Edtech Insiders

Week in Edtech 4/24/2024: Meta's New VR OS Strategy, Anthology Raises $250M, Insights on the Future of Educational Funding and More! Feat., Rebecca Kockler of Magpie Literacy and Kristen Huff & Amelia Kelly of Curriculum Associates & SoapBox Labs

April 30, 2024 Alex Sarlin and Ben Kornell Season 8
Week in Edtech 4/24/2024: Meta's New VR OS Strategy, Anthology Raises $250M, Insights on the Future of Educational Funding and More! Feat., Rebecca Kockler of Magpie Literacy and Kristen Huff & Amelia Kelly of Curriculum Associates & SoapBox Labs
Edtech Insiders
More Info
Edtech Insiders
Week in Edtech 4/24/2024: Meta's New VR OS Strategy, Anthology Raises $250M, Insights on the Future of Educational Funding and More! Feat., Rebecca Kockler of Magpie Literacy and Kristen Huff & Amelia Kelly of Curriculum Associates & SoapBox Labs
Apr 30, 2024 Season 8
Alex Sarlin and Ben Kornell

Send us a Text Message.

Join Alex Sarlin and Ben Kornell as they explore pivotal topics in this Week in Edtech episode:

👓 Meta's new VR OS strategy—will it create a new tech rivalry?
🤖 The growing role of AI in Meta's apps: innovation or irritation?
🏆 Age of Learning earns a spot in TIME's Top EdTech Companies of 2024
💼 Anthology raises $250M for strategic growth in EdTech
🌱 Insights from the Co-Founder at JuneX Capital Partners on the future of educational funding

Plus special guests, Rebecca Kockler of Magpie Literacy and Kristen Huff & Amelia Kelly of  Curriculum Associates and SoapBox Labs 

Don’t forget to subscribe to Edtech Insiders for more updates and insights from the forefront of educational technology!

Show Notes Transcript

Send us a Text Message.

Join Alex Sarlin and Ben Kornell as they explore pivotal topics in this Week in Edtech episode:

👓 Meta's new VR OS strategy—will it create a new tech rivalry?
🤖 The growing role of AI in Meta's apps: innovation or irritation?
🏆 Age of Learning earns a spot in TIME's Top EdTech Companies of 2024
💼 Anthology raises $250M for strategic growth in EdTech
🌱 Insights from the Co-Founder at JuneX Capital Partners on the future of educational funding

Plus special guests, Rebecca Kockler of Magpie Literacy and Kristen Huff & Amelia Kelly of  Curriculum Associates and SoapBox Labs 

Don’t forget to subscribe to Edtech Insiders for more updates and insights from the forefront of educational technology!

Alexander Sarlin:

Welcome to Season Eight of Edtech Insiders where we speak to educators, founders, investors, thought leaders and the industry experts who are shaping the global education technology industry. Every week, we bring you the week in edtech. important updates from the Edtech field, including news about core technologies and issues we know will influence the sector like artificial intelligence, extended reality, education, politics, and more. We also conduct an in depth interview with a wide variety of Edtech thought leaders, and bring you insights and conversations from ed tech conferences all around the world. Remember to subscribe, follow and tell your ed tech friends about the podcast and to check out the Edtech Insiders substack newsletter. Thanks for being part of the Edtech Insiders community enjoy the show.

Ben Kornell:

Hello, Edtech Insiders listeners. It was so great to see so many of you last week in San Diego at the ASU GSB Summit. Many of you turned out to our once in a lifetime rooftop party in San Diego on that Monday. But whether it was lunches or panels or passing by in the hall, it was just so great to see the Edtech insider community out in full force. We have so much great stuff to follow up on from that conference. We're gonna dive into it today. But before we do that, Alex, what's coming up on the pot.

Alexander Sarlin:

So we have a really interesting episode of the podcast coming up with Jonathan Lee buff who's the CEO of antimatter, which is an incredibly innovative AI tool that basically uses things like memes, text messages, and Socratic dialogue, the chat GPT you know, where the GPT asks the student questions, to really flip the script on what AI education could look like, and use his very popular sort of pop culture methods. It's a really interesting interview. So definitely check that one out. And

Ben Kornell:

as for events, we've got a happy hour coming up on May 9, it's going to be at the Salesforce Park in San Francisco, which is awesome urban park, on the fourth floor, and excited to have all of you come out and enjoy the San Francisco sun. And then we also have a baseball game that we're co sponsoring with workshop and a couple other sponsors, where it's at Tech Night at the ballpark. And it's not for the Oakland A's, the major league baseball team, it's for the Oakland B's, which are the independent league team that's actually taking over the Oakland baseball scene as the Oakland A's move out of town and eventually down to Las Vegas. So it'll be really fun. We're gonna get our name on the scoreboard, Alex. So we're finally hitting the big time for AI tech insiders. It's great. Well, speaking of big time, ASU GSB. Man, it was packed this year, let's just do a brief debrief of is your DSP. What were some of your major themes and takeaways? Yeah,

Alexander Sarlin:

