Transforming Education with AI: Insights from Stephanie Palmeri and Caleb Hicks

None of us want to send our kids to school to be on a laptop all day.

Ed Tech, broadly speaking, is going to play an important role in things like upskilling and reskilling of workforces.

Schools are usually five to 10 years behind every wave.

We're not going to disrupt all of education. We're going to bring the best that Silicon Valley tech has to offer.

This is the most mission driven industry.

In the world, creating environments for students to thrive. Right? And if you do that, everything else follows.

Now it's not that hard to become a rocket scientist.

Students have to know how to use.

This stuff, create a great end user experience.

AI is not just an assistant, it is a.

Welcome back everybody to another episode of the venture capital podcast with your co host, Peter Harris. That's me and John Bradshaw. We are excited today to have as our guests Caleb and Stephanie. Caleb is the founder and CEO of School AI, a fast growing edtech startup reimagining how teachers and students interact through AI powered learning spaces. With a background as a schoolroom teacher, Apple education team member and co founder of Lambda School, Caleb brings a rare blend of frontline experience and tech innovation to the future of learning. His work focuses on restoring human connection in classrooms while using AI to give educators back their time and insight. From turning middle school keyboarding into a startup incubator to raising one of Utah's largest series A rounds, Caleb is reshaping education with empathy, creativity and bold execution. Thanks for joining us, Caleb.

Yeah, of course. Happy to be here.

Stephanie Palmieri is a partner with Next Few Ventures where she backs visionary founders building the everyday economy, businesses that leverage AI and software to reshape how we live, work and thrive. With over a decade and early stage investing, she's led seed investments and hits like Poshmark, Clever Class Dojo Rap Book and Carrot Free Fertility. Stephanie has been investing at the inception stage in companies for the past 15 years and brings deep expertise across SAS marketplaces, education and healthcare. She's also a certified coach and a founding member of All Rays and serves on the board of Columbia Business Schools Lang Fund. Thank you Stephanie for joining us.

Thanks for having me.

Welcome, welcome. Awesome. Well, to kick things off, you know, one of the things that we saw in the last few years is this whole like transition from like Covid to AI. Right. So education, particularly from an investing perspective, you know, like our firm, we were early investors and Instructure, so you know, been watching like Ed Tech for a long time. We're shocked I think as many to see like this huge, huge like boom, during COVID in ed tech is all these schools scrambled and then as those but like to solve virtual learning and so forth. And then as that like fell off, then you've had this new wave of AI. And so maybe where we started is like Caleb, I'd love to understand and for our audience what you're working on with school AI and how you came up with that idea and how that kind of fits in this broader landscape of what's happened in education over the last, call it five years.

Yeah, I think you, you mentioned this. I was a classroom teacher. I taught middle school. I think a thing that most of us don't know or appreciate about middle school and high school teachers is just how many students they actually have. So I had 42 desks in my classroom and I taught seven or eight periods a day. So if you, if you do that math, I had almost 300 kids walking in and out of my door every day. And I was like expected to know who they are and be able to do something for each one of them every day. And teachers are put in this like impossible choice every day of hey, I have 10% of kids that like really love this. They want more. They, they, they would come in during lunch or after school and they deserve that extra time attention for school to be for them. And then you have the 10% of kids that are really struggling. Maybe it's things at home, maybe it's learning disability, maybe it's, they just don't like the class. There's, there's like and, and they also deserve someone to dial in some attention for them. And then you have everyone else. There's 80% of kids and they're somewhere in the middle that they also deserve school to fit them. And I think a lot about that 80%. Most teachers get pulled based on the school that they're at to that top 10% or that struggling 10%. Almost nobody gets to think about the middle 80%. And I think about that group because I was in that group. I was a good kid, I was a good student. I got A's and B's all the way through high school and college. None of my teachers knew who I was until high school when one teacher almost convinced me, convinced me to become a doctor and I fainted. The sight of blood. So like it was just like complete non fit. But there was like one teacher like had it dialed for me. So I just think about how do you create connection inside of these massive groups of people that are supposed to be working together. My work At Lambda School we did that. I had a 1 to 8 teaching assistant to student ratio. Every single thing that happened with, we did a one on one with every student. Every day we were surveying students, every hour we would send them a slack with six emoji and they would pick one and tell us why. And we would go and handle every student that needed that help that can't exist in classrooms today. And when you look at AI, we all get excited about, hey, AI saves us time. It helps me like in schools write a lesson plan or differentiate content or whatever. I think we can, we can think about how AI allows us to save time or we can think about how AI can do something new. And that something new for me is how do you close those feedback loops between people? How do you help students feel seen and heard by all of their teachers? How do you help teachers see the student that needs them today, right now? And so I can talk more specifically about how we're doing that, but that's ultimately what we're building. We like to say we're making school awesome for students and, and AI is a tool by which we're going about doing that.

