Edtech Insiders

Week in Edtech 12/16/2024: OpenAI’s Sora Launches, The Era of Multimodal AI, AI Wars Heat Up, Texas Curriculum Changes, Declining Public School Enrollment, $175M AI Investment in HE and More! Feat. Henrik Appert of Magma Math and Alex Linley of Cappfinity

• Alex Sarlin and Ben Kornell • Season 9

Share episode

Send us a text

With the holiday break just around the corner, join hosts Alex Sarlin and Ben Kornell for a timely look at the latest developments in education technology. From exciting AI advancements to shifts in curriculum policy, this episode captures the key trends shaping the future of learning.

✨ Episode Highlights:

[00:03:16] 🧠 OpenAI's New AI Tool "Sora" Makes Waves in Video Generation
[00:06:01] 🎥 The Era of Multimodal AI: Google, Meta, and Microsoft Compete on Video and Agents
[00:08:05] 🔍 Who’s Winning the AI Wars? Exploring AI Specialization
[00:18:14] 🏫 Texas Education Agency’s New Curriculum Policies – What It Means for EdTech
[00:25:55] 📉 Declining Public School Enrollment: Implications for K-12 and EdTech
[00:27:05] 🌍 Global EdTech Trends: European VCs Raise Big Funds Amidst U.S. Challenges
[00:28:27] 🚀 $175M Investment in AI for Higher Education by Element451
[00:32:13] 💼 Workforce Upskilling Revolution: SaaS Startups Receive Major Investments
[00:33:40] 📊 EdTech Year-End Reports Show Optimism in AI-Driven Upskilling

Plus, special guests:

[00:35:33] 🎙️ Alex Linley, Co-CEO of Cappfinity, discusses the future of skills-based hiring and education.
[00:54:34] 🎙️ Henrik Appert, CEO of Magma Math, shares insights on their award-winning math platform and recent $40M funding round.

😎 Stay updated with Edtech Insiders! 

🎉 Presenting Sponsor:

This season of Edtech Insiders is once again brought to you by Tuck Advisors, the M&A firm for EdTech companies. Run by serial entrepreneurs with over 25 years of experience founding, investing in, and selling companies, Tuck believes you deserve M&A advisors who work as hard as you do.

[00:00:00] Alex Sarlin: Is this the type of tool that is trying to compete with a, you know, small gap in what the current models do that they're probably going to fix in the next launch? Or is this something that can actually benefit from the very fast evolution of the underlying models? When video comes out. Is your company going to get better or is it going to get wiped out or a very hardcore competitor or a bunch of new competitors because you're filling in a gap?

And I think it's a really important question, even though it's hard to define.

Welcome to EdTech Insiders, the top podcast covering the education technology industry. From funding rounds to impact to AI developments across early childhood, K 12, higher ed, and work. You'll find it all here at EdTech 

[00:00:49] Ben Kornell: Insiders. Remember to subscribe to the pod, check out our newsletter and also our event calendar.

And to go deeper, check out EdTech Insiders Plus, where you can get premium content, access to our WhatsApp channel, early access to events, and back channel insights from Alex and Ben. Hope you enjoyed today's pod. Hi

everyone, and welcome to the Week in EdTech with EdTech Insider co hosts, Alex Arlen, Ben Cornell, we're so excited to have you here. One more show till the end of the year. My gosh, it's going to be our famous End of the year wrap up and future forward predictions episode. I for one am super excited and we've got our recommendations and our predictions pouring in from the ed tech insiders community, lots of people sharing their insights.

So if you have any, please shoot them over to us. You can also check out our. AI generative map that's edtechinsiders. ai. Before we jump into the news, Alex, what's going on with the pod? 

[00:01:58] Alex Sarlin: Yeah. So we've had just an incredible suite of guests for this December. We talked to Laura Ibsen from Ellucian. You talked to Trish Sparks, the CEO of Clever, and we just put out a really cool episode with Gavin Cooney, the CEO of assessment giant Learnocity.

And then we're doing something very special for the Christmas week, starting on the 23rd. We are putting out an episode. Each day of that week about Google, and here's why, you know, but you and I were at this incredible Google event. I think you were sort of instrumental in getting it off the ground, and it was just unbelievable.

And we got to talk to all of the learning leads at Google across the entire company as well as other people. A number of people sort of on the ground. And what we're going to do is we're putting out our postcards episode on the 23rd, that will be sort of bits and pieces, like a montage of what it was like to be at this amazing Google AI learning event.