I mean, what a fascinating experience. Everybody I talked to there and since has said, wow, it was incredible and exciting and also overwhelming. You just have so many conversations with so many interesting people about so many big, big ideas in education in edtech. And then you sort of zip off, like 10 minute conversation, 20 minute conversation, then you zip off, he goes here and there. And here and there. And at the end of the day, you sort of look back and you're like, I think I talked to 40 people about 100 ideas. It's such an intense experience, but incredible. And you know, I came out this funny combination of super energized about the space with so many big ideas about what could happen. And also just like I've been sleeping a lot more than normal over the last week. And I've heard that that is pretty, pretty normal for others who are there. It's just such an intense experience. Some of the things that stood out to me, and I, you know, I want to get to yours, because yours was so interesting. But some of the things that stood out to me, I talked to a lot of educators and educator groups at the conference and is really interesting, you know, just we always do sponsors 2000 educators to come to the ASU GSB Conference, which is very expensive otherwise. And this was the inaugural Airshow, which was a free open event, where they had lots and lots of educators coming in to learn about AI offerings. So there were I felt like the educator voice was more visible than usual. And one thing that really struck me is, you know, educators are so excited about AI, especially the type of educator who's, you know, at ASU GSB or at the air show, but they aren't yet convinced that the sort of suite of tools that Ed Tech has put out as an industry are superior to just sort of them sitting with an open model is a you know, like a teacher sitting there with Gemini or Chet UBT can do some pretty amazing things. And you know, that was not something I've sort of heard people say in the past. You know, there's been a lot of polls about AI teachers embracing it, are they afraid of it? Is it about cheating? Have they even used it? And I felt like among some of the cutting edge educators. They're like not only we are embracing it, but we don't yet see ed tech, you know, making a difference here yet. That was kind of a little bit of a jaw dropper to me. So that was really interesting. Lots of other takeaways. But, Ben, I want to hear yours first, what are some of the things that jumped out to you? I think you were looking at it through a very investment heavy sort of business lens. And you had some really interesting findings. Yeah,

Ben Kornell:

the highest level, I mean, the reality is, by the numbers at Tech investing is a 10th, of what it was at peak. And as much as I think, last year's analysis was, hey, we're actually back on track to where a tech investing was pre pandemic 2024, has actually been even lower. And so it's quite concerning. And then on top of that, you have many headwinds in the economic environment. Schools and universities are facing big crunches in terms of a declining enrollment, the stimulus funding all going away from the pandemic, and increasing staffing costs. And when you ask, honestly, the leaders and administrators of these institutions, where are you going to spend money? Are you going to retain that employee versus retained that ad tech tool or product, they're going to retain the employee? Yep. And even in the workforce, which, you know, workforce, pre pandemic was kind of the Darling space, there's just a big sense of corporations after doing layoffs and so on, that they're going to be tightening their belt around any products or tools or services that they're going to be using. And so l&d budgets have been particularly impacted. And so that backdrop by the numbers was a stark contrast to the vibe and energy totally, which was, hey, we're transforming the world and AI is going to be this new opportunity to change teaching and learning. And so I think what we may have is actually, you know, the pithy frame, I would say, is, we actually have hit a moment where the technology is meeting the needs of the teaching and learning at K 12, higher ed, and corporate levels. And that is incredibly exciting. And we are also hitting a trough in terms of the appetite of these industries to pay for adapt and change. And so it's going to be a particularly tricky period of time to figure out which tools will scale which ones won't. Yeah. And then the other kind of big takeaway is that there weren't any bombshell announcements normally is huge. ESP is the time where it's like, this big company is getting acquired and merging or releasing this big new product. And you and I are kind of getting a lot of backchannel information, there are people ready to do big announcements, but everyone's being sensitive about maybe let's hold off a little bit. And let's wait until we're 100% Sure this is really going to happen. And so, you know, I just think we're in a more cautious, more plodding era of edtech, despite the fact that the kind of tools are evolving. So rapid, yeah,

Alexander Sarlin:

I was keeping an eye out before the conference, one of the big things I was going to try to sort of figure out as much as I could, you know, with one person's eyes is, as you said, it's sort of what's the vibe, you know, given the really negative news, I ran into the whole on IQ team who had put out this incredibly dismal funding report. And they were like, oh, yeah, it's rough out there. But I agree, the energy was very high. Obviously, AI is a huge win behind people's sales, as well as things like the acceptance and embracing of apprenticeships and work based learning. And there were a few really exciting trends that people were embracing. And I felt like overall, there was a little bit of a willful ignorance of the complexities of the business side. The other thing I would say about that, and this is, I don't know if this is even going to make sense, but let me put it out there. One thing I felt just so much at this conference is that we are not that enormous an industry right ad tech is not that huge, an industry. And there are so many players, you know, we did a presentation about the hundreds of AI tools that have come out, there's so many different folks in this space, even though it's not that big. And it it makes me a little nervous that the fragmentation is actually becoming a real bug of that tech world not a feature. It's uh, you know, people talk about competition being better and everybody having to work harder to be better if you're in a competitive red ocean kind of space. But that's not really the sense I'm getting I feel like if ed tech as a field wants to sort of regain its footing regain its interest from generalist investors, you know, become a must have for schools and companies rather than a line item that they can just sort of cut out like you're mentioning Ben. I feel like we kind of have to band together more than We have been and you know, you're in a room like the Manchester, the ASU, it feels like we're all banded together, it doesn't feel like they're people, you know, eyeing each other across the room, because they're competitors or huddling and trying to strategize, it feels like it's all one big group pulling in the same direction. But then we go back to our respective corporate headquarters are working from home, whatever we're doing these days. And don't feel like you know, there's enough people thinking about the entire sector, and how we can sort of prove our worth, and really be part of the conversation going forward. And I just wish there was a little bit more. And I hope, you know, that the community building we're doing the community building, future higher education community is doing, you know, some of these great community builders, I really hope that there's a little more of a pulling together to really take this moment that, like you said, it's a really split moment, the potential is huge, but the money and the customers may not be there, you know, how do we not only you know, survive this moment, but actually turn it into something fantastic. I think it's going to take more than one company, maybe even more than, you know, one company acquiring for other small startups, it might real collaboration across the field to really make a difference. And I don't know what that looks like, but I want it to happen. Yeah,

Ben Kornell:

totally. And you can even see that funders are actually, you know, philanthropic funders are saying, Hey, I'm not willing, or able to pick the winners when you all work together. And let's figure it actually,

Alexander Sarlin:

I just wanted to pull like, you know, 10 different AI founders and be like, you do this, you do this, you do this, you do this, like make it into one tool. And then it'll be a killer tool. But like, you don't want to be competing for the same dollars, because you're all kind of being point solutions. And I know you don't want to be so I don't know, it's such a wacky moment, especially in the AI space, which is just, you know, so frothy. Yeah.

