We speak with legendary Edtech investor Jennifer Carolan, Managing Director of Reach Capital, and cover topics like:
ASU GSV Takeaways
AI Tools continue to Proliferate in Edtech:
The Edtech Websites That Power LLMs Like ChatGPT
Guild Rebrands and Expands
We speak with legendary Edtech investor Jennifer Carolan, Managing Director of Reach Capital, and cover topics like:
ASU GSV Takeaways
AI Tools continue to Proliferate in Edtech:
The Edtech Websites That Power LLMs Like ChatGPT
Guild Rebrands and Expands
This episode of edtech insiders is sponsored by magic edtech. Magic Ed Tech has helped the world's top educational publishers and ad tech companies build learning products and platforms that millions of learners and teachers use every day. Chances are that you're probably using a learning product that they've helped design or build. Companies like Pearson McGraw Hill, imagine learning and the American Museum of Natural History have used their help to design or build some of their learning products. Now magic wants to bring its pedagogical and engineering expertise to make your key learning products accessible, sticky and well adopted. Check them out at Magic Ed tech.com, which is Ma GIC. Edie te ch.com. And when you get in touch tell them Ed Tech Insider sent you. Welcome to Season Two of edtech insiders, where we talk to the most interesting thought leaders, founders, entrepreneurs, educators, and investors driving the future of education technology. I'm your host, Alex Sarlin, an edtech veteran with over 10 years of experience at the top edtech company. Welcome everyone to the weekend ed tech. I'm Ben Cornell, along with my co hosts, Alex Sarlin is an exciting post ASU GSB week, we're going to have the full wrap up for all of you who are there who weren't there just tell you everything going on. We also have an incredible interview with Jennifer Carolyn, who announced a brand new fund from reach capital. So much on the pod. Before we dive in, though, Alex, what's going on at edtech insiders this week. Oh, so we have some great episodes coming up. For one thing we recorded 10 Count them 10 Different on the spot interviews with amazing people at the conference. Those will be coming out probably a standalone separate episodes later this week. We also have a great long form interview with Nancy Sony, the CEO of path match, which is trying to define hire ability for college students. And with Audrey wish the CEO of curious cardinals, which does mentoring for between college students and high school students really cool companies and really cool conversations. Incredible what we're going to start with our first headline, just a recap of ASU GSB. We were there with 7500 of our favorite people in a facility that was probably ready for 3000 people. So it was jam. Both the lobby and the conference sessions. And this year's theme was ai ai everything all over the place. What were some highlights for you, Alex? was definitely the theme. It was really, really interesting seeing so many conversations about that we're really open ended just about this being a very crazy moment in edtech. It being a moment where we think that a year from now we'll come back and things may have changed more than you might expect. I had a great time watching the ASU GSB cup pitches. And congratulations to Kenzie Butera Davis from Morrow for winning the ASU GSB Cup this year. It's a mental health startup is really good presentation. Cash. I mean, it was just so fascinating to be in a space with as you say, like twice as many people as could have been there. I mean, I was overwhelmed. Every time I was in a hallway, I just felt like I was in Grand Central. But every time I got into a conversation with anybody literally sitting next to you at the, you know, at the lunch table, it was fascinating. Everybody's doing amazing, interesting work. So I had a great time. Yeah, we also had just a great experience at our happy hour. So thank you for everyone who came out to our AI social, it was prompt imagine and tech. And we had two other 50 folks there. Thank you magic ed tech for sponsoring for we also had started as a sponsor, Ed Tech MBA. And we also had good harbor partners. We also had good harbor partners sponsor the event, a couple of themes that stood out for me, one was like old school, new school, there was like a degree of like reconnecting with old friends. There were a bevy of events that were invite only dinner, so on so forth. And it was kind of like the old guard. And I think for those folks 2020 2021 was the high watermark. And so there was a little bit of like reminiscing of that and the new reality. And at the same time, there was an infusion of tons of just new entrants. Many of these folks were are either students or they're they worked in ad tech before even education and they're seeing AI applications around learning. And it really felt like this gender rotational split at our happy hour was really fun, because we just had everybody together. But so many of the events, the rest of the week really felt separate and distinct between this, like new school, and the old school veterans. And I just think that imbued a lot of life into the event because of the new optimism, but also, you know, the saner heads of the veterans, you know, we're prevailing, so it just makes for an interesting, I almost like concert going experience. It's like, you've got the opening band, and lots of people want to see the opening band, and then you've got the headlines. And it's Bill Gates, a couple practical themes that really I've been mulling on from from the conference to first is AI intelligence and how it shows up. And I think there's, you know, there's a lot of conversation around AGI. And is there a super intelligence or a singularity moment coming. And I think people keep perceiving AI intelligence evolving in similar ways to human intelligence. And as I started looking at the products that were out there, and just connecting a lot of the dots, I'm seeing a world where AI super intelligence looks much more like a bee colony, or an ant colony, where actually there's not a singular embodiment of the super intelligence, but it's actually the hive structure working together with each bot doing distributed things. And so what you found was like, basically, in November, we were talking about single player AI, with a company that could, you know, ingest your your question or query and spit something out. And now we're looking at multiplayer, and even networked AI, where people are pulling together three or four different AI components, different machine learning part pieces, general AI pieces, and putting them together and really novel concepts. And so I think that number one like bodes really well for the resilience, scalability, and potential impact, but also makes it really hard for end users to understand everything that's going on behind the curtain. I had a couple others, but I don't know, you know, as you talk to a lot of people. How did that theme has that resonate with you, Alex? Yeah. 100%, it does feel like we're at a point where there's now multiple generations of companies that have that have been in tech unicorns or been edtech flavors of the month, or have been edtech success stories or poster children for various things. And you sort of see at a giant conference like this, all of it on top of itself, all the different generations, all the different types of groups on top of each other, as well, as you know, all people from lots of different takes on Ed Tech, a couple of things that stood out to me. So AI was on everybody's minds. There were a decent number of sessions about some other important topics that were not AI like skills based hiring, I saw a lot of there were a few things about VR. And how to Metaverse and just sort of thinking about is what is the next step there. There were ones about simulations, there were ones about mental health, there were ones about equity. It wasn't that it was, you know, all AI all the time, there was some web three stuff, too. But it did feel like it was impossible, just with the timing of Chattahoochee coming out so recently, and then edtech companies just starting to pop up left and right. We've compared Chet up to electricity in the past, we've compared it to calculator compared it to the internet or the study of the iPhone. And I think what struck me about this conference is that you saw everybody talking about AI, but a relatively small number of mostly new entrants, actually really leaning into it, and like baking it into their product and making it all about what they're doing. I stumbled on this little AI hackathon booth in the middle of the big floor. And it was like three or four really cool cutting edge AI companies that were doing