And then each day we're going to sort of unveil a little bit inspired by open AI is sort of week of shipping, you know, each day we're going to unveil one new episode, which is going to be full interviews with two Google leads. And. They were so incredible. I'm so excited. And then right at the end of the year, we cap it off with an awesome conversation with the great David Yeager, who is basically, I think one of the leading social psychologists in the country.

He's basically the heir apparent to Carol Dweck and has just published an incredible book called 10 to 25. That's incredibly relevant about motivating young people. In ed tech. So we still have a bunch of interviews to publish. And if you have some time over your Christmas or holiday break to go for a treadmill jog, or if you're somewhere where it's warm enough to go outside, get outside and listen to some of these amazing podcasts.

Cause I mean, you know, I'm not somebody who likes to sort of pitch a lot of these episodes, but wow, these are some really, really good ones. So excited for the break for everybody. And Ben, how about you? What is the new year look like for you? Well, 

[00:03:51] Ben Kornell: for upcoming events, we are going to have the Bay Area EdTech Summit.

It's our seminal event in San Francisco, hosted at the Cooley offices, and that's going to be February 12th. So please mark your calendars. We're doing a little bit of a break here over the holidays. No events here for the next couple of weeks, but February 12th. It will be a half day and then followed by small group dinners, which I think everybody loves reconnecting the new year with that.

Let's jump into the news. Maybe we start with our around the world in AI and let's start with open AI. Open AI had their kind of 12 days of Christmas with a bunch of launches. There's also a great article about how they crushed Chegg. I don't know whether that was self inflicted or whether that was open AI, but you know, there's a way in which the narrative around AI disrupting answer platforms, I think is a very strong narrative.

Sora came out. And so for those who weren't following the story too closely, Almost a year ago, Alex and I interviewed Sam and Sora was just coming out at that point. And he even mentioned it on our podcast and they edited it. And then here we are a year later, one of their focus groups, which, you know, they've been creating these advisory communities and one of the advisory communities of artists.

Leaked the Sora model. And so in some ways you wonder, was this a planned rollout or is this reactive? But it is really a big moment in terms of video AI generation. And we see that so many of the other players, and we don't often talk about mid journey here, but. Mid journey is crushing it on the generative AI video as well.

You've got stable diffusion with an AI video model, Gemini, which we'll come to in a second with some video creation. So the world of multimodal, we've been talking about it and talking about it and talking about it. Now the form factor is not just a text or image output. We're seeing video too. So that to me was the big headline in OpenAI land.

What did you have Alex? 

[00:06:01] Alex Sarlin: Yeah. I mean, that was back in February when we talked about Sora and it is exciting that it's finally here. Partially, I think, you know, as with all of these AI moments, all of the big model creators, as well as the more specialized companies like Midjourney or Runway that have focused on video, as well as ones in the education space like Colossian or Profgym, you know, basically all of these different companies.

Look at to each other very carefully. They're all very much locked in, not quite, I mean, in competition, but also just watching the space evolve. So the reason I think it's a big deal that Sora is finally coming out, as well as all of these other video models from Meta, from Gemini, from Google, is that now they're getting out in the world.

People can try them and they can continue to ratchet up the competition and they're just going to get better and better, very, very quickly. Quickly, which is exciting for all of us in education, especially who have been waiting for sort of the moments when video becomes a very cheap and easy way to deliver information.

We also saw Google announced this week that YouTube has a new model that can basically do AI dubbing into nine different languages. Speaking of. You know, video AI, which means that if you have a learning video on Google, it can move into nine different languages, not only with perfect, you know, overdubbing, but actually sort of look and feel very much like it's in that native language.

So we are just seeing the sophistication of the platforms around video ratchet up very quickly. And I think it's going to open up all kinds of opportunities for the ed tech space. 

[00:07:28] Ben Kornell: Yeah, you know, over this, you know, past two week period, it does feel like the race on AI has pitted Google versus OpenAI quite directly.

What do you think about the other players, Alex? You know, Microsoft had some news this week. Meta also has a new model coming out, a new LLAMA model coming out, which has multimodal or omnimodal capabilities. We haven't seen Microsoft as much in the forefront, but it does seem like their co pilot is generating some revenue.

What's your read on, like, who's winning the AI wars and where we're at here around this turn of the year? 

[00:08:05] Alex Sarlin: Yeah, what an interesting question. I mean, I think that they all are. Finding their lanes. Let me put it that way. I think, you know, we've talked a lot this year about how these big model creators are sort of in direct competition and they're just trying to sort of claw past each other and be the most sophisticated model on some of the AI benchmarks that are out there.