Ben Kornell:

So let's talk a little and folks can read our newsletter. I think there's more in depth analysis of ASU GSB. There, also tons of great posts on LinkedIn with others perspectives. So it's really good to kind of get a syndicated view of the ASU GSB experience, because it is multisurface. Yeah, I think we should do a little bit of a round the world and AI, as well as talk a little bit about what's going on deeply in the education and tech space. On the AI world. Any big headlines that popped in your newsfeed.

Alexander Sarlin:

So you know, you mentioned there weren't any big, really big announcements at ASU GSB. And I agree, they're the one thing that really jumped out to me other than that, you know, LAUSD has been doing huge publicity push for their ad program with all here and really trying to sort of lead the way in terms of how to incorporate AI into a school district. But was meta I mean, you know, so we had Nick Clegg very, very high up very old school, you know, guide and meta at the Ed Tech conference, you know, at ASU GSB announcing, basically that meta is going to really try again, for lack of a better word in education, they're going to release a whole slew of new apps, with partners such as Roblox, and other third party vendors. It's not about them making their own stuff, but it's really about an ecosystem. They're going to create teacher tools so that teachers can actually manage multiple headsets and sort of create a VR classroom for lack of a better word, rather than having to have you know, two sets in a classroom of 30 and figure out what the heck to do with them. And I imagine, although I don't think I saw this, I imagine also really have a clearer sort of business strategy for what it takes for a school to incorporate the oculus at scale at what is the price point that actually makes sense for them? What are the outputs that they can actually measure? And point to? That was a surprise to me. I mean, we had talked to Monica as a long time ago about some of the big challenges about getting VR into schools practically. And it feels like, I think that the real takeaway is at Mehta has sort of realized they've been they put a huge amount of money and effort into VR. And they don't want to just do gaming education is a pretty close second to gaming in terms of the amazing use cases. And I think they're really trying to lean into it, whether or not it'll work, who knows. But it was interesting to see them sort of doing a hard reset, and basically trying to say, hey, VR education is not only not dead, we're like, we're turbocharging it, we're putting tons of money into it, and we're going to make it work. So that jumped out to me. Yeah,

Ben Kornell:

I think the challenge that we have here is the capabilities that we have in the education space, make it really, really hard to scale what works. And so when I am hearing some of these big announcements, especially the metal one, I'm thinking, Okay, now we're creating basically the infrastructure that allows us to scale and expand. And so just back to your original point about we need to all come together. There is a way in which like, we need these vast infrastructure providers that include AI companies that include headset provide There's that includes school districts like LAUSD creating infrastructure that allows for the application layer really to scale to all people.

Alexander Sarlin:

Yes. 100%. And, you know, we talked to Quinn Tabor from the murse. And they've had this really, you know, nice relationship with metta because they've been sort of a poster child application as a VR language learning platform and a social platform. But you know, it's funny, I mean, Merce was not at this conference, as far as I know, they certainly weren't showing anything at the air show that I know the founder just had a new baby. But it's like, this seems to be operating on different tracks, right? It's like meta saying, We gotta get into school. So we're gonna go talk to Roblox and immerse and then you have dozens of people trying to do various types of really interesting VR and AI content that just would cut off their left arm to be included in the in the Oculus, something's sort of not gelling enough here to make it really stupendous.

Ben Kornell:

Yeah, totally. You know, on the AI front, I think the headlines that I'm hearing about are a lot of the downside. Yeah, there's, I think, a lot of optimism about what AI can do. But we're also all aware that, you know, technology in the hands of humans with Mal intent is incredibly negative. And so we've seen a couple of articles around deep fake news kit produced, as well as now an article about people basically creating deep fake news of kids and then trying to extort them. I think there's a really, really big challenge here with image generation, the use of photos of anyone and AI and the capability of creating false pictures. I also in seeing a lot around the AI Literacy Day, which was on Friday is right horse. So it was a national AI Literacy Day, there were five organizations that came together, the White House participated, you know, First Lady Kamala Harris spoke, it was really around how we need to have aI literacy for all people, not just kids, and that kids are actually the natural place to start as a channel to teach parents and families about the dangers and challenges and opportunities with AI. But some of the examples that you heard about, which included you know, AI, girlfriends and boyfriends, kind of active misinformation, the inability to tell who's human who's, who's not. And that erosion of trust overall, it really was a glimpse into the future around when we take AI to its logical extremes, the kinds of complex decision making and evaluation that kids and really all humans will have to be, they're going to have to have great metacognitive skills to understand is what I'm looking at human is when I'm looking at AI, to what degree does it represent real perspective or an artificial one? And how should I feel about that? It's just a really interesting ethical space that we're headed to. And I don't believe that we have, you know, frameworks as a country or as a global community to deal with all this. I agree.