clever things with it. And it was so funny because they were like, everybody already knows that AI can generate lesson plans and personalized content and this and that in that we're going for like the next thing and I was like, everybody knows that like, I don't think so I you go into a panel and people are like, can you imagine that AI can maybe make lesson plans for teachers and they were already trying to move to the next step. So you even with this thing that is so new. It reminds me of early days of the Internet and that everybody wants internet pop once email popped, everybody was talking about it, but a few people actually were like really trying to drive it quickly. And realizing how fast this thing is going to move. And I sort of would put my money on those people, for lack of a better word and you really felt it in a room in a giant place like that. You saw every different take. Yeah, and I think this is where like this idea of single player multi player networks, or whether it's like single LLM versus multiple layers or even the tuning and training, we have a distribution of people that are at different phases of the innovation cycle. Yeah. And what we know is also that what used to be the highest end of Maslow's hierarchy, you know, synthesis or something. There's ways in which the AI does those things really well. And so it's like people are the acceleration in product, I think, is one where the market is actually not fully caught up where product is. And so my second theme would be, there's an interesting wait and see going on. So product is accelerating incredibly fast. And I don't believe we've ever seen an in general in tech, the kind of acceleration that we have around product. But for buyers, there, actually, it's made, forcing them to slow down. Because then they're like, Well, what if you solve all my problems today, but next month, someone sells even more, and they're half the cost, right? The paralysis among buyers, especially b2b buyers, where it's like, I'm not going to buy your product. And then there's churn on the b2c, where they pick up a tool, use it for a month, throw it away for the new tool, saw this great meme online, where it's like, all these drowning children, and it's like, AI application, and then the parent is throwing up their child in the air, like playing with my newest AI tool. And it's people are churning so fast. And I think the funders are in the same boat. And, you know, I saw some graphs around like Lindsey, and its growth and then crash, I saw Jasper growth in crash. So I think as interesting as the product acceleration is you actually have a market, there's so much confusion, or the mark is not sure what to make of everything. And on the ad tech side for K 12. I will also say, in particular, there's rising concerns about data privacy, and, you know, is student information protected? If it's an open AI? We've talked before about the idea of an edtech, LLM? I think there were a lot of, I had a lot of great conversations where people were like, this is something that we need. And I'm like, okay, who's going to build that? Okay, so I will also just go to the third topic, the other sub strand going on that wasn't headline news is that there are a lot of companies that are under nine months of runway now. And I think, you know, I had a couple of Heart to Heart conversations with founders that I know and love, they're doing great things, their products are awesome. Some of this buyer slowed down and some of the AI is actually hurting their ability to grow. Partly because there's a saturation of tools out there. Part of it is like today's value proposition pales in comparison to the what one might imagine tomorrow's value prop could be right, I think there's a bunch of companies. So just to give you a couple examples, one company is having a cramdown where one investor is leaving a bridge round and none of the other investors are following and it's at a lower valuation. That is a really tricky thing to go through as a team, and your employees stock options, everything gets reset by the cram down. The second was somebody who's like, basically, I've got like three months, and then I've got to make a call. And I think there's everybody's made layoffs, everybody's kind of cut costs. But my advice to that founder was set some benchmarks. And if you don't hit them, rather than having the company peter out, like make an affirmative decision, like, hey, we made it, we took our best shot. It's not working. We can't grind on this anymore. Let's call it and then I think the third are these distressed company, fire cells. And man, the m&a people at this conference, I know many of them, and they're great people, but man, are they getting a high volume of like, companies that just need to sell for whatever price, and it would shock you how big in terms of users some of these companies are. And so I just think we're in for a big reset, if all the product acceleration is going so fast, and yet companies that are providing core value are running out of business, you've got a little bit of a market mismatch here and either you're gonna have a culling of the herd and a dying off and these new companies are going to fill the space or you're going to have both sides actually have to face the music, which is that the buyers are not ready to pay, and everyone's going to have have to reset expectations. So I came away, don't get me wrong, I think we're really excited and amped about AI in our space learning is a prime example use case. And the novel utilization of AI and user experience is incredible. But I do think underneath a sub headline of that is actually a pretty big reshuffle in the ad tech space. It makes sense. But it's also a little bit. I mean, we've gone through the pandemic bump, and a little bit of this ad tech winter, the funding is still way down. You see these giant rounds still being raised, we saw reached announce a big round, we saw new markets. And that's a big round, we know that I'll raise a billion dollars last year. So there's, there's money in the VCs pockets, there's certainly money in the private equity pockets. And then you have all these idealistic are really exciting small startups, but they just haven't quite found their, their feet. And then and it's not an environment where they can really keep going. And we we saw, we know that the reach round included a sort of bridging function, the opportunity fund to keep some companies in their portfolio get moving. But yeah, I heard a little bit of that, too. I don't know. I mean, we've also seen these weird things happen, where some of the acquisitions that have happened over the last few years, have sort of not necessarily felt like they're super synergistic, some some have, but some feel just a little bit random. And I wonder if there's also a feeling among if there are a number of small companies on sale right now are trying to sell themselves to bigger players, I wonder if some of the bigger players are going to try to be more strategic, even if they have lots of options? They might want to be more strategic in what they're purchasing? Or how they can put the pieces together? How can they put the users from one tool into the flow of another one? Or how do you combine, you know, a financing option with a bootcamp? I mean, you can imagine that there's lots of potential strategy there if people are selling themselves, but I'm hoping that people that they're learning from the sort of growth by acquisition, I won't say failures, but the sort of like over estimations that we've seen over the last couple of years. One other and I think this is related, a trend that I think you and I both saw is that there are a lot of companies being founded by first generation immigrants by first generation college students. There are a lot of female founders. At the conference like you, it really felt that the GSB cup pitch competition, I think it was probably so much more diversified in terms of founder profile than you would have seen in the past, it really felt there at for investors to, there was a lunch with 70 Female investors, there was a Deborah cazzo had a lunch with over 200 female founders in ed tech, like it really felt to me like there was more diversity and sort of more, just more people in the room, more voices, more perspectives than I've seen before, even in tech. Yeah, I totally agree. And yet, at the same time, had some conversations that as the market is tightening up, you're seeing generalist investor shops, or limited partners, really signaling off of somebody's pedigree or like stereotypes. So I think one thing that you mentioned is just how there is some capital in the market, I didn't quite fully get this, and y'all will hear Jennifer's interview here at