Yeah, I love that. One of my like central thesis is around AI is that it enables. And part of this was influenced by you in conversations that we'd had before as you were kind of describing this, this, this new power that a, that AI was really unlocking for education. And that is AI enables like customization and personalization at scale, which I think is just super interesting and all the different verticals where that's applicable. Um, Stephanie, I'm curious on your side what, cause you were an early backer of this, what, what got you excited about what Caleb was building and, and how does that fit into, you know, the question I think a lot about when I look at deals, which is like that why now? Question.

Yeah.

Why is the timing now?

Yeah, well, I mean, I think Caleb probably even just opening up for us explains exactly why I wanted to work with Caleb. Right. Because you know, he has such amazing both intuition compare along with that on the ground experience as a teacher and a product leader. Right. And that just, that's a really, really, really unique thing to find in a founder. And here's a founder who, you know, deeply understands education and deeply understands technology and how to bring those two things together in the classroom. Right. So like, you know, first and foremost, like Caleb is the reason I backslide AI. I think the other point though to make is I did was fortunate enough, I could say, to have backed A number of ed tech companies in sort of the previous wave of ed tech around the time that mobile and tablets were starting to emerge in classrooms and the schools were going through that transition of we have Internet in the classrooms now we're about to get to, we're moving towards one to one devices. And that was the time when I backed um, class dojo as teachers were starting to bring like these into. They had their computer in their pocket and then clever, you know, shortly following that. But then I didn't lead an edtech investment for a decade and in part because I had invested in these companies that had created a certain amount of momentum and mind share and, and were arguably leaders in their categories and in part because there wasn't this next wave or phase. And I think the unique piece here is certainly you talked touched on this a bit like there's that Covid bubble, like first and the whiplash, but that was really about like solving for an immediate problem and moment in time as opposed to solving for the real needs that were happening in classrooms, whether those are virtual or physical classrooms. Right. And so what happened with AI is a new technology is unlocking the ability to bring a different kind of transformation to all industries, all aspects of life and absolutely, certainly within education. And so for me personally, I looked at this as a time to say as a seed investor, there's going to be a period in which companies will emerge that have the chance to build truly transformational platforms within the education category. And you know, I kind of had to bet ahead of how quickly our schools could adapt AI. The thing I'll share is that I have been really excited from an investor perspective to see how excited educators and educational leaders are to bring edtech into their schools and into their classrooms. And I think some of that is the initial foundations are already there. Right. You know, we're now 10, 15 years past the point of mobile, tablet, Chromebook, you know, distribution in classrooms. And so the access points are there and we're able to do something with the access points. We don't have to build that infrastructure. So. So now we can create the really interesting transformational value.

Yeah, no, I love that.

I think. Hey Peter, real quick, the, a couple of other why now thing like the, the what AI actually unlocks that Stephanie mentioned and a little what I talked about kind of zoomed out.

Yeah.

AI can finally fulfill the promise of remembering how every student is doing and adapting to that. Like that's the holy grail of education. Yeah.

Truly personalized.

Yeah, it's, it's, it's why it's so hard to be a teacher with 300 students or even an elementary student with 26 students. Imagine, you know, the people listening to this. They've, they've, they're, they've, they're either working in companies or managing people and things like that. It's like imagine having 32 direct reports that you're working with eight hours a day in the same room. Right. And AI can actually help you personalize to them, but also when they're working with it, give you the right degree of insights. I think the, the other important thing for education at Tech why now is schools are usually five to 10 years behind every wave. Ever since the personal computer, personal computer schools were on the front line, but since like the last 30, 40 years, schools are usually 5 to 10 years kind schools were very much on the front lines of AI disruption and, and changing the game because Chat GBT came out. One kid that I was working with won $20,000 in scholarships within 30 days of chat GBT coming out. Just having it write his scholarship essays for him that next week, right after New Year's hit that, you know, New York City bans all AI and basically every school district in the United States followed suit immediately. That's not a sustainable solution. You can't just ban it. So what do you do with it? That's a, that's a. Why now is like schools are on the front line that students are either cheating or schools have to adapt to the fact that like the way we have measured progress in student learning is their ability to write a pretty basic essay for 20 years and so. Or longer. So that's, that's another like, for this. Why this exact moment.

Yeah, no, I, yeah, I like that a lot. And I want to come back to that because I think there's some interesting discussion points around. How does education evolve in the world of AI? When to your point, so much of education has been based on your ability to write an essay and submit it, and now that's, that's being challenged, if nothing else. Right. But before we kind of dive back into that, I do want to spend just a little bit of time talking about ed tech investing because I think it's interesting what you said, Stephanie. Right. Made a bunch of investments. I did the same. Right. Made some of those investments, turned out really well, and then 10 years, didn't really do much in ed tech. And I know, you know, I was talking to a friend of mine, another vc, and he's like, he's like, yeah, every, every VC has like their gaming, you know, like a video game startup that they backed that failed and their ed tech investment that they backed and failed and then they don't do anymore. I'm curious, why do you think it's been such a, a bit of a graveyard over the last decade or what have you. And, and, and I guess going back to like timing, like why is now different? Right?