Alexander Sarlin:

And I you know, I'm not going to get on my my same soapbox about this, because I feel like I do it every week now. But I really think these types of stories are a huge threat to the adoption of AI, especially anything involving children, or in a official environment, like schools or universities, because they're just so scary. They're incredibly concrete. They're incredibly specific. They are just they're haunting. You know, when you read any of these, these stories about what are the things that can go wrong? It just feels like, you know, it behooves all of us to make sure that we are very vocal about the things that can go right. I mean, one metaphor that always jumps to mind to me is the early days of internet dating. when the internet first came out, and they first started, people started meeting in real life after meeting on the internet, it was considered like the craziest thing to do the news every night was like this 10 year old met somebody who they met on the internet and horrific things ensued. Right. And now we're at look here in 2024. And like 80% of marriages come from online dating, it completely changed everything about dating and romance in the world. But at that time, people freaked out about the sort of the potential and the catastrophic potential. And it's not that those things are real, they are real, but I think they're going to be the edge cases once these technologies are mainstreamed, and used in really amazing ways on an everyday basis. Like, you know, for every one couple who got, you know, dated on an online site and gotten married. There are 10,000 couples that dated and got married on an online site. One person had, you know, an abusive situation or something bad happened, but it's so easy to imagine and so scary and you talk to the victims of that and it's freaks everybody out. So it just, you know, we need the other side of the story. I, I'll stay off my soapbox, but we just need the other side of the story, especially in education, this deep fake stuff and literacy. I mean, people mean really well when they're trying to do AI literacy. But what they end up doing is highlighting all the horrific, crazy things that might happen. And that actually, I think, has the potential to really scare people.

Ben Kornell:

Yeah. And your analogy also points out that, is society better off for having online dating? Or is it not better off writ large, and I think different people would have different perspectives on that. So kind of coming back to where we are on the AI headlines, though, I think the train continues with open AI announcing new features and speculation around GPT five release. And there's also a growing sense that we're going to be getting lower and lower costs, per compute or per token coming really, really soon online, some announcements from Nvidia at the TED conference, and so on. So ultimately, I think the main takeaway is, you know, ASU, GSB great optimism about what AI could do really tough business environment. AI is really great optimism about how fast the technology is evolving and how cheap it is. And a challenging adaptive change for humanity. Like it's going to be about the behavior change, not actually about the technical change. At the end of the day. Let's dive into the education space. What are some other headlines that are peeking your interest? Yes,

Alexander Sarlin:

I mean, we saw a really big round of for a very mature ad tech company we saw and fella G previously get it sort of the owner of the blackboard system raised $250 million for its strategic initiatives and core ed tech solutions. That is a kind of number we are very unused to seeing this year. Or recently. It is a you know, that is a seriously large check and their anthology blackboards been around a long time, hundreds of hundreds of institutions. It's one of the bigger LMS is in the university space and been losing market share. But it still is one of the biggest ones. It was just nice to see a number that big tech round, even if it is somebody who has been in the space for quite a while Time magazine also put out their list. First time I've ever seen this have top ed tech companies of 2024, including also a shortlist of top sort of emerging ed tech companies. I might have missed this in the past, but I have never seen a Time Magazine list of top edtech companies they did this in connection with statistics. I mean, my first impression is awesome. Wow. Okay, Time magazine was thinking about edtech. Then I actually looked at the list. And I'm like, I have no idea what they're talking about here. I mean, it was a very, very, very odd list of companies, the especially the ordering was just bonkers to me. So you know, I think next week, Matt tower and I are going to are going to dive into some of their findings, because everybody should take a look and see how you feel about that. For the people who are sort of in the field. It felt pretty wacky to me. Yeah,

Ben Kornell:

that one really just made me both do a wow, and an eye roll. And that'll be fun. And by the way to all the companies that made the list or did not make the list like you know, we support you and we just think that it's pretty cheeky, that time would launch such a list without consulting, tech insiders. Right. One headline that caught my eye was $100 million dollar evergreen fund launch in Europe by the founders of the Ed Tech X festival. They do London Ed Tech Week, and they've been behind several education specs. It's Benjamin via dren croquettes, who started this new fund called June X capital partners. And what's interesting about it, it's 100 million euro fund, but they're using different mechanisms not just venture. So it is debt. It is, you know, short term capital, it is long term capital, it is an evergreen fund, basically, meaning that any returns get reinvested back into the fund. And of course, there'll be a management fee and all of that, but I think it is a signal of things to come where many entrepreneurs that I'm talking to, are looking outside the typical VC pattern of, you know, precede seed A, B, C, D, and the growth metrics that are associated with that. Part of that is the VC requirements to fund are like I need to see a path to profitability and I need to see positive unit economics and I need to see all of this growth and if you've got all those metrics, why are you going to dilute yourself with venture money? Why not get a loan, but it also shows that new capital coming in is kind of assuming the kind of shark ish venture debt and going for More like small business loan style financing of funding, but at a much bigger scale. And so I think that's a really interesting thing to watch, especially in Europe, where this kind of grow at all costs thing is actually culturally anachronistic. And there's a way in which people really want to see solid slow grow companies over a longer period of time. So I'm, you know, I was very impressed with one ability to raise 100 million euros in this environment, and then to this kind of new fund structure.

Alexander Sarlin:

Can you imagine a world where part of the role of the VCs in this space is actually sort of combining ideas so that they become something that is actually venture fundable rather than looking at, you know, 100 ducks in a row that are not, which is where we've been for quite a while? That would be interesting. Yeah, I

Ben Kornell:

think so. And, you know, also some things that we can't talk about publicly yet. But you know, in the space, we're seeing people moving towards in house m&a, in house operations support with this idea that we need to consolidate around our bets, not make more bets. Yeah. And so that will be another really interesting trend to watch. Well, with that, I think we can go to our guest. So why don't we jump into our interview?

Alexander Sarlin:

Yeah, we're talking to Rebecca Kockler, CEO of Magpie, and stay tuned after that, for Ben's interview at SXSWEdu feels like ancient history now, but it was only a month ago with incredibly interesting, folks, the curriculum associates and soapbox labs, merger that has recently happened. It's sort of big news in the AI space as well enjoy. We have a very special guest on the weekend ed tech this week, Rebecca Kockler CEO and founder of Magpie literacy, which is doing incredible things, very new things in the reading space. Welcome, Rebecca.