the end of the pot, is also a really tricky time for the investor community. Because in the past, they were raising a new fund, like every two years, and my attack has been booming and growing up into the right. And now it's half the amount of money being deployed or raised. And, and it's unlikely that you're going to be able to raise every two years. So there's, even though the font sizes are healthy, they're slowing down their deployment rate, to make sure that they're doubling down on those things that are winning. And so in your economic textbook, this would be like the classic moment for a market consolidation. Or this is why they believe that down markets are actually the best time to start the most enduring companies. But that all that's nice economic theory, if you're the winner, for those who are pushing the ball up the hill, the vast majority, it's gotten harder and harder. And I think particularly when we're talking about the diversity of our ed tech ecosystem, I think it's particularly that people are seeing that it's particularly hard for female VCs and female founders outside of the kind of intentional funds that are really focused on bringing more diversity to the field. So I think it's just a signal of the potential and also still a lot of work to be done to make sure that the barrier to entry I mean Look, we we want to transform education, there should be no barrier to entry for any entrepreneur that has great idea. We're for any entrepreneur that or for any investor that has a thesis that could really change the game and return capital. Yes. But that said, some of these Latin American startups were getting a huge traction, at least in the market. There. Yeah. So a lot of left him. So it's a funny push and pull. I mean, yeah. Indian entrepreneurs there as as in years past. And, you know, I think we commented last year that the Chinese edtech community, almost nothing, I think most of the energy around the Asian and tech scene now has shifted to Singapore as a hub, and a lot of exciting stuff going on in Asia. But man, the the Chinese contingent of like three or four years ago, right before the pandemic, that that was the theme. And I'd say last year, the theme was probably tutoring to be honest. Yeah, no tutoring companies hosting like big dinners or anything like that. It's like, hard time for tutoring. So it's just do you think that next year, it will still be about AI? Or will there be some new like news? Or trend du jour? That's a great question. So my personal prediction, and we can revisit this next year and see how to rewind the tape on this one. Yeah, my personal prediction is that it will be technologies fueled by AI. But I think we'll be talking about it in a very different way. I think as of right now, it's literally like, you know, the electricity just got turned on. I mean, we there 10 Different companies, at least, announced their big new AI feature at the conference, everybody in edtech is saying, This is what we're using AI for. And some people are using it for many, many things. So right now, it's sort of like, let's use AI. And it's just AI. And you know, for the most part, it means generative AI, sometimes it means other forms of AI like classification and classic stuff, or reinforcement learning. But I think that what we're gonna see within the next year at this point, is AI use cases like aI fueled use cases that almost you don't even won't even talk about them as AI as the core thing anymore. Like, yeah, virtual completely virtual teaching assistants have with you know that we heard that from a few different people this year. And it's like, once, those are actually pretty sophisticated. You don't have to say, Oh, we have AI, you can say our teaching assistant has this amazing new thing. And your teaching assistant has this amazing new thing. I think you're gonna see a bunch of things like that a bunch of ideas, where AI just fuels whole new categories. Yeah, like the AI Coke, the copilot concept of copilot for teacher copilot for kids copilot for? Yeah, that could be a little category. I mean, I think that AI is enduring like the internet, like electricity, but it is the how, and it's not the what. And so I do see things shifting more towards the compelling use cases that also have defensibility. I also think that there's going to be a lot more conversation about infrastructure and shared infrastructure around AI just so that this degree to which like, there's an explosion of tools, which also makes it really hard in the marketplace. What are the trusted infrastructure plays here? Because one thing that ad tech people always pay for and love is the infrastructure. I mean, think about clever thing about all the LMS is, and this is, I mean, I'm not sure how attractive those are in terms of their return profile. But in terms of scale, and impact, those often become the defining elements of a more mature landscape. So you mentioned a bunch of companies launching their AI products. We've assembled a big list, and we'll include all of them in the show notes. I'm sure we've missed a couple of two. What were some that stood out for you, Alex. So gosh, where do we even start? I mean, one that excited me, mostly because we've been sort of seeing it coming in exciting ways. You know, where Kara, who we've interviewed on the show a number of times, is launching skills benchmarking for large language models for generative AI there as well as starting to bake AI into their actual product. But we saw it across the board we saw brainly announced new AI features that's the Polish Homework Help site that you can imagine how much they can do that's for personalized learning. We saw Chegg announced an AI companion, a Scott school lyrics build an AI co teacher, you see a lot of these co teachers labs, stir announced an automated UX upgrade that basically is trying to big AI into its virtual labs that's Norwegian or Nordic company that does all sorts of interesting things. We saw a company called high link unveil a lesson planning tool like we've been mentioning Honestly, this was these were the ones that came out during the conference. And already, it's Monday conference ended on Wednesday, I'm sure there are many more, because it's just a matter of everybody is looking for how to use it. And the API is make it so usable. It's not, you don't have to spend three months and customer discovery and development to launch something, you probably do have to spend three months in customer discovery to launch something really good, just to be clear. But to launch something that sort of works alongside your existing product, it's pretty easy. And you're just seeing every every established edtech company bake it in, partially probably for the reasons you just said, Ben, which is that if you're Chegg, and you're in many different schools, and you're in many different, you have huge tutoring service, you have a whole bunch of different things going on. But you know that there are 100 AI companies about to start that are trying to pull automatic answering textbook annotations or AI for for tutoring or AI for connecting people together around learning topics, like it's coming from every direction, you don't want to have to tell your existing customers that you don't have anything, you need to feel like you're ahead of the curve need to show that you're ahead of the curve. And I think you're seeing a lot of the companies just, it's almost like a defensive measure against the the inevitable wave of new tools. Yeah, a more cynical view would be like, people are using smoke and mirrors to upend AI on to their thing. And so part of what I got really excited about was any combination of data plus AI, where somebody basically has some sort of unfair advantage in terms of access to data, and then leveraging AI for it. Because ultimately, if you're just integrating chat, GBT, or open AI through an API, there's no defensibility in that, because others get essentially access to that direct, same thing. Yeah, you know, I've always thought quizlets move into the space is really incredible, because they number one are sitting on a treasure trove of assessment data. And assessment drives so much in that tech, but too, they have like a technical team that knows what to do with it, and create the right kind of learning opportunities, closely tied to assessment, which is just where there's a lot of pain and a lot of opportunity. One that also stood out to me was that school addicts one, so it's a tiny startup, but they get all of this data from Google Analytics. And then they make it meaningful for the classroom teacher for the school leader, and so on. And so for several years, they've been super, super close to all the Google Data, and they really understand the Coppa