. Another you know, that people have thrown around the term AI very loosely for probably the last like 10 or 15 years, typically almost synonymously with just sort of personalized learning or like somebody that feels like a conversational or smart like human computer interface. But, you know, really AI technically more like a machine learning loop, right, where there's some sort of feedback loop that actually makes the system smarter and smarter. And the the advent of generative AI that sort of has been consumerized you know, recently through chat GBT and other interfaces that allow users to actually create content at scale, because real AI models have been trained, you know, with machine learning, develops new opportunities for both Brainscape to develop certified content ourselves and, you know, with with our certified content partnerships, in a faster way, maybe that you know, we'd still have to obviously be very, very heavily checked with humans in the loop but you know, might actually enable us to scale and we've been playing with that. But as we you know, train our own models and do our own trumped engineering for Brainscape certified subjects for which we now have about 140. As of this month, we're going to be able to take those prompt engineers and kind of apply them to the user generated flashcard engine. So if I'm a user, I create a thing for, you know, chapter three, or you know, lecture Week Three lecture or whatever the deck of flashcards is going to be, I could just take my class notes, right, my google doc by slide deck, my you know, maybe maybe some content, you know, from the original instruction, paste it into, you know, some kind of Canvas and say, Make me flashcards from this. And so it could save me a half hour just for, you know, one chapters worth of content of having to make my own flashcards, and it'll spit it out as like, actually very good, high quality questions and answers that are asked in different varied ways with, you know, rich language that kind of forces me to think about it, and can, you know, hugely accelerate the pace of user generated content development, which will both, you know, accelerate users learning, but will also make it more accessible for other users who are studying similar subjects to discover more and more content on the Brainscape marketplace. So those are three, three areas where I'm really excited about AI, and the tools that are increasingly available for us to make use of it.

Alexander Sarlin: 41:24

It's such an amazing moment in edtech. And it feels like this, as you say, the consumerized generative AI, I think that's a great way to put it is just making people's imaginations go a little bit wild about what could be next. But some of the use cases I'm hearing from you are exceedingly, you know, realistic. They're right here, right now, the idea of being able to pull, you know, entity extraction and pull things out of existing content to make confident flashcards, definitely to create new content to be able to take these datasets have the users the information, the confidence, the time, and being able to sort of take the complexity of that and start to find out the kind of thing you're saying, you know, what, which of these concepts is actually the hardest on average, and the easiest on average for people to learn, and which are the hardest and easiest for this individual to learn? Because we can learn that over time, too. It just really boggles the mind. It's really It's exciting to see it from all these different sides. Brainscape has been around for over a decade, it's added so many amazing features. And 140 certifications, almost I can't even imagine that there's there's so many tests out there. Like what is next for Brainscape? Where are you headed right now? Clearly, there's some AI plans. But what do you think is the sort of next chapter in the Brainscape story we've heard about the VC era and the bootstrapping era and the profitability and the expansion and the distribution from COVID? What do you think is the next chapter for Brainscape