Rebecca Kockler:

Thank you so much for having me, Alex, great to be on.

Alexander Sarlin:

It's great to have you on. You know, you and I have known each other for a couple of years now. And you've been doing unbelievable work. But before talking about Magpie tell us a little bit about your education background, you've worked in a lot of different contexts in education and brought you to this moment in edtech.

Rebecca Kockler:

Yeah, I've started by teaching, I taught middle school reading in history. And New York, New Jersey, which I loved. And to this day was the absolute hardest job that I had. And then I spent my time honestly, most of my career working in public education focused on helping teachers be great teachers and helping kids learn. And so I was the chief academic officer in Louisiana at the Department of Education for six years as Chief of Staff at Los Angeles Unified. And before both of those spent a lot of time coaching teachers and principals on instruction, and how to teach reading and writing and history and all the things. Awesome. Yeah,

Ben Kornell:

can you tell us a little bit about the founding journey of Magpie, obviously, through all of those roles, you know, teaching kids how to read and literacy has always been core to what you've been doing, but you decided to start a company. Tell us a little bit about that. Yeah,

Rebecca Kockler:

it's so unlikely in some ways that I'm doing this and that I started this work in this way. In the big school systems I was working in, we saw a lot of success, we saw growth on almost every metric we measure. In Louisiana, we saw massive growth with our most struggling students, most underserved students in our system. And, of course, like none of it was sufficient, and especially our work and reading, we just could not see the change I wanted to see in our middle school reading results. And it was so frustrating, because it's what I taught, it's the thing I'm most passionate about, I know the most about and I couldn't understand why the strategies that we were leveraging in other areas weren't transferring, especially when we're looking at sort of fourth grade and above reading. And so I really started this work out of that frustration, feeling like I wasn't getting good answers from research, the research community, I wasn't getting products from the market that were really actually delivering on results for kids maybe short term here and there results but nothing deep and sustained. And so we started this work. We have really two organizations. I lead our research arm at reading, reimagined under air def, and our product development arm under magpie. And we're really focused on funding better answers to why so many of our students especially our black, Latino, Native American students struggle for so long, and what we need to do differently about how we teach them and we really believe that technology can help us see much faster pace and more equitable piece of ensuring that more of our students get access to the kind of teaching and learning that they deserve. So we believe products and tools deeply seeped and built in connected to the most up to date research and tackling the bias that exists in so much technology can have a real outpaced impact on our kids. So we always say we ultimately collectively are pursuing the eradication of illiteracy we believe absolutely we can eradicate and should eradicate illiteracy in our lifetime. I definitely think technology, if we focus it correctly, can help us achieve that even faster. And we believe that, again, this sort of coming together have deep research partnership immediately in from the beginning with our students and their communities. And technology is going to help us solve this problem in a really new way.

Alexander Sarlin:

There's been a lot of focus in recent years on a really hard shift to the science of reading. But you have known that for a very long time being in this space, science of reading is sort of old news for you and for your work. And I know you've been thinking a lot about, you know, really pinpointing the exact places where reading goes off the rails, especially for your target populations. Tell us a little bit more about, you know, what you found in your research and what Magpie is aiming to do, yes,

Rebecca Kockler:

science of reading is so important, we should all be implementing the things we know work in research. And specifically, I think science of reading translates to our understanding of the foundational skill sequence to coding, fluency, vocabulary. For younger kids. It's really how we think about it. And again, we were doing that in Louisiana, we've been doing that, you know, in some places for decades, I think the movement around science of reading is pushing us to make sure we're doing it everywhere in our country, which is great, and we support. And again, we tend to look at third and fourth grade reading results as the suggestion that we figured it out that we implemented this quote unquote, science of reading, and kids are doing fine. And yet, it's actually after fourth grade when we see so much kind of regression where fourth grade tends to be our peak results in this country in how kids do in reading. And we tend to see worse results after fourth grade. And that's the thing we're trying to understand why even when we do meaningful decoding, which is just you know, the work of putting sounds to letters and putting those letters together to form words. And we need to teach it systematically when we do that, why is it not transferring and helping students pass fourth grade, there was a sort of simplistic explanation that it was all vocabulary. And I think our research is finding this focus on something called the decoding threshold that decoding continues to really hold our students back beyond third and fourth grade, which is really not understood well. And it's a unique kind of decoding, it's decoding and multiple syllables. So beyond just we always say, you can learn how to pronounce cat. That's super simple. But when you put cat in the middle of education, the cat changes, that's a completely different set of sounds, it goes from cat VAT, that seems VAT sound goes through case, that makes no sense in the English language. And we don't actually help kids with that transition from single syllable words into the complexity of multiple syllable multisyllabic word decoding. And that is the thing we really think is holding so many of our students back we know, when students cannot decode, they will never show growth in reading comprehension. We think about 40% of middle schoolers and high schoolers in America cannot perform multisyllabic word decoding at the rate that they need to to be able to access reading comprehension, they cannot make the transfer from cat into education. And yet there are almost no assessments on the market. And really no tools on the market to take this challenge on. Even though we think it's the biggest thing holding most of our students back who are really deep struggling readers, again, we need to keep doing vocabulary work, we need to keep doing this definitely known to do in kindergarten through second grade. But there's this massive hole in what we talk about in research based reading instruction. And that's what we're really trying to promote and take on and what we're building and supporting kids with. Yeah,

Ben Kornell:

it really flies in the face of the kind of Maxim which is learn to read and read to learn, you know how most people focus on the kind of decoding and core literacy elements earlier on, and then it's let go, it's actually incredibly parallel to math, when you actually find that people learning algebra, have foundational numeracy misconceptions, or, you know, base understandings of fractions or something. And so they end up struggling with algebraic thinking on algebraic equations, when conceptually they can understand algebraic thinking, but they don't have the building blocks skills to access it. So very, very, very interesting. In terms of, you know, how you imagine this playing out a lot of people in the AI space are championing these level reading capabilities. Where, oh, you know, I'm in eighth grade. So now I can do this at a second grade level and I can do things at a fourth grade level all the way up to a 10th grade level, and what the champions of that strategy are saying, we need to provide access To the information and the material at a level that's accessible to students, do you think that that's helpful or harmful? Or how would you kind of coach people on that? Because it is like a hot topic in this space?