limitations. And yet, they're if you see what their their demo, I mean, it is basically from data to action, which is like the Holy Grail of data driven assembly instruction. So you know, it's that combination that I think was really thrilling. I'd also say the other one that really stood out chyron learning kind of had a coming out party, and I went over to their booth, and it was just packed with people. And this idea that machine learning, which is really hard tech, technical stuff, can leverage generative AI to kind of fill things out or create more content or create more user acts and so on. It's this idea of almost creating synthetic data for you to analyze. I think that that's a fascinating space. And so anyone who's bringing a strong Machine Learning Team AI, like deep AI team like Kyron has, and then basically able to use generative AI to extrapolate data in a way that can train the ML system. That's been a barrier for decades. I mean, ml has been around for a long time. And this is really like a breakthrough moment for that. In their case, they're doing what you mentioned, this idea of virtual tutor virtual as a teacher. So those were one of some of the big ones. I did also notice them flops. We don't throw shade here on our podcast. But there were a number of really large traditional companies that share demos on the floor, or in some of their sessions around how they're thinking about AI. And we look at a company like Duolingo, which is huge, but so nimble and innovative using it. And it really draws a stark contrast between some of the like legacy companies and how hard it is for them to fully integrate AI. And so I do think as much as we talk about defensibility being a challenge. I do think that many of the old guard companies are either ripe for disruption or we're primed for partnership because they're just not moving fast enough compared to the upstarts. But I think we're gonna see a whole bunch of b2b AI companies that sell themselves to other ad tech companies. I mean, we've talked to profit gym, for example, right? Prop gym is all about turning either text or, and they just announced that the conference, text or even voice data into automatic lessons, and they're like, and they're gonna, they're gonna shopping around to all the publishers and saying, You have a whole lot of text data, you have a whole lot of voice data, you have images, like we can bake AI right into it. So I definitely think you're gonna see that. It also makes me think of places like Teach effects, which we've talked about on the part of our while in terms of like creating a new dataset. It's like teachers did this weird thing that I think is Yeah, at first, I just didn't even get it. And now I think it's kind of genius, which is like, start thinking of the actual talk that goes on in a classroom as untapped data. I talked to somebody at the conference, they were like, Oh, yes, it's making surveillance very normal. And it's like, yeah, it kind of is surveillance. But at the same time, if you keep it safe, if you assure and rightfully assure everybody, you're not using it for any strange purposes, or it's not going getting the wrong hands, then suddenly you have this data set that nobody else has, or at least, has been collecting systematically. And you can use it for action, again, goes back to that formula of data, exactly connected with the AI. So I just think that that's where the votes are. And yeah, Jamie was on our pocket, telling us the CSIS. And we were kinda like, we were the numbskulls who were like, Okay, sounds great. I guess reach sees something bigger than this. And now here we are. And it's like they were really right on the cusp of this movement. Yeah, no, it's true. And I think you're gonna see other tools sort of follow similar kinds of ideas, using mood data, taking pulses or exit tickets, basically creating new datasets that happened in the classroom, and then using those in combination with existing AI tools to do things that nobody else can do, because you've just have this new data. It's funny, though, when I hear you say, the bigger companies aren't nimble enough, we're talking about a small amount of time that has happened so far. I mean, the fact that bigger companies are even showing anything that has to do with AI. I mean, imagine if three months after the iPad came out, you had the big publishers all trying to show how they had these iPad, things ready to go. That took a while. And that was already faster than things that came before it. Like the speed of innovation, I think is so fast that you're right, nimble companies like Duolingo and Quizlet. And Khan got a little bit of a head start and have done incredible things. But given the short timeframe, I don't think you can judge any company on what they've done so far, but the whole thing is hardly begun. Yeah, I mean, the ironic part is that many of these large companies are actually fueling the LLN so let's take us to headline three outlets. You did some deep diving into a pretty incredible and I think like overlooked Washington Post article that details the list of websites that make AI like chat GBT workable, so what did you find? Yeah, it really wasn't interesting isn't Washington Post did this big deep dive and basically said, what you know, you hear this large language models are trained on the Internet, what parts of the internet the internet's a huge place. And some of the headlines of this of this deep dive of this expos a were about sites that might be controversial, like ar t.com, or Breitbart or religious sites are various things that you might not expect to be used to train a bot like this or, you know, Reddit or 4chan. At the same time, I was pleasantly surprised at how many interesting edtech companies and sort of ed tech content repositories are really fueling this. It's a funny split, though. So some of the companies that came up really high on the list you have Wikipedia is number two, you can sort of see why that is Wikipedia. You could argue if it's an ed tech or not, I'm going to call it an ed tech for now. You see Scribd as number three that's often used as an edtech tool, sort of as a presentation tool, easy slideshows, Coursera. Number 12. Which is really interesting, because I think this has a lot to do with the fact that Coursera has made almost every lecture into its own separate page for SEO reasons. So it's actually much more scrape bubble. There's many more tokens as they call them out on the web, not behind any kind of paywall or login wall than other tools. And I think that's part of why this is happening. Instructables. Number 15 Wiki how 41 EdWeek the publication it was in the top 100 It blew my mind and we got number 91 Wow, yeah, ahead of YouTube. 104. Yeah, ahead of YouTube. This I couldn't get my head around, actually, YouTube 104 on the list. YouTube is as far as I know, the biggest repository of content on The Internet of any kind, but maybe it's content that is not that trustworthy or hard to scrape or comes through transcripts that are not translated well, I don't know. But under 100, for YouTube, I would have thought that would be up there with Wikipedia number, you know, two and three, but not even the EdWeek blogs site is in the top 200. Because EdWeek has many years of all sorts of blogs. And I don't know, I guess they're being looked at as the sort of representative education publication. If you were going to ask Chad GBT or other MLMs, you know anything about education, they want to look through the history of education through through that, then there are a bunch of companies that were lower than I would expect, but not that low. So you see, like Quizlet is about number 1500. On the list. Quizlet is a ridiculously large repository of data and content. As you just said, Ben, it's one of the, as far as I know, I think it's still in the top 50 sites, top 50 most visited sites on the Internet, and his billions and billions of things. So that was a little low for what I thought it was. But again, the site probably some logic behind it. Companies like Udemy, and Skillshare Udemy is number 5000. Skillshare has about 15,000 Course Hero about 15,000. Pearson about 16,000. And I mean, I don't know you think about the amount of meaningful educational correct content on Udemy Skillshare. Course Hero and Pearson and Quizlet. They're not being used as much as you might think Yun Academy 26,000? I don't know before we go into the really loved ones, what are your thoughts on something? Yeah, I mean, look, you're training, an LLM to make predictive inference. And you've got RedHat in the top 10. And then you're going to take Pearson, and it appears that his basically at 16,000. So I think this raises questions about the traceability verifiability of the data. And I think everyone can tell you like the bigger the data set, the higher the accuracy, but it is interesting, the disproportionate nature to which kind of user generated content versus expert curated content plays a role in the LLM. And it's just swings very heavily to user generated. I would also say there's some surprising ones that you called out that were even more like off the charts low in terms of their hierarchy. What were some of those that that spike for you? Just a quick correction. Reddit is not in the top 10. Reddit is actually number 540 on the list, okay, 540. But I hear you it's still orders of magnitude higher than Quizlet or Pearson or things that are established expert content. So I'm totally with you. A lot of it has to do with the number of tokens, they say the number of basically like individual, I don't know, data points that they use, but yeah, so some of the ones that were really low, so you thought Pearson is low at 16,000, Chegg. 