And yeah, I mean I think first off there's a lot of, you know, let's use the frame of K12 ed tech, right? Because I think there's, you know, different ways we can lenses by which we can look at this. But I think that, and this is true for startups in general, right? There's a lot of startups that get started that are really great products, but they're point solutions. And I think within the education world there's the opportunity to use technology for teaching and learning. But a lot of those solutions, like the best way to learn seventh grade math is not necessarily a venture scale company, right? And so, and I think, and I think the other piece that's interesting here is for founders, a lot of founders, I think get started in universities building tools to adjust the needs of students, right? And one of the unique things we all have, I bet the one thing all of us on this call have in common and probably your entire audience is we all have some experience with school, right? So every, we all bring our own, you know, we all, we all have is a universal experience, primary mostly right in the, here in the U.S. right. We all have this experience with school, very different experiences, but we all know what it's like to be a student. We all know what it's like to be a learner. And so I think it's an, I think the ability to build technology solutions to enable students is actually a really interesting gateway for a lot of entrepreneurs. And I beg you, if you were to go back and look at a lot of founders who now have venture scale companies, I bet you they all built tech products when they were in school that were probably used as students, right? I bet a lot of them had companies that maybe they didn't raise venture for, but that were actually. Right. And so I think this is like an interesting framework to think about which is the barrier to launch some of these products is relatively low because one thing we're all experts in is being a student because we've all spent, you know, many, many years being students, right? And so, so I think that is, is one element of like why the graveyard? Well, a lot of the companies started are not necessarily companies that can or should be on a venture scale path. Right. And so what is a venture scale path? A venture scale path is a company that has the ability to, you know, get to hundreds of millions of dollars in revenue, you know, that is going after markets, you know, in my mind, that are billions or more. Right. And there is the ability to access those budgets within K12 and the broader education universe. But there's very few companies that actually build products with the kind of scope and scale that actually can be venture back. Right. And again, I think it's easy to point to edtech as a category where there's a lot of smaller companies, very viable companies built. But, and that's because again, you know, we all kind of can relate to those products like we as students, ourselves at home. Right. And so it's, there's this interesting universality to education that I think maybe we're more aware of. Hey, this should work. And there's lots of solutions out there, but the reality is like, I don't know, there's like something like 3,000 edtech products out there that every individual school might be using. Right. That's a lot of products. But, but those products alone are not venture scale companies.

Yeah. And until, until 3, 4, 5 years ago, you, if you were selling technology, like ed tech products, software products into schools, you're, you're, you were selling into two or three classrooms per building.

That's right.

Right.

Maybe the library.

Right. And that's, that's, that's changed. Almost every kid in the United States, at least once you get to middle school and high school, has their own Chromebook. Even most elementary kids, there is a rack of Chromebooks in their classroom. So that just the ability to actually reach students is where is where a potential for venture scale products is because you go from, you know, 5 million potential users to in, in the United States at the teacher level to something like 50 to 60 million. And yeah, it's the, you go from being able to address one very specific line of funding for schools to many more of them. So that's, that's another thing that has changed. You know, we talked about the COVID wave and things like that. That's, that's part of that transition.

I think one of the other things I'll add to that is that one of the reasons it's so difficult to build really big scalable companies in edtech is also that you're serving a lot of different constituencies. Right. And especially for products that are sold with a top down district level sale. But even one sold at the school level sale. Right. They're still being adapted if they're, if they're classroom products by teachers and students on the ground. Right. And so think about the complexities of building for an end user who is different from the buyer and what are the individual and collective goals that all of those folks have. And then you know, add to that, you know, the kind of historically known and oh, super scary long sales cycles, district sales, you know, again. And I think that's where you, the rubber hits the road and you separate the folks that start companies in edtech that deeply understand the complexities and have the tenacity to, to like build teams that can weather the fact that these are long complex sales cycles with a lot of different stakeholders and the need to understand stakeholders at all levels. Right. And that's what makes it hard. But that's why when you can build a company that's truly transformational, it can be incredibly valuable.

Yeah, I think like back to when we invested in structure, I think another piece of that that you kind of mentioned is I don't know when, when we invested in structure, I remember seeing at the time there were a bunch of Silicon Valley based startups that were like, we are going to disrupt all of education. And, and, and Instructure kind of had this interesting perspective of like we're not going to disrupt all of education. We're going to bring the best that Silicon Valley tech has to offer. We're going to marry that with how education buys today and how they think about making purchases and we're going to marry the two and kind of walk that line. And I think that's one of the things that I really like about what Caleb's building is he comes from the education side, but he's also bringing kind of best of breed tech and can kind of speak the language of educators while simultaneously giving them access to technology that they haven't had access to before. Because so much of the products to your point kind of sit on either side of this. It's either like we're going to disrupt the world and they don't know how to sell into that, that industry or it's, they're just not like that compelling of products or their point solutions. Right. Because they're built by an educator. So anyways, I think, I think about that a lot when we're evaluating some of these, these ed tech companies.