Rebecca Kockler:

Yeah. I think our research would suggest that's it's not an exact yes or no answer. So do we think you should take the Romeo and Juliet and level it? And give students a third grade version of Romeo and Juliet? Absolutely not. And our research would suggest that you should absolutely not do that. Do students sometimes need access to texts so they can practice some of the skills they're working on in the context of texts where they can read it so that they can practice that? Yes, that's not leveling, that's alignment of practice books, which we call decodable. And we think even older kids need mature versions of decodable holes in the form, you know, of sort of graphic novels and other things. Not a different version of Romeo and Juliet, a specific thing for them to practice the specific decoding multisyllabic decoding skills, morphology work, other work that they might be working on. We are funding research right now on how to teach complex decoding and foundational reading skills in the context of grade level vocabulary and grade level text. So we think AI can be such a powerful tool, because in reading, kids can think and access Shakespearean language, no matter their reading ability, and they love it actually did a survey once a Louisiana, and most kids told us they wanted to read Shakespeare, like, you know, it was sort of interesting to us. And surprising. They love it. They love engaging in it, they should not be getting the cliffnotes version. We're actually building tools and AI can be a great partner where you can actually understand what is required to read components of Romeo and Juliet. And how can you use components of Romeo and Juliet based on the research we're finding about how to do even early stage decoding work in the context of grade level vocabulary and books with it. And that's such a great use model for AI. That's a real instructional strategy based on research, teach single syllable and multisyllabic word decoding in the context of Romeo and Juliet. And with that vocabulary and help pair those things together. Doing that individually for lots of different kids who have different needs really hard to do for a teacher, technology can play such an incredible role in that kind of a space. That's how we believe AI can be really leveraged in powerful ways right now is targeting things we know and research or how kids learn to read. But really difficult for any human to actually deliver on at scale, bring technology in to help us solve those kinds of problems. Versus technology just sort of being seen as the hot button thing to do right now. We're just trying to find any hole to fill to leverage AI. And that we think is not likely to lead to real contributions that push the envelope and how kids learn. One

Alexander Sarlin:

of the foci of your thinking, and your research that really fascinates me is that you know all the statistics on who learns to read who is held behind in reading who, you know, really struggles and you've worked in school districts that have very high populations of underserved students. A lot of people sort of give lip service to this, but you've actually really leaned into it and thinking you actually lean into the research and say, there's actually differences in how different students learn to read. And instead of just, you know, flattening it out, let's actually lean into it and make a difference and try to address the populations that are at the highest need. That includes black students, Latino, Native American, and students living in poverty. Tell us about those target areas and what the research has been saying and how you're incorporating into magpie. Yeah, you know,

Rebecca Kockler:

we believe that to eradicate illiteracy, we have to figure out how to help every single child in America learn to read. And one of the things that we have found in the research base is that so much research was done with mostly white middle class students. And that might seem fine, because of course, we know scientifically every person's brain learns to read in the exact same way, right? So like, Okay, that sounds great. But the challenge is, is that we don't have a sort of monolithic way that we speak in this country. And it turns out the way you speak has a lot to do with when it's easy and hard to learn to read. And specifically what we're finding is that your brain is pre programmed from the moment you're born to learn to speak, you literally have neural pathways built in your brain to learn to speak. It's such an implicit thing that we do, which is good, you know, for evolution, all the things right so we can function. Your brain is not pre programmed to learn to read, you have to teach it to you have to build neural pathways. That's what we do when we teach kids to practice these skills. Well, one thing we know is that different kids need different amounts of practice. Some kids will build that neural pathway in about five practices. As the average will be about 60 practices, some kids will need upwards of 200 practices to build a neural pathway. That alone starts to suggest you need some real differentiation in the app paths approach for different kids that has nothing to do with racial background. That's just all to do with just different kids seeing different things. And different components of that, that's fine. And technology can play an incredible role in providing that in a way, it would be very difficult for a teacher. But the other thing we know is when you're building those neural pathways to learn to read, when you're learning to read, leveraging the patterns in which you speak, building those neural pathways is much easier. And so when you're learning to read patterns, that are the same dialect and the same pronunciation and same cadence as your spoken language, it's easier when it's different, it's harder, and that has not been well researched, historically and understood. And there's some real leading experts on this in our country, not me, we partner in particular with Dr. Julie Washington at the University of Irvine, who's a leader in this field, there are some others in our country doing amazing work here. But what we're finding is, we have to figure out how to leverage the linguistic patterns of our students in the ways that they're learning to read. And it turns out, this plays out even more in multisyllabic words, cat sounds like cat in most dialects in this country. Education does not sound like education in most dialects in this country. And so again, we have this multisyllabic word challenge, where it plays out even more for students in some ways when they're younger, but especially when they start to hit this late elementary, middle school. So that's an example of the ways that we haven't done enough research in the places where our English language learners and our black students or dialect speakers have real differences. And we're not applying that so that we can leverage their linguistic strengths as they're learning to read, versus it feeling like a hindrance and confusion to them, which is often what happens right now. And we're teaching them to read.