Andrew Cohen: 43:00

instruction, there's sort of an education there's, there's like a continuum of you know, where the knowledge comes from, and then how it sort of gets validated. Right? You, you start with the instruction side of the continuum, so that you're, maybe you're reading a book, maybe there's a lecture, maybe there's some sort of, you know, project that you originally can do to have to seek out that knowledge. And then there's on the other side of the spectrum, there is assessment, and how do we actually validate that the person knows it? And maybe it's a skills based exam, maybe, you know, it's a paper, maybe it's oral, but in the middle between kind of the original instruction or exposure and the assessment sits studying. And studying it typically is like, just been ignored by a lot of the establishment, right, like 95% of the of the institutional budgets are in instruction and assessment, and studying, like, go figure out how to study on your own right? Yeah. Parents and students, they'll go to Staples or Office Depot to buy their their notebooks to take notes to study their highlighters, their flashcards, right? Oh, you want a Brainscape account? You want Quizlet you want Chegg? Well, schools aren't paying for that. That's the student. So studying is just this like bastard stepchild. That is this sort of, you know, up to the learner. Yeah, and, you know, really, for Brainscape to really make a bigger difference in education than we already have. Right? We're lucky we've had 10, almost 10 million registered users, you know, sign up over the years, we've you know, made a big impact on people through their own volition because they're, you know, a very self motivated learner who chose to use Brainscape. But if we want to be part of the system in a in a bigger way, we need to better integrate, you know, either with kind of the, the assessment side of the spectrum, right, become more of an Assessment Engine ourselves, somebody that could be used, you know, institutionally or alternatively kind of integrate more with where the instruction comes from. And so yeah, I mean, transparently. Like years ago, I was in conversation With a lot of, you know, publishers, right trying to get the big guys to to buy Brainscape, right tack on Brainscape with your textbooks, right? So you're reading the E textbook, and you've got the flashcards already, right, they're aligned with the content, rather than having to, you know, make your own or put them on a on a platform. And that didn't work out. And actually, I'm so glad that it didn't. But now that, you know, we're on much bigger footing, and you know, we're becoming independently stronger, you know, might that be something that we can find better ways, whether it's through partnerships through integrations, and we're through kind of becoming a little bit more of where the instruction comes from ourselves. And it's sort of evolving from flashcards to more of kind of a reading or a flash book, you know, kind of fusion is something that we're thinking a lot about, you know, it's not, we're not doing it tomorrow, but we're increasingly well positioned to make some big plays and integrating Brainscape, you know, more with kind of where instruction comes from,

Alexander Sarlin: 45:56

I love that. And it's a very interesting way to break down the landscape. And I 100% agree, I'm sure many of our listeners do as well that you just don't hear a lot about Ed Tech for studying. And when you do, it's, it's considered a fully consumer optional product. And as you say, people are still going very, very analog for a lot of this.

Andrew Cohen: 46:14

No, and it's where all the learning actually happened. That's like, where you internalize this stuff for good. And it's like the education system has just forgotten about it, there's no institutional budgets for it. Like you can't sell Brainscape or Quizlet to like a state of Georgia, there's like a $3 million, you know, license for Quizlet. Like, they're not making these deals. And it's because there's just no line items for it. And it's ridiculous. Yeah, either that needs to change, or we need to just figure out how to bring studying where the content comes from,

Alexander Sarlin: 46:42

it's a call to action, I hope people listening can start moving the needle on that for procurement or for state budgets. I mean, I think there's a big opportunity, as you're saying, to just really lean into the active learning piece of it. I mean, the idea of, you know, reading an online chapter, versus consistently answering questions as you're learning, we know from years and decades of research that the latter is significantly better, especially if you're doing retrieval practice and the way that Brainscape does, you get very little from reading and rereading. And there have been some great studies about that, and you get a huge amount from testing. So it is really odd that these things are seen so differently, I think it's just a very antiquated way to sort of break down the different pieces of the learning cycle that I've never thought about it that way. It's really, really interesting. We're coming up the end of our time, and I want to ask you, the questions we end with you have been in this space for a long time, you've observed it very closely. Outside of AI, let's put that on the side table here. What is the most exciting trend that you see in the landscape right now, for edtech that you think our listeners should keep an eye on?