There's a, there's an aspect to that when, when I started so I was working with a bunch of High schoolers. One of them won $20,000 in scholarships, right? And I was like, okay, AI's here, Cat's out of the bag. This is a moment to, to go and figure out what is going to change in schools and how we can do that in a productive way versus a destructive way. So January of, of 2023, I set aside what I was working on, went all into discovering this, talked to everyone under the sun, superintendents, kindergarten aides, everyone in between. And I actually left that moment thinking what these, what we actually need here is data lakes and zapier workflows. Like it's like nobody, nobody's asking for AI. What we actually need is across all of these different systems, how can we put the data in one place so we can actually tell what works? Right? And then what I saw was what I felt when I was a teacher, which was like, none of these people are on the same page as each other. Teachers don't feel seen or heard by parents, by their students, by their admins. Right? Admins don't feel seen or heard by the parents, by their teachers, by the students. Students don't feel seen. Right. I can keep going, but it, but it's like, and at the center of that all you have students. But this is, this is the most mission driven industry in the world. Like the most well intended people in the world. So why is school not better today than it was 20 years ago or 30 years ago or 50 years ago? And it's, it's like it might be, hey, you can't actually do data driven improvement because either the data doesn't exist in the first place or it's in 50 different places and you can't actually pull it together. And no Silicon Valley data scientist or data engineer is going and working for Alpine School District in Lehigh, Utah. Right? And the. So, so how do you actually like. Oh, and the feedback loop for schools is insane. Most, most schools, they, they work all year for, to, to improve test scores on end of level testing. Well, you do those tests in May, then you send those off, you get the results in June. All of your teachers are now out of school and so you will see them again in August. But guess what? August school starts. So you didn't do anything about any of those scores. You aren't able to adapt to that. So at best you're on a two year improvement feedback loop on the summative like final score on how you did as a teacher, as a school. So how do you take the data that doesn't exist and get that how do you put that the data that does exist and put it in the same place? And then how can you use modern technology, whether that's AI or just good practices of data engineering and things like that, to surface the right data in a tighter feedback loop and let these well intended people do their well intended work to, to change again how, how students experience school. And like that's just talking about like modern technology practices like a lot of the rest of the industries in the world have figured that out, like data lakes don't exist in education. Some very, some, some select few districts in the country have started with the idea of like we're going to get all of our data in one place. A couple of states have started saying we're going to get all of our data in one place. And then you have schools saying well we don't even have that data in that format, so good luck. Right. And then I would contend we don't even have the right data in the first place. Attendance in grades do not tell a student story or give you any inclination of how to improve it. So yeah, lots, lots of opportunity and just hey, take what's worked in the rest of the world and, and bring it here. But the incentives have not existed because it's been so hard to sell into schools with a price point that matters. So if you can go and build a data solution for every enterprise on the planet or a data solution for the hardest one to sell into with the lowest ACVs, which one are you going to choose? But, but AI gives you an opportunity to get that data. It gives you an opportunity to process that data and surface insights in new and interesting ways. And so yeah, using, using that to get new insights and surface those actionable insights to pull teachers, students, parents, school leaders together is what I think is the biggest opportunity here in, in it's, it's not just, you know, build me a lesson plan.

Yeah, yeah. I'm hearing a lot of like AI can help in some ways like drive down the cost to something that makes sense and then also provides personalization at scale. Right. It's kind of that combination that unlocks like these big market opportunities here.

Think, think about what a tutor costs. Yeah, let's, let's go on the very conservative side that a K12 school might be able to pull off 20 bucks an hour for 20 hours a week. You're talking about a few thousand dollars, upwards of $10,000 per school year. That is on the extreme high. That's one tutor for 20 hours a week. That is on the extreme high end of what schools are used to paying for a classroom product like a point solution classroom product. Now imagine giving every single kid in your school their own tutorial that they can use on and off across classes, remembers how they're doing, remembers what homework they have, and is able to tie all of that together, but also address the students goals, the teacher's goals, the parents school goals and the school leaders goals all in every interaction and then start pulling those those like tightening those feedback loops between those people is like, okay, now $10,000 feels like insanely cheap. So there's also like this opportunity to do more and have a way higher impact and do it at a more affordable cost. Like it's things that weren't even possible today. Not just how do you do the same things, cheaper, faster or better, but how do you do things that aren't even possible today at all. We're working on a product for admins and the way that we're talking about it is like imagine you had 100 analysts going and digging through every system you have and drawing out actionable insights for you based on how your students are doing right now academically, but also experientially what they actually did in class yesterday. And they gave you a detailed report tuned to your interests and your like opportunities that you're trying to solve in school every night. By the way, forget the fact that your school district couldn't even hire an analyst unanalyst in the first place. It's just, it's, it's. Yeah, I've said this now a few times, but do the same thing faster or cheaper or do entirely new things. That's, that's the promise of, of AI in schools.

No, I love it. That's great. Any thoughts, Stephanie? Or.