Ben Kornell:

Wonderful. Well, thank you so much for joining. I feel like we have so many entrepreneurs that come from the tech space. We're trying to solve problems, and you're so deep in the educational space, and bringing forward some pedagogical misconceptions, I would say, and really addressing it through an innovative enterprise. So can you tell us if people want to find out more about Magpie where they can go?

Rebecca Kockler:

Yeah, if they want to learn about our research work, we're reading, reimagined at Arif, a er D f.org. And if they want to learn about the tools that we're building and the work we're doing in schools, you can find us@magpie.org We would love to engage with folks, partner with researchers and school leaders and students all over and yes, mostly just deeply believe we have to marry instructional precision understanding of the complexity of our kids and the schools that they live in every day. And the best of technology, we think it's that combination of things that's really going to help us build the best tools for our kids, and most importantly, can pursue the eradication of illiteracy in our country.

Ben Kornell:

Wonderful. So inspiring. Thanks so much for joining us. Rebecca Kockler, the CEO of magpie.

Rebecca Kockler:

Thank you.

Ben Kornell:

All right, Edtech Insiders. We have special guests today from Curriculum Associates Soapbox Labs, talk about AI and Ed Tech. Super excited to dive in one of the great panels. Welcome to the show.

Kristin Huff:

Thank you. Happy to be here.

Ben Kornell:

So why don't you tell us a little bit about your journey, your professional journeys and how this marriage of curriculum associates and soapbox came together?

Kristin Huff:

Great. Should I introduce my Yeah, Hi, I'm Kristin Huff, I am VP of assessment, psychometrics and research at Curriculum Associates. I've been there about eight years, I've spent my entire career in large scale assessment. I've worked on a lot of the larger tests like TOEFL, LSAT, AP, and a lot of state tests. And I got to a point in my career where I felt I really want to help teachers and students, and really dive into the learning journey. And that is what brought me to CAA. And one of the things that I have been most excited about in my eight years at CEA is that we really focus on that intersection of assessment and instruction. And that is really where the magic happens in the classroom. So this partnership with soapbox, I couldn't be more excited about it. So box, and their technology is right in line with our mission to make classrooms better for teachers and students. And so I'd love Amelia to hear from you and really ask you to talk about the wonderful work that Soapbox is doing.

Amelia Kelly:

Sure, thanks. So my name is Amelia Kelly. I'm the CTO of Soapbox Labs. I joined soapbox as one of the first employees back in 2015. I have a long history of speech recognition and linguistics. So my background actually originally was in physics and astronomy but then I moved to linguist X for a master's in a PhD from Trinity College in Dublin. After that, I worked for a start up in California. And I worked for a while up in IBM Watson. So, over that time, I got a lot of experience in a lot of different facets of speech technology from speech synthesis, to natural language processing. And, you know, working on IBM Watson as well, in the early days, as you know, very NLP. And, you know, we know a lot better now from GPT, and all that. But basically, you know, what I'm saying is, I've been building language models for years. But you know, it all came to the fore, really with soapbox labs, because when we started, we had a Data Set of Children's speech. And it was different from every other Data Set of Children's speech that was available at the time, because it was real world stuff. It was children's voices speaking naturally in noisy environments. And we were at the intersection of or this kind of inflection point we're talking about yesterday, of computational power was increasing. We had GPUs access to GPUs we didn't have before. And we had massive, massive data sets. So we were able to build very sophisticated models using very real world data. And the result of that was that we made very accurate child speech recognition. Over the years, as we developed it further, we realized that the use cases for children's speech recognition were different as well. So that's kind of how we got very hyper focused on Ed Tech. So as we developed our platform, we had a lot of customers. So box labs, before we joined curriculum associates, they were all building really, really good content, they were building really good ways for children and teachers to engage in learning activities like letter naming letter, sounds, segmentation, even multiple choice questions in gaming, and of course, oral reading fluency. So over the years, we built up our speech recognition to be able to do specifically those tasks, which really set us apart from off the shelf speech recognition, there's no way for example, you could use an off the shelf adult based Speech API to recognize letters sounds, children's sounding out words when they're four years old, early literacy. So we paid a lot of attention to those use cases. Then, when we met curriculum associates, it was a match made in heaven, really at soapbox with a very, very strong culture, everybody in the company is very mission driven. Everybody is building the speech technology, a because it's cool. But the because it's a means to an end, we want to help children, we recognize the role that voice AI can have in solving the literacy crisis, for example. And that is what we want to achieve. So we are very, very excited that we are now a part of the curriculum associates family. And we can now reach for 13 million children and rising and

Ben Kornell:

rising. So as a CTO, you know, every CTO I've ever met, they love seeing their product in the hands of users. Can you tell us a little bit about how speech and assessment specifically fit together and given your two specialties? One thing that we've always believed at ad tech insiders is assessment really drives the flywheel of innovation. And if you shift how you assess, or you expand accessibility of assessment, it really opens up opportunity for teaching and learning. So can you tell us a little bit about that?

Kristin Huff:

I would love to talk about that. I love your question that it's so insightful. And I believe the same thing. I'm going to take a broader perspective here for a moment, we've got to do better in both assessment and instruction. We see data coming from, for example, Nate that shows year after year stagnant growth, and now dips post pandemic, curriculum associates own data based on millions and millions of students show that we are not seeing the acceleration and learning that we need to see post pandemic for the students to catch up. We're even seeing impacts at grade k one and two. And these are the critical years when students are learning how to read. They are learning their basic numeracy skills, both of which are critical for the future. So when we think about the power of voice AI, to not only make this assessment richer, deeper, more authentic of these early literacy skills, it also provides deeper, richer feedback to teachers who can then accelerate learning. So this use of technology where we have centered the needs of students and teachers, rather than just trying to create a technology that can then be kind of sprinkled on an existing product. It really does open up a world of impact on teaching and learning that we haven't been in before.