75,000, Kahoot, 130,000, find us 330,000 330,000 in post access site for these larger language models. Duolingo over 500,000 con almost 600,000 at 570. And edX, which this is this is right, this really shocked me. I mean, you have Coursera up there at number 12, number 12. And then edX is at 646,000. And I have no idea why that is my best yet. I'm super open edX is open. That's the whole point. I think it's open from a free perspective, I'm not sure it's open from a data scraping perspective. I mean, there was a great article by DualShock, from class central a couple years ago about how all these different how a lot of the different sites did different things for SEO. And they talk about how Coursera had this very specific strategy of basically making like hundreds of 1000s of individual pages for every class for every lecture, putting the text on them. So that like, no matter where you are and the Internet, you're going to whatever you search, you're going to end up finding something on Coursera that brings you in that was their strategy. edX did something totally different. And a lot of companies had things that are totally different. So I think it might have to do with that because the number of tokens that they are saying they can get out of edX, according to this article is 36,000. So Alex, is that how the tokens are counted? That might be an issue like if the entire edX library is on and the entire Coursera library is on open AIs? LLN that, okay, maybe we're just counting things in a different way. I'm not an expert, but I don't think it is the entire library. I mean, you look at Coursera Coursera has 53 million tokens, and edX has 36,000 tokens. Wow, that's such a huge difference. And how do you feel about it? If your data is fueling open AI? Is that a good thing? Is that a bad thing? It's probably a bad thing from competitive standpoint. Now, your training ability on top of your training and tuning has already been essentially baked into the model in the first place. I know the folks that I work with that common sense, saw that they were in the top 10,000. And they're like, basically, anytime you want to do a movie, or TV show review, you can type it into jack to BT. And it just does a synopsis of the common sense rating. You don't have to go to the rating. Yeah, I mean, I talked to the Coursera, Product Manager for quite a while who's just been launching their AI content. I mean, I don't know if they're thinking about it quite in this way. But it is true that if Coursera would have had this proprietary advantage of having this incredibly amazing data set that nobody else had access to, you could argue that that would be a big competitive advantage that said, both edX and Coursera, specifically, and places like Khan, really, I think of them almost as like public goods. I mean, edX and Coursera are publishing University content. It's fueled by taxpayers, in many ways. Like, I don't know, I actually, to me, maybe this is just naive. But part of the shame of seeing edX at that low is edX is Harvard. It's Dartmouth, it's Berkeley, it's Georgetown. It's some of the best universities in the world, putting really well structured data out into the world great teaching tools, and then somehow just the way it's being indexed. This AI isn't obviously able to use it in any meaningful way. Like what a waste makes me think further, like we should build an ad tech insiders LLM that is really affordable learning. Exactly, seriously. We need an LLM that really prioritizes factual, accuracy, intellectual veracity. And also like pulls from all of this incredible publicly available stuff that felt really low on the opening Eilis. I guess the other thing that I think we're gonna have to watch here is how does IP rights litigation go going forward? Yeah. And like, will there be a licensing fee kind of all Spotify, where people, once you find out this blackbox gets opened up and says, here's all the websites we index on? Is there some sort of payment that's going to have happen, I heard that Reddit is now you know, basically saying, We'll license our Reddit content for like $100 million a year to the LLM that wants to pay it. And they're gonna end up having triple the amount of revenue from licensing out their data than they do from any advertising and the platform. Yeah, and just to put a fine point on it, some of the sites that are being used very high up in the ratings to get really good data, our journals, PLOS or frontiers, or Springer, are all in basically the top 20. You have Elsevier about 700. But like, I think that, again, is amazing for the world. The NIH is number 22. The NIH journals, this is information that is sitting there in these impossible to find often pay walled ridiculous places that where only doctors or geneticists or whoever specialty can even begin to see them, let alone use them. I love the idea of all these academic journals being part of this, and you can ask an AI something and it'll tell you about a study that happened in 1993 that nobody ever, even the very few people even cited, but it has exactly your answer. Like that is amazing. But it hits exactly the issue you're talking about. I mean, these journals have been paywalled. Nature is at number 48. These journals have been paywalled for a reason for a long time, are they just going to shrug and say, hey, it's great for the world? So take our information, I don't know. Right? Well, then how did open AI scrape it to if they're paywall, and so this is a story that is going to be written here in the next six months around, I would also just say, for our listeners, read the nutrition label, you are now getting a sense of what's in the LLF. And so whatever is going in LM, that's what's coming out. And you can basically see me opening eyes huge dataset. So the idea is that the bigger it is, the less risky it is to be swayed by one potential data source or another. But I do think that it is a great chance to basically say if I'm going to build on top of one of these platforms and Quebec, let's look at what's under the hood, what what the nutrition label is, as I said, I mean, it would be so easy. I'm being glib, a little bit here, but it'd be so easy to look at these 200 Top Sites and say, Nope, no Nope, nope. Yep, yep, yep. To the ones that are like peer reviewed or like really deeply or not even journalistic. Right. I mean, you see things like, a lot of the top ones are our news. But often their news that is considered bias you see, like the sun, the UK is tabloid up here in the top 100. You see, you see some really wacky things along with AP, and PBS and things where you're like, Yep, I'll take that. I mean, obviously, everybody has their own ideas about media bias. But yeah, so I think probably a good AI person would say, ultimately, what LM is doing is just predicting the next word or phrase, it's not necessarily constructing full ideas. But we wanted to kind of do this deep dive and share it just because I think it does raise more questions than answers. And I think that this will be an interesting story to follow up with last thing real fast. So like Fox News number 34, CNN, number 162. Now, is that a political choice? Probably not. But to your point know what's under the hood? Because who knows? Who knows? It's an interesting moment. Yeah, well, and when Elon makes his own post truth. What his status. You had mentioned before that there were a number of non AI announcements at ASU GSB. And before we jumped to the interview with Jennifer, Carolyn, we did want to cover one more, which was guild announcing that they're dropping the education from their brands. So it is guild period. And they're essentially repositioning themselves as a vertical talent solution that basically helps employees grow their careers with companies. I