Let'S see, what would I add to that? Probably not much. I think, I think the other, the other piece, and I know we'll probably talk in this later, is like, you know, we, we talk about this technology in classrooms, but at the end of the day, like we're not talking about removing the humans, right? Yeah, we're, we're talking about a better experience and a better experience for all of the humans in the classroom, both the students and the teacher. I don't think teachers, you know, to Kayla's point, teachers don't come into the classroom saying I really want to only focus on the top, the top 10 or the bottom 10%. I want actually to give the best experience to every student I touch in my classroom. And the ability to have a personalized relationship with, with each student is actually, I think, a real win for teachers.

Yeah. Yeah.

Well, I think about it too. And just, you know, you mentioned this earlier, Caleb, about that teacher that inspired you to be a doctor. Just how impressive, right? But like, but like, you made the comment earlier, like, oh, well, like we, we have all these people that we manage. And what would that be like if you had to manage all these people at once? But then you layer on, these are students, they're at their, like, most impressionable. Right. And, and the things you do with them now will have impacts throughout their lives. Right. For decades.

Yeah.

And so, like, the stakes are so much higher. And I think coming back to your point, Stephanie, is I also think about, like, will venture capital be disrupted by AI?

We could have a whole other podcast on that. I'm spending a lot of time thinking about that. So come invite me back and we'll talk about that one.

I will definitely take you up on that. But one of the things I think a lot about that I think is tied back to education is this like person to person piece of it. And, and that's really, I think, critical people want to engage with other people and have real relationships. And I think the AI can be an augment to that, but not necessarily a replacement for it.

So none of us, none of us want to send our kids to school to be on a laptop all day. No, none of us. Stephanie. Big day for Stephanie. We told us in the pre show, Stephanie, what happened today?

I dropped not one, but two new kindergarteners off at their schools this morning. So two very excited five year olds to be entering the classroom. And it's super exciting. And another, another podcast. They're Twiblings. They're biological siblings born 15 days apart.

Wow.

Ask AI.

Wow.

Okay. I'm a twin. So you are.

Oh my gosh. Awesome. It's such a cool, unique experience that they have kind of growing up with the same age kid. But I'm really excited for them to bond with, you know, make new friends in the classroom and I'll sort of bond with their teachers and, you know, all of the other adults in the school. Right. And that's a, that's a big part of what school is. And, you know, we're fortunate they're going to a smaller school, but even within a small, you know, how do you have that really, truly personalized experience with kids? And I think, you know, the ability for all children to have the potential to access that is really important.

And how excited are you, Stephanie, for them to Go in, put on their VR headset, and, you know, just shut off the world for.

Absolutely not. Absolutely not. And we actually talk about, you know, in my house, we talk about technology as a tool, not a toy. Right. And it's a tool, and it's a means by which to accomplish something. And a lot of the way that we think about how we might use it at home, but also how we want to use it, you know, in a classroom, is to further learning. Right. And to reinforce what we're learning, but not necessarily to supplant, you know, the experience of being in that classroom or of learning something.

So.

So I think you have to think, hey, how does this change the experience between the people? Will the job of teaching change over the next 10 years? The answer is yes. Right. But is it going away? No, it's. How do you let the teacher be more of a coach, more of a guide, more of the person that can look you in the eyes and say, we got this. Like, what do you want to do next? Let's go and make that happen.

And how, you know, think about it.

Fun opportunities there.

Yeah. I mean, think about it like, you know, my son loves space right now, and if. If he's learning to read and he can read about space, he's going to be really excited about reading. Right. And. And think about that. For every. For every kid, the thing that was your thing and the ability for you to kind of interact with an astronaut, if that's your thing, or the, you know, in a classroom setting to learn the lesson that you're learning, and the kid next to you might be really care about, you know, tigers, and, and. And it doesn't really. It doesn't matter what the thing is. It matters that you're. You're kind of, you know, progressing, learning. And I think the ability to kind of meet kids where they are, whether they're, again, like, very young, elementary age, all the way through high school. And obviously there's the. The experience beyond that.

Yep. So I want to push this a little bit further. I read a New York Times article yesterday, actually, about some of this, and.

I want to talk a little bit.

About the challenges of AI in the classroom, because they are real. Right. I. I think to Caleb's earlier point, the whole reason why school is adopting AI faster than ever before is because they also are probably the most challenged of any, you know, group out there. So I think there's a lot of different ways that they're being challenged. I actually am curious, Caleb, what are you seeing are, like, the biggest things that educators are concerned about. Your, your customers. Right. Whether they're educators, administrators, what have you. And, and, and how are they solving for those problems outside of AI?

You're, you're just saying generally speaking.

No, I'm saying because of AI. Yeah. What are the AI related challenges that schools are facing that they're most concerned about?