Amelia Kelly:

Yeah, and it's great to have been able to talk about this as well yesterday at the panel. We're here at South by Southwest edu We had a panel yesterday called The Future of assessment is invisible. And that's really the word assessment and voice AI meet. It, as he said, allows this extra data stream of information that the teacher has access to, that gives them information about how the children are doing as they practice. So you can tell using the soapbox technology, how well they're pronouncing words and reading words right down to the phoneme level, for example. And with this information stream, class of 30 children, the teacher then can see every time that they practice, how well they are doing, whether they're improving or not, and then choose to intervene, where it's appropriate, or where it's most needed. That allows the teacher to use their time wisely to do invisible, that assess assessment. And really, you know, so

Ben Kornell:

we're seeing a broader shift from summative assessment to formative assessment, you know, more and more emphasis on formative formative, formative and rapid cycles. From a product standpoint, now your team that has been acquired really has to work kind of cross functionally with the curriculum associates team. For other people out there who thought about like, being an m&a or deep partnership. How do you manage that what have been some like keys to success for working across the entire product portfolio?

Amelia Kelly:

I think we've done this from the beginning at so bucks labs, we have a very engaged very knowledgeable product team at soapbox labs, and they work very closely. First of all, with our engineers and speech scientists, and our computational linguist. If anyone has ever worked with the speech scientists like myself, we can be hard work. But know that the team are absolutely wonderful, and they work together. So we already have that experience of working cross functionally from product to tech. But also when we were engaging with clients, we would work with them. So we were building the voice AI that powers product subs in the past, other people were building now its curriculum associates products. So tech team and our product team have experience working with people who build content, we've gained that experience over the last 10 years, and now working with curriculum Associates, and it's, you know, more of the same with they've got wonderful product teams, we're able to work closely with him whether the experience felt very seamless.

Ben Kornell:

Today, curriculum Associates has exclusive rights to soapbox labs in the past, you scholastic was one of your partners, and you had many other scale partners. How does that change your outlook on where the company goes and where this speech recognition goes?

Amelia Kelly:

I think that it's all about reach for us. So I suppose when you're dealing with multiple clients here spread quite 10 different clients are at different stages of development, working with curriculum associates, were able to focus all our energy on getting our voice tech into products that will reach kids at scale. And

Ben Kornell:

just to be clear, the entire team from Dublin has come and joined with curriculum Associates is all over but the Boston like hub, it is kind of a marriage that I think everybody in edtech would like to emulate where you're basically doing exactly what you're passionate about, but at a scale that is, you know, 13 million and growing. Last question, just on the assessment side, for the learners that are younger. So I read is the famous curriculum associates product. But now more and more schools are having TK and there's things like early diagnosis of dyslexia or other reading disabilities. How do you just in the strategy of curriculum associates overall, how do you see voice playing a critical role?

Kristin Huff:

I think that the voice AI technology that has been developed by soapbox, I mean, one of the reasons that we are so enamored with this technology is what Amelia was saying a few minutes ago, the fact that we are able to authentically assess each student with their voice, regardless of dialect, regardless of linguistic or rather than say, regardless of linguistic diversity, we're honoring linguistic diversity with this technology. And so it helps us with our commitment to equity and making sure that we are measuring all students from a strength and asset based perspective so that we can give this rich data back to the teacher, not only about where students are in terms of their strengths, but also where they need to grow. And this is that formative assessment process where teachers can then use that information to further all students along the learning trajectory. And we know from decades of research that when you do formative assessment, right when is integrated and seamless in instructional practice, when it's invisible, that that has a huge significant impact on student learning. I

Ben Kornell:

love that I love that framing of it being invisible. And it goes along with this idea that it's formative that it's seamless. It's integrated. This is not an API sprinkled on top. It's actually built in the form factor. I would also just say, we get questions all the time around AI defense. ability in education in edtech, what's to stop? You know OpenAI from launching some new product or whisper or something like that. But the nuance of child speech is actually quite, quite defensible. And it's just another example of where, you know, purpose built AI products for kids, for learners, for families, and educators often can scale in ways and indefensible ways where large AI can't make that impact. I don't know, Amelia, if you have any thoughts on that. But, you know, it is an exciting time in AI. And this does seem like a really defensible space.

Amelia Kelly:

Well, I think you said it better than I could have actually, that's exactly our approach, instead of, you know, using the foundational model approach, where you're pretty much just taking a sample of the entire internet and every voice that is available. The problem with just using data that's easy to get is that it's very heavily skewed towards one particular demographic, when we're building our models, we do it intentionally we do it with consideration, we make sure that we're in the training set of the models capturing the voices of as many different students and speakers and accents and dialects as possible. And then we test our models very, very rigorously for those different sites, different accents, different ways of speaking. And we update our dictionary, so it has different pronunciations available. And we want to make the product as equitable as possible from the ground up from the very beginning when we make choices about what goes into the models. That's why we have we're a company the only company in the world still to affirm the digital promise micro credential for prioritizing racial equity in AI design.

Ben Kornell:

I'm super excited to see where this goes. I've been falling soapbox, since like an NPR story, I don't know six years ago about children's speech and voice recognition, Alexa and all that so it is really amazing to see the journey. Kristin from Curriculum Associates, Amelia from Soapbox Labs, what a partnership coming together. Thanks for sharing with us today.

Alexander Sarlin:

Thanks for listening to this episode of edtech insiders. If you liked the podcast, remember to rate it and share it with others in the EdTech community. For those who want even more Edtech Insider subscribe to the free Edtech Insiders newsletter on substack.