thought that the subtlety and the nuance in the messaging really spoke to how hard it is to build a double sided marketplace for community colleges and employers, in part because part of the the goal of the employee is to get a better job independent of whether it's with that company or not. Right. So if you're upskilling people that often doesn't align with their employment, for example, at Walmart, like how many store clerks versus store managers do they have it, the pyramid is very, very wide at the base and very narrow at the top. It also just spoke to this idea of being more consultative and services based, which I think guild is what I've heard, Gil, you know, as you get larger and larger clients, you really have to create customized products that really support them in their industry. I see it as a positive move, I think they're going to be going deep in different verticals. And, you know, I also heard some great stuff from Cengage, which also recapitalize and brought their debt down, but they're doubling on upskilling and career pathways. I think a lot of people are seeing kind of lifelong learning vertical pathways as a space that goes beyond just an educational offering, but really is like a consultative spine that helps either the employee navigate their career across organizations, that would probably be more the Cengage model, or the kind of corporate talent management hype, which would be more like like guilds is doing it just feel like while everybody else is chasing after the AI, I think these folks are really seeing that hiring, retaining and managing turnover as a component of the upskilling value prop is really where the most defensible value is. What was your take, as you read all this? Yeah, so I agree with everything you're saying there, I have one additional point that I'd make here, which I think is relevant. There's a company called Gleann. That is sort of, it's been funded by amazing people, including slack, and it's basically trying to be slack for your intranet. He can go through everything inside the company's information, architecture, no matter where it is. And you can just ask it, conversationally, I need to know how to expense this, or I need to know, whatever, and it'll pull it out and serve it to you. And the reason why I'm bringing that up in this context is, I really agree that there was a very clear recognition at ASU GSB that people want hardcore skills, sometimes I mean, soft skills, but relevant skills to careers is just like accepted. I think in other years, it's people have had to convince others of that, but it was accepted. And when you look at what guild is doing with their career accelerator, so think about from a 10,000 foot view, guilds original model was providing education as a benefit to frontline workers. And what they meant by education as a benefit was college degrees, or at least certificate Coming from colleges, then Amazon comes along, or instride comes along and they, they do something kind of similar. But Amazon says, You know what we'll acknowledge credentials from from lots of different colleges. But we're also going to acknowledge credentials from alternative providers. And that's kind of interesting because alternative providers could be very stackable are very short or don't have to go through the sort of any kind of the formal learning concept that colleges do. And then Google this thing. I mean, listen to some of the stuff in this press announcement. It says, career accelerator includes four modules designed by career experts, meaning internal guilds, career experts, you know, people they hire, to address the most requested career readiness skills, navigating a job search, developing resumes, and cover letters networking and interviewing. Now, these are not things that colleges necessarily teach, you could find them if you look real hard, but that's what they're they're learners are looking for. And then the other types of content and accelerator on demand bite sized, interactive, mostly made pies, the companies, so three years ago, if you're a Walmart worker, and you're with guild, it means you're getting a degree from Brandman. University while you're working at Walmart, as a frontline worker, what it's going to mean in six months is Walmart has a huge library of skill based short micro content that if you're a frontline worker, you're taking the Walmart degree or you're taking the Walmart training, on demand in ways that can be elevated internally. And they say this person's learned everything about accounting, while they're while they're doing this, maybe they should be an accountant, or you can even do it towards an internal promotion. I mean, internal career mobility, they call it long rant, but I see it as actually a further push in the direction of universities just being sort of continually left behind, as people just get so freaked out by the rapidity of the changes into society. Everybody wants skills, they want them as short and fast as possible. And even guilds original model can't keep up with that, because they can't get people through the programs. And it's just too big, it's too formal of a learning opportunity. Instead, there's like, why don't we give people you know, 35 minute videos, and that qualifies them to get a promotion, and not that many people are ever going to do that. So give them that. And it's just a really, it's a very big change in sort of whose information matters. Let me put it that way. Right. And also this idea of, is it a compliance based thing that you just need to go through the hoops? Or are we actually training real skills that are applicable the next day? I just everything you said, when we first started talking, I don't know a couple of years ago who was like, this is where it's headed. Is it is it not? Yeah, it's like a fait accompli now, it's like, up to the fact now. And it's really a matter of time before small and midsize companies are jumping on that same train to and please like Springboard or have entire corporate division now selling the company. So well, all of that happening leads us to our incredible interview with Jennifer Carolyn coming up here. So we will come right back from our break. And hear from Jennifer. Hi, everyone. We are so excited today for our very special guests. Jennifer, Carolyn, of reach capital. This is an honor to have you on our show. You have some big news that you just announced this past week in terms of a new fund. We're so excited to have you on the show. Thanks so much for joining. Thank you, Ben and Alex, it's really a pleasure to be on. Before we talk about the new fund. I'd love to just rewind a little bit and just take us through the journey of reach capital when you started it. How did that come into play? And then also, did you ever think that you'd be here with this exciting $250 million plus fund? Sure. So we got started back in our story starts at NewSchools Venture Fund. I'll give you the kind of quick version of it. We choose a partner at reach. And I started there in 2005 2006. We started working together there we began investing in education technology companies. She ran the new school Summit. One of my first investments actually was better lesson Mastery Connect. We also invested other elements. And we left to create the new school Seed Fund in 2011. Our first hire was Shawn Tao, and shortly thereafter Esteban and we've been together since and we're still the four original founders, managing directors of reach. And so we started new school seed. It was a $10 million early stage ad tech fund that was still under the NewSchools umbrella. It was primarily funded by foundations and high net worth individuals. The success of that fund we were very lucky in terms of timing, and when we were investing and the sorts of things that we're investing in the app store had just I'm online, iPads had just come out AngelList had just launched, and everything was kind of coming together at just the right time. Out of that fun. We invested in companies like Ed grade gold Buck Class Dojo, Nearpod. And it was turned out to be a great fund. On the success of that fund, we then created a fully separate fund called REACH capital. And that was incorporated in 2015. That was our first fund, it was 50 million. And now we've subsequently gone on to raise four funds. And so we just announced our fourth fund 215 million, alongside our first ever Founders Fund, which I'm super excited about. Yeah, so let's unpack some of the different aspects of the new fund, you have a few different focus areas and you're thinking about AI, you're thinking about founders opportunity, can you break down each of the pieces of the fund? Sure. So in terms of scope, or or just the different funds structures, let's do the different funds