Cheating? That's, it's, it. There's the whole list, actually. I'll come back to it. I'll come back to a couple harder ones. Is the big rock because it, it impacts everybody that that shows up in different coded language, right? Some people will just come out and say, like, my, my students that are using AI are cheating. Some people will say it's plagiarism. Some people will say, I'm worried about critical thinking and, and losing the ability to think critically about the world. I think we should be worried about attention, right? You're, you're 12 years old, you get your cell phone and you're now tik tok tuned, right? Like it, it's like your attention span is now 90 seconds max, right? And so there's, there's like real challenges there. If you go and ask a teacher that started teaching before COVID students, especially middle school and high school students today, it's so hard. Like the amount of. They'll use the word respect or they'll use the word like they, they, they don't care anymore about anything. And like, that's, that's a really hard challenge in the classroom is getting, getting students to care. That's less AI. But it is related to AI because if a student cares about playing Fortnite tonight and they don't care about learning math, then they're just gonna have AI do their work for them, right? And whether they have access to it at school or whether they don't at the administrative level, like when you start talking about principals or districts, you're actually just having like, really pretty simple policy conversations. And when I say simple, it's like, it's like, what is, what do we even allow with AI? It actually gets into complex stuff. It's around cheating, it's around plagiarism, it's around ip. But then you start getting into student privacy. You start getting into. I actually thought you were going to talk about a very different New York Times article. There are increasing numbers of stories of, of people, not just students, but people having AI induced mental health episodes. Whatever you want to, whatever version of that, that is mental breakdowns, psychosis, where your best friend is, is now AI. You know, we can all think of people, and maybe even some of the same people that are going through, like, serious challenges as it relates to. To, like, relationship with AI. That's only once you get like. Like I talk about this frequently is like, there's a journey that people are going on in education as it relates to AI. It starts with permission and policy. That's where most of the conversations candidly still are. A lot of people have kind of moved on. They've got a statewide policy or a district policy or something like that. Then it's about productivity. How do I. How do I use AI myself? A lot of students are really annoyed that they're not allowed to use AI, but they know their teachers are using AI to grade their homework and to write all of their lessons and things like that. And it's like, okay, why am I just going to AI school then? Like, some school, some students feel that, but productivity is that next step. Then you start talking about AI literacy, meaning students have to know how to use this stuff. If you don't, if you are not a student that is using AI when you are applying to college, you're competing against that kid three years ago that I was working with that was winning scholarships over you, because he was. Whatever you want to think about what that kid was doing, the fact is he was doing it. I thought it was brilliant. Good for him, using the tools that he has. But also, it's a complex. It's a complicated access and equity question. So, but, but AI literacy, AI literacy. I already can't hire people that aren't using AI productively compared to those who are. And so that is now a critical skill. But you can only get to thinking about that if you figured out policy and if you figured out how you're using AI yourself. So you're not even. You're still not even talking about how AI can change anything. You not talking about how AI can improve learning. Probably the fourth, the foremost thinking 1% of school districts in the United States right now are, Are even open to having that conversation right now. And then, yeah, so this, it's. It's challenges up and down the board, but it, it starts with, with cheating, and then you get into mental health, student privacy, student literacy. Like, how do we teach our kids how to use this in a safe, productive, managed way? It's a lot.

Yeah, it is a lot. It's a lot. You know, it's interesting.

We.

So we, we recruit a lot of students into our program every year as interns, and we have them go through this assessment process where they have to analyze a company and then pitch us on whether or not we should invest. It's kind of this case study. And I was feeling like this is the semester where the submissions we get should be by far the best we've ever gotten because the students have access to these incredible AI tools. And I would say it was actually probably one of the worst. And most of the students submissions fell into one of two categories. They either fell into I didn't use AI or I didn't know how to use it well. And so what I submitted was subpar or I just basically outsourced it to AI and they couldn't, they would, they would have things on their presentation that they didn't even know. They couldn't even explain what it was. Right. I mean we had one student who was like, well, the pre money on this company is X. And I said, cool, what's pre money? And they were like, I'm not sure what that means actually, you know, and so I don't know. It raises this question around this whole cheating critical thinking. And I'm curious, Caleb, how do you think that gets, how does, how does that get solved? Right. How does a school AI help with that? Or other solutions where we want people to be able to leverage these tools because if you, you can't, you're toast in this economy. Right. But at the same time still develop like true critical thinking because you also need that too. Right. In order to perform.

Yeah. Think about how you learned to ride a bike. Did your parents just say, here's the bike, good luck. Did they put you on the bike and just push you and say figure it out? No. Right. We all had training wheels and I, I'm using training wheels. But I'm actually going to flip the script here in just, just a second. But like you have to give it to them in a way where they can use it in a non destructive way. Let's not put them on the bike and let them fall and break their arm. Let's give them the tool where they can use AI but in a way that, that like they can't do that. Right. I think OpenAI and Anthropic did the most predictable thing and the obvious thing for most people, which is AI is an assistant. One of the biggest things we've spent the last two and a half years doing is making AI not an assistant. Now it's an assistant for teachers, but with a student it's a tutor, it's a guide. It says no, it says how Right. Let's go through this together versus sure. Here's the essay. Right. One of the very first things that teachers wanted us to do for them was I want a writing coach that won't write the essay for them. And I think for a while we might have been some of the best people in the world at getting AI to not do the thing that it wasn't supposed to do. Now that you know, there's entire research teams and red teams and things like that that are doing that. And, and that's not like our core competency, but it, it's an important thing that like AI is not just an assistant. It is a foundational technology that can be used to be an assistant, but what you employ it to do in your school. Yeah. If you just give it to every student and say be an assistant, then they're not going to be very good at, at the surrounding skills that make having an assistant be leverage instead of destructive.