Andrew Cohen: 47:53

I'm actually really excited about the innovation in new K 12 school models. Overall, that has been happening over the past several years. So just throwing out everything that you that you thought of as a school being this, you know, big building with departments and sciences over here. And this is where my social studies classes, and oh, the bell rang after an hour. So now I'm going to this other class with another teacher and just totally changing on its head and making it a lot more student driven. So two interesting models I've seen when there's Alpha schools, right, Joe Lee, man, legend, tech entrepreneur, started a network of schools out of Austin, and now they're expanding to other places where it's all project based, it's self directed, right? Students don't have teachers or periods or classes, but they have coaches, and they determine what technologies that the students are going to use or what content that they're going to use to actually pursue their own goals, but still aligned to the curriculum requirements that they that they have. I think that's really interesting. And I'd love to see more models like that, you know, get funded and you know, maybe, you know, become kind of, you know, more charter schools or even or even new public school models eventually, even homeschooling So outschool is a really interesting platform, you know, out of Silicon Valley for small live classes that really facilitate homeschooling models. And you know, I actually just right before this podcast got off of a the Brainscape community team, right we have a Discord server and this kid Denver, he's a high school student who has been homeschooled and I was like, Oh, I homeschool kids you know, you have this like concept of like, you know, homeschool are they are they're really getting a full education. He and several other you know, recent homeschooled kids I've had interactions with have impressed me more than almost any other, you know, young person that I've met in the last couple of years. And it's because just the tools for homeschooling is just so much different than, you know, the kind of, you know, far left or right religious families that you know, we're kind of taking their their kid and sheltering them as sort of what we used to think of as homeschool. So there's a lot of really progressive things that are happening and these new models and I think that it's the In smaller environments with fewer rules that are going to really, you know, not only allow kids to better explore the paths that they want, which may or may not even lead to college, but will still lead to, you know, huge success and leadership and everything, but also the lead to more experimentation and innovation. Because, you know, they're able to try things that maybe the establishment might not have been able to try as much and a lot of other innovations that, you know, whether it's math, whether it's AI, whether whatever the tools are, are probably going to, you know, have a lot that are evidence based kind of trial through those simpler platforms,

Alexander Sarlin: 50:35

plus one times 10. For that, that is a really, really interesting trend. And, you know, the recent NWA MAP scores just came out, I think, just today or a few days ago, showing that we are just not bouncing back from the pandemic, it's really amazing schools have just think they say, students are four months behind, on average. And in math, it is really bad out there. And it's I agree with you that it's so refreshing and exciting to see people who are willing to sort of toss some of the old models of what school has been for all of these decades and centuries, and say, you know, what could it be in this amazing new world where we have so many technologies, and I agree with you, I think homeschool, this sort of homeschool? 2.0 Maybe it's even 3.0 is really exciting. Because I think, you know, people, especially a lot of families who left schools during the pandemic and realized, Wow, there's so many cool things I could and I'm doing with my students right now, let's just keep going there. I hope there's some real, you know, epiphany moments where the school system starts looking around outside of itself and says, Oh, wow, people are really are have really moved the needle, and we're just continuing to try the same pass. I'm with you. And last question, what is one resource could be newsletter, blog, Twitter feed a book, of course, that you would recommend for people who want to dive deeper into any of the many topics we discussed today,

Andrew Cohen: 52:01

I might suggest Brainscape. So branding is very easy to find out about Brainscape website and mobile app, obviously, to learn anything with the product and really get a better sense of you know how all these principles work, you want to follow the company in our news, you could follow us on Twitter at Brainscape or LinkedIn at Brainscape. But outside of just following Brainscape if you're interested in just the principles of cognitive science, for learning, for motivation for focus for productivity, and brain health, and everything you know, related tangentially to you being a more successful learner, then check out the Brainscape academy or Brainscape YouTube channel where we're just constantly pumping out lots of great thought leadership and interviews and research about how to better rise to your challenge and be a better learner in general.

Alexander Sarlin: 52:50

Fantastic. We will put links to the Brainscape app to Brainscape Academy on the show notes for this episode. And you can definitely find them in the app store. I hope this podcast if you hadn't heard about Brainscape before this, you realize how much is going on in there that you've been missing. Andrew Cohen, thank you so much for being here. You're an old ed tech colleague. I think you had the term ed tech insiders many years ago, if I am correct, and I've sort of siphoned it off over the years. I really appreciate chatting with you. And I think you know, it's been a terrific conversation.

Andrew Cohen: 53:22

Thank you. Likewise.

Alexander Sarlin: 53:24

Thanks for being here with us on Ed Tech insiders. Thanks for listening to this episode of Ed Tech insiders. If you liked the podcast, remember to rate it and share it with others in the EdTech community. For those who want even more Ed Tech Insider subscribe to the free ed tech insiders newsletter on substack.