structures. Okay, so we have the 215 million core fund, and that will be early stage focused, continue to do primarily seed series A and some follow on Series B, and C, that will continue to look very similar to our three previous funds. We then have a sidecar Fund, which was raised entirely by reach founders. And this is something we've wanted to do for many years, which was really bring in the reach founders that made this all possible. Many of them are sort of our OG founders from new school seed and reach one that are paying it forward and investing in the next generation of tech founders. So they're, we're kind of knitting them more closely to this ecosystem. So they are sitting on boards and investing in the next generation, and supporting them and mentoring them. And it's, I think, a real sign that the ad tech ecosystem is at, you know, has evolved and matured in a way that now we have enough founders that have exited and have are in this place where they can now join this fund is LPs and give back to the the next generation. And then we also have the reach Opportunity Fund, which we don't talk about too much. But as is really a growth fund. It's a pro rata fund, where we can continue to double down on our existing investments, and do our due our pro rata investments. And we also have this new AI catalyst, which is really a carve out of reach for and is a way for us to move quickly and do small amounts of capital to very early stage AI education projects. As you may know, we have this this thing called labs to outreach where companies that are under 250,000 investment needs, we can move quickly and invest in those companies without going through the traditional investment process. So just for our listeners, this basically means that reach you can go at the very earliest stage, you're just starting, you just have an idea and you want to pursue it all the way through Opportunity Fund means doubling down on those that are winning in your portfolio. And reach for really represents such a huge scope that you're able to fund people all the way through their lifecycle. What was it like raising reach for versus reach one? And we're hearing from a lot of folks that institutional investors are rebalancing their portfolio away from venture? How hard was it to raise reach for given that you've had a great track record? I'm sure that spoke for itself. But what was that process like? And how does it feel now coming into a market where it looks much different than 2020? Or 2021? Yeah, that's a great question, Ben. So on this first point about us being able to invest across the lifecycle of a venture, this is a really important point, because even though our fund has increased in size, we wanted to make sure that we had the structure and the systems in place internally, to do what we have always done has been part of the Reach DNA, which is invest in the earliest ideas and ventures in education. So we have always incubated companies that reach many times we have founders that are working out of the reach offices. And we're doing EIR type of structures. We have two right now, in fact that were that were incubating. And so we want to continue to have that muscle if you will, to invest in the earliest ideas and support those those founders, but also to stick with our companies as they grow. So it's not unusual for us to invest many times in some of our companies. We invested seven times in Nearpod four times in springboard, and that's not uncommon for us and then on and how we have the funding environment. So we actually closed this fund, I think it was November of last year. But we just announced it recently. And the first fund was very challenging to raise, I think it took us a year. And Shawn and I were flying all over the world trying to raise little little bits of capital and trying to convince people that this sort of unlikely group of venture capitalists could invest their money wisely. And it was hard because we had the track record of newschool seed, which was helpful. It was very nascent sector, a lot of people were skeptical about education. And it took a lot of convincing this time was was much easier, it's never easy to raise capital, I don't want to come out and say that it was easy, because it's not, it's a lot of hard work, there's still a lot of convincing people. But we did have the fund track records, which were very strong. And we were able to raise off of that. You mentioned this unlikely group of venture capitalists. And you know, one thing that stands out about reaches, it's a really diverse group, and really known for that you have multiple females, imagining partners, three of the four people of color, I'd love to hear I think you were doing that before it was cool. People were paying a lot of attention to that kind of world. Can you tell us about what that's been like? And how that sort of evolved as the world has evolved around you? Yeah. It's interesting. When we raised reach one, we used to get so many strange comments, like when we pitch to investment committees, they'd say, Oh, you look like the United Nations and, and things like that. So we to be honest, we have come together in a, in a way that was very natural, we have all been impacted significantly by education, it's transformed our lives. And so education often transforms the kind of non majority demographic. So I think it's not unusual that that we are a diverse fund, because we choose people based on kind of who they are, their background, what education means to them. And we have all been our lives were changed by education. That's why we want to work in this field. So I think that's how we came together. And, you know, I grew up in Chicago, I went to my sophomore year, I went to Whitney Young high school there. And it was a diverse school. And I think many of us are accustomed to growing up in diverse environments to I will say, I'm hearing from folks raising their first fund, or maybe it's a second fund after they had a raise or very small initial fund, that they're still feeling like the dynamics favor white men, but also people who have kind of these textbook pedigrees. And so it still feels like the investor space, especially if you're just starting your investing career, or you're trying to do something solo is really challenging. Do you have advice or recommendations for folks that are trying to navigate that journey? Yeah, I think that limited partners are going to often seek to look for evidence that this team can execute. And if there's any way that you can show your track record, maybe if even if you've done any angel investment, which I know is privileged and hard to do if you don't have that capital. But if there's any way that you can come show a track record or show that you have investing track record, one of the things that we did, I think which was helpful, is we warehoused our first three investments. So we made three investments that Silicon Valley Bank actually backed us before we had actually done our first close on the fund. And I think that was helpful, because then they can see oh, this is the types of things that they are going to be investing in. So it was risky, and we were nervous, I had many sleepless nights, not sure if we were going to actually be able to close. But in retrospect, I think that was helpful, has great advice. Let's shift to the entrepreneurial side reaches known for your partnerships with Stanford, you're doing these events with Union Square Ventures and Andreessen and really being at the forefront of AI and innovation. You also have a very strong reputation as the best edtech Langham investor and really knowing we're just seeing so much creative energy coming out of LA Tam, and the education environment there is innovating in ways that it's challenging in the US, what's getting you most excited in front for and what are some of those focus areas that are entrepreneurs out there should be talking to you about? Yeah, that led an opportunity is super interesting, and one that we've been looking at for a number of years and you know what half of our team are Spanish speakers, and we have Esteban softneck, who ran a seed fund you In Buenos Aires, previous to coming to reach and so we have that those connections and network. And we have some LPs, also from from Brazil. So we have been investing in that space, we have a company called coder house is a core investment and a number of other investments. We had a bunch of events that we held two weeks ago in Brazil. And it was so exciting to see the energy there. It was like the early days, when we were just getting started on C fund way. And I and Michael Stanton and some of the early investors in the space, Michael Horn, we would go around to different cities and host of these ad tech