Right.

So how do you, you've got to give it to them in some sort of way. Well, we do that by saying, hey, you can create these fun learning activities or tutors or like there's lots of different ways schools or teachers can use school AI but like third grade teacher built a, A, they were teaching the water cycle, right. And they built a magic school bus style water cycle adventure where the student was, was going on the journey with a water droplet. Right. And so AI is not playing the role of an assistant at that point. It's playing the role of a storyteller. It's playing the role of a question asker, the role of a, like a creative outlet for like, what should we do next as a water, as a water droplet. Right. That is a way where AI can be engaging, captivating, pulling you into a thing rather than pushing you out of work. Right. All four of us on this call have used AI in a way where at the end we were like, I didn't even read that. It just did the work. And I was like, first paragraph, last paragraph. It was fine. Copy paste, right?

Totally.

It's all about how we, how we deploy. It's all about how we use it. So the biggest thing is we have to give it to students. We have to do it in a way that is like training wheels where it is not just stripping away critical thinking or stripping away the work, but having you develop those skills around using that tool so that when you do take those training wheels off, you know how to interact with it in a useful way. And you can do the thing where you get really good at saying like, like doing deep research with, on competitive analysis, things like that. It's amazing. I love it. It's so good at it. And if I know how to use it as that tool, it's a superpower. If your students had done that, but they actually knew how to prompt it, knew how to ask it knew and it knew what they knew and didn't know and was able to coach them on it, that's that that you would have had that result of. This would be the semester where these were the best submissions.

Well, and it is kind of this awkward transitionary period, right? In that we're all, we're all figuring out how does AI work, right. And there, there are crazy opportunities like the one you mentioned earlier, which is if you're the first one to figure out that AI can write your essays, you have this huge advantage. But that, that advantage today is gone. But there are other like interesting opportunities.

You'd be surprised how not gone it is, but that's.

It'S diminished.

I was doing a PhD in my life right now is exactly when I would want to do it. Not two years from now and not two years ago. I'd want to do it right now. That's right. Anyway, sorry Peter, I cut you off.

No, no, no, no. It's a fair point that we're still in this transitionary period where people are figuring things out. And as we come to the end of this, this conversation has been super fun conversation. Where are the opportunities? So clearly like you're building some really special stuff at school AI. But I'd like to talk a little bit about maybe where are the opportunities? Maybe within scale AI, but also without the AI is enabling an ed tech that you're seeing. And maybe we start with Stephanie on the investor side.

Sure. Well, first off, I mean, I'm excited to be involved in school AI and you know, obviously, you know, I think the exciting, one of the most exciting things about what, what Caleb and the great team he's built are doing is that they're, you know, creating environments for students to thrive. Right. And if you do that, everything else follows. So I think like, you know, the lens that I tend to look at things through is like create a great end user experience. Right. And then, and then, you know, hopefully you can find a big business around that, but create that great end user experience. You know, I think a lot of the lessons that, you know, Caleb is, is, is, is, is putting to work at school AI can be applied to both higher ed as well as Continuous, you know, learning and upskilling. Right. And so I think there's going to be tremendous opportunities on that front. If you think about some of the challenges that universities are facing, right. They're, they're trying to address things like enrollment, you know, driving ROI for back to the schools. Right. And how do you know, like if a student doesn't enroll in classes, you're losing, you're losing a very valuable student at the kind of university level. So there's a lot of things that can be done and lessons to be taken from what's happening in schools to understanding where a student who is a much more independent, in a sense, learner, who actually has the choice to opt out of a system in a different way. Our younger students can opt out by not fully engaging and participating in classroom. Our older learners, our adult learners can opt out by just choosing not to participate in general. And so I think there's a tremendous opportunity around for AI to transform how we think about engagement both in kind of traditional university settings and giving data back to schools and helping schools make better choices about how to support those universities, those students in those university settings. And then obviously, I think it's clear that AI is going to change the nature of work. It already is. You know, Caleb talked about it's really hard to think about hiring someone that isn't, you know, AI forward thinking, if not AI native for the team. Right. And we think about that a lot, you know, at next View. And I think about that with the teams that I'm backing now, you know, whether or not a company is truly an AI company or not in the same way was a company, a mobile company 15 years ago, you know, these teams are AI native, right. And they're, they're using AI to build their organizations and they're thinking about how they build product and organizations differently today. And so, you know, I think that Ed Tech, broadly speaking, is going to play an important role in things like upskilling and reskilling of workforces. As we think about how, what participation in the workbook looks like. And one of the cool things about AI that again, you know, we can do this with, with our younger learners, but think about the possibility for an adult learner to be not alone in a classroom or not just, you know, doing this, you know, very rote, boring online instruction, but to get real feedback that's related to their job at the right moment in, you know, and again, whether that's used it in corporations or whether I'm seeking that to get a job at another Organization, whether I'm using that to upskill myself to get the promotion. You know, there's a lot of different areas where I think AI is going to transform education and it may be fully baked into some of the other products that we use. Right. So, you know, I have a pretty broad lens by which I think about, you know, education and technology, but I'm, but I'm really proud to say that my first ed tech investment in 10 years is AI.