dinners. And these pioneers would come out and talk about their ideas. And it was really early days and felt very scrappy, and nobody really knew about us or believed in a space. And it feels a little bit like that in some of these new emerging markets. And so in these events, they were it was just fun to see the energy there. But yeah, we for some of the things that we're going to be investing in, we'll continue to we invest across early childhood, k 12, higher ed, and lifelong learning are the future of work. And we're still going to continue to invest along those different factors. And we've we have different people on the team that focus on those different areas, we have 10 investors now at reach. So we have our core investors. Now we have the venture partners as well, that act as sort of operating partners that reach we still think it's early in the evolution of digital print to digital. And so a lot of our companies that achieved scale and had great success in our earlier funds, were really leading that evolution from print to digital. So we'll continue to invest in that. And primarily in the US, US was the first country to really go one to one. And now that will be true soon happen around the world. As learning this technology devices will be used to deliver learning in formal education and informal settings. So we're going to continue that. We're also super excited about AI. And that's another area that we have invested in over the years we had right lab and grave scope and mainstay. And those founders are serving as the advisors for the AI Catalyst Fund. And they have always continued help us that companies they mentor founders, they serve on boards of some of our companies now like Matt, Matthew Ramirez is on the board of math picks one of our AI companies. And so we're just seeing a lot of activity in that space. And I particularly am interested. And you'll appreciate this Ben as a former teacher, the ability to automate and optimize the work of the teacher, so that they can really focus on the instruction and the quality of instruction and that and the connections with the students. So I'm looking at a bunch of sort of teacher preparation, lesson planning, AI tools, we're also looking at some ports for teachers around helping kids learn to read, just seeing a whole lot of different things. I mean, we really look to the founders to show us what is possible. And so we have these sort of general areas that we're interested in. And founders are really innovating and showing us kind of the breakthrough ideas that we meet them back. Yeah, I wanted to ask about exactly that these breakthrough ideas were in this amazing inflection point moment with AI. And this AI Catalyst Fund seems like it's it's designed specifically to sort of catch the rising wave of people are going to be using this in so many different ways for teachers support for tutoring for reading for you know, anything, I'd love to hear you talk about how companies that are trying to use AI right now can use it in a way that's defensible that goes beyond what three other companies down the road might be thinking at the same time. We talk about this a lot on the podcast, how do you do something with AI? That is a unique value proposition that somebody three people in a garage can't replicate down the block with open AI API's? Yeah, that's a great question. Alex, I really liked your you guys conversation on it was it last week or a couple of weeks ago, you talked about the different ways that these niches sort of smaller language models can be used to be a differentiator in the space. And so I have seen some companies that are really interesting that have partnerships with proprietary datasets, for example, are partnering with the research journals and creating these training on the the education research journals and therefore can surface findings or insights from education research, which I think this is so exciting because that information and knowledge should not be lost. tucked up in these journals that you have to pay for, but should be made accessible to all teachers. I mean, I remember when I was teaching, I would look for what are the best ways to group students. And then I'd get stopped by the paywall, that JSTOR, whatever. So I think that that's where we make progress is when we liberate knowledge and information. And so that's always been an overarching thesis at reach is that opportunity will become more accessible, if we can liberate the knowledge tools, mentorship, resources, that help you achieve social and economic mobility. And that has been a story of, of our history, right back to Gutenberg Press, you know, when the scribes were not the holders of knowledge, but it became liberated, and then literacy improved, and, and so on. So I think that there's opportunities like around just these unique partnerships where you can train on certain datasets, certain models that you can get access to. And that's where education entrepreneurs have an advantage. They know where the that knowledge is that content is, and can develop those unique partnerships. That's one area that I'm really I'm really interested in. And I think that there's unique advantages for education entrepreneurs. Another is education, entrepreneurs can understand and know their customers in ways that the larger companies cannot. And I think you see this happen, like with Google to Google education, totally, yes, it's an enormous platform that a lot of teachers use, but they're not able to do they don't know they're their customers in the way that some of the edtech companies can. So a competitive advantage is just understanding the space and understanding that customers deeply. Jennifer, I could spend hours just picking your brain on all of this. And if people want to find you and reach, what's the best way for them to reach out? Is it through an individual? Is it through your website, because I know with this fund for it is going to have explosive impact in our space, right at this inflection point for industry. Thank you, what's the best way for them to find you just my email Jennifer at reach capital.com. All of us at reach have that same convention and they can reach out to any of the investors or anyone on our team. And then there's a separate submission form for the AI catalyst. And it's on the website. And I just got done looking over the submissions from yesterday. And so there's that's a place that we are monitoring very closely and looking at those every day to wonderful. So we'll include those links in the show notes to the AI catalyst. Fun, Jennifer Carolyn, reach capital, so inspiring to hear you talk about past, present and future of reach. Thank you so much for joining us today. Thank you, Ben and Alex, pleasure to talk to you. That's been great. That's it for us on this special supersized edition of Ed Tech insiders. Thanks to Jennifer Carolyn from breach capital, one of the most informed, interesting, thoughtful, experienced investors in all of edtech before we go, one other article we didn't get to talk about was a really interesting one from l&d leader, Josh Burson, he was presenting at ASU GSB. And he came back and said, published something called ed tech is going crazy for AI, basically saying that he hasn't seen anything like this, this sort of love affair with a new technology and a long time, and also basically saying the l&d market is about to get disrupted like never before. That's a direct quote. So we'll put that link in the show notes as well. He's a smart guy and has a really interesting take. Thanks for being here with us on this extra long episode. If you've made it all the way here and remember if it happens in ed tech, you'll hear about it here on at Tech insiders. Thanks for listening to this episode of edtech insiders. If you like the podcast, remember to rate it and share it with others in the Ed Tech community. For those who want even more Ed Tech Insider subscribe to the free ed tech insiders newsletter on substack. This episode of Ed Tech insiders is sponsored by magic ed tech. Magic Ed Tech has helped the world's top educational publishers and ad tech companies build learning products and platforms that millions of learners and teachers use every day. Chances are that you're probably using a learning product that they've helped design or build. Companies like Pearson McGraw Hill, imagine learning and the American Museum of Natural History have used their help to design or build some of their learning products. Now magic wants to bring its pedagogical and engineering expertise to make your key learning products accessible, sticky and well adopted. Check them out at Magic Ed tech.com, which is Is ma GIC Edie Te ch.com and when you get in touch tell them Ed Tech Insider sent you