I love it, I love it. I think it's. Yeah, I think it's so great. And I love, I love the thoughts around retraining because, you know, I mean, like, I just was thinking about Sam Altman's quote recently. He was like, the, the cool thing with AI is now all of a sudden we can start like pushing the boundaries of the new types of jobs that we can have that we haven't been able to push before. And people are going to need to be educated on that. Like maybe now it's not that hard to become a rocket scientist and we can have millions of people going to space. Right. And pursuing education that enables those types of jobs and opportunities. What are your thoughts? Caleb is. As we wrap up here, where are the opportunities that maybe school AI is not currently addressing that you think founders should be taking a hard look at?

It's an interesting question for me to answer because I could name lots of different things, but we're going to do them all. What, what I mean, here's what I mean by that is like we're, we're fun. We are fundamentally looking at what does school look like? If AI is just like if compute is free and so how do you get AI to be invisible? Like if you're, if you're thinking, hey, AI is I'm chatting with a chat bot and I'm doing something the same thing faster I. There will be really impressive businesses made now and in the next five years that are doing that or that are using AI for detecting cheating. Right. People will buy that today. There's lots of opportunities in that. It's very red ocean. If I were a new founder, I probably would not want to get into that space unless if for a venture scale opportunity. If you want to build something for your community schools and go and sell to people, you know, lots of opportunities for that, go do that. But you'll be competing with every vendor on the planet, including Google and ChatGPT and Claude directly. Right. Maybe you don't want to be competing directly with them, but when you start thinking about like do new Things. I think I alluded to this earlier. I think data driven education is a lie. I don't think we even know the right things. I think we track grades and we track attendance and that's about as granular as it gets. Think about every experience that we have online. If you go and do a Google search right now and I do the same exact Google search, we're going to get completely different results because it's not, it's, it's not, it's personalized to you and your source of career. All the things that Google knows about you. 16 year old opens up Instagram, that feed is perfectly tuned to them, including the ads that they're going to see. That's because the data is extremely granular, the pattern matching is extremely good and it's just getting used for selling ads for the most part. We think that can change is like if you know how students are doing in the moment because they're working with your AI assistant or they're using your AI NoteTaker, they're using like these other AI native tools that people are wanting to be using in classrooms and you can understand the actual experience and attribute that in really granular ways to what works and what doesn't work. Then you start saying okay, let's make this system better. Let's do new things altogether. If I were a venture capitalist and I were in the last five years of my career, I'd be willing to invest in do things, same things, faster opportunities. There's lots of those. If I were betting in 10, 20 year time horizons, I would be thinking about what are new things we can do. Asymmetric upside and finding the things that can actually be real and that people are actually going to be want that, that people are actually going to want, but that are new things. If you know how, if, if your new atomic unit of data in a school is how every student is doing in real time, you can start pulling these four major stakeholders back together in really interesting ways. You can start changing how students experience school in a way that's way more personalized. You can help teachers actually love their job again because the reason they got into it is a light bulb moment they had with a kid during college and they now get those all day while they're teaching. Instead of once every three months when the kid might come in after class and say I really need help with something thing, you can give school leaders an opportunity to actually lead again instead of be reactive to whatever's in the moment is it's so we're we're, we're building a new operating system for school, which is why I'm kind of jokingly saying like we're going to do all of it because I think, I think so many of the opportunities are in that, like you just have to think about what does AI as a core technology, not AI as chatbots, but AI as a core technology, unlock within a thing that at any given time, 25% of humanity is in. Right. And I'll make one more plug for ed tech here. As we close and I know we're over time, a lot of people think about AI. EdTech is like there's no money in ed tech. There's no money in education. K12 kindergartners through 12th grade, not even college in the United States alone. We spend the same amount as the size of the entire e commerce industry like this. This is, there is, there is opportunity to make real impact and there are resources dedicated to doing so. Finding the right way to support and magnify the impact of the investment we're already making in public education is globally transformative in a way that like Stephanie was talking about, most edtech products can't be because they're, they're too narrow is like, I don't know. If you can find a way to leverage a trillion dollars a year in spend to make a, if you can lift the impact of that, you've got a pretty special opportunity on your hands. And not just business, but actual socially impactful as well.

Yeah. Creating a win, win, win for individuals and society at large. Love it. Awesome. Well, Stephanie and Caleb, this has been a super fun conversation. Really appreciate you both joining us today. For those of you that are listeners, please like and subscribe. Throw in the comments, some of your questions and your thoughts on edtech education, AI, where are the opportunities? Love to hear from you. And with that, we'll wrap up and see you at the next one. Thanks everyone.

Thank you.

Thanks.