But I think we're starting to see. Different companies sort of find areas of expertise that they feel like they can sort of own and really lean into. And I think your example with Microsoft is a great one. Microsoft, we don't talk a ton about it. It's been very close to open AI for a long time. It's put a lot of money into them.

Some people question how close they really are. But Microsoft has found a real lane in the coding space and the Microsoft coding copilots have been incredibly sophisticated and working very well. And I think they've really. You know, behind the scenes started to really change the coding landscape for professionals.

They just launched something called copilot vision, which basically allows the Microsoft AI tool to read your screen and actually sort of operate on things on your computer. We've seen Anthropic do something similar recently, where they offer an opportunity to basically Let the anthropic model take over your screen and complete tasks for you.

So in one way, there is a continued competition in the agentic space, right? You know, these models can do more and more and as they can do more and more, they are each sort of reaching into our systems, making more interoperability between tools, allowing it to search the web in more sophisticated ways.

Open AI launched a really good web search as well and starting to really, you know, just try to sort of close the loop, which I think is something. Those of us who use AI a lot really actually appreciate because you can get great answers from AI, but the fact that they can then put them into spreadsheets or put them into a code base or do it in context is very powerful.

We saw Anthropic this week double down on what I think increasingly they've seen as their specialty. On one hand, their specialty is ethics, right? That's Anthropic's sort of raison d'etre is splitting off from open AI to try to become the Ethical model creator, the one that really tries to keep humanity at the forefront that really tries to keep ethics at the forefront.

And I think they've done a pretty good job of that, even though it's a little bit hard to tell at this point, frankly, but they've also been trying to do transparency. So they've done a lot of studies to basically try to figure out how their model is actually thinking to open the black box. And now they've just provided something called the model context protocol, which is a new architecture for.

Basically, helping models get context from external systems. Sometimes there's a technique called retrieval augmentation generation, which basically means, Hey, go out in the world, get some additional data and use it to tailor your request. They're sort of doubling down on this and starting to think about it in an even deeper way and saying, well, what if you wanted to attach to a whole big database?

What if you wanted to attach to a code repository? What if you wanted to attach to your entire, you know, Google drive or your computer's file system? Well, that suddenly it's not just. An article that you're using, it's a lot of data that's being used to supplement whatever you're asking the agent to do that matters a lot.

And I think they're being very systematic in thinking about that infrastructure. So I sort of think the anthropic is doubling down on sort of the infrastructure, transparency and ethics. To some point, open AI is continuing to sort of become a product company. They launched a new one model. They've launched a new voice model.

They launched Sora. You're seeing Microsoft sort of start to focus on the coding space. You're seeing meta. Focus on the open source space, as we know, to create all of these very powerful models that can be built on top of. And then Google, as we've talked about all year, is continuing to put their AI into everything they can touch, which is a lot.

If you're Google, they're putting things into the pixel space. They're putting things into YouTube. They're playing really hard with agents who can sort of work across the Google ecosystem. So we're in a really interesting moment. There's sort of the second tier competition around video and agentic.

Models. But I think even within that, you're starting to see each leading developer sort of start to find a lane that differentiates it and where they can sort of double and triple down on their development. That's my read. What do you think? 

[00:12:09] Ben Kornell: Yeah, I mean, I think, you know, we can save some predictions for the next episode for our year end prediction one, but I totally agree.

And I also think that this idea that you as a user or you as a company will only engage or interact with one model. unlikely. You'll probably, like you said, they'll have different lanes and you'll use different things for different purposes. So it may end up being that chat GPT is your model that you use as a consumer for fun.

And then you're building infrastructure for education with Anthropic because you like how you can collaborate with your team and how you can. Tune it for an education purpose. I think for the entrepreneurs in our audience, I think the reality is that you're going to have to have API enablement for multiple models, just because the rolling nature of these releases allows you to continue to experiment and test with the quality of the outputs over time.

And that that is really what you should be, you know, exploring and testing. Last thing I would just say is. I think there's an opportunity for smaller bespoke models, not from the big tech platforms. And you know, my hope is that because we've seen these sub models, ones that can even live on your phone actually be quite effective and that they can live in relationship to these larger models.

I'd love to see a new move of R and D where some of our companies that are leading edge. are actually building their own models that allow for, you know, agentic workflows to be as effective as possible. So it's going to be, you know, a sneak preview. It's going to be exciting to see how this plays out, but I do feel like we've now established who the big players are coming into the ed tech space.

There have been a number of rounds to close the year, some of which we do know about and don't know about, but this idea of leveraging AI and some of these advanced reasoning models does seem to be playing into our space. One would be magma math, who we're going to have Henrik here on the show here in a moment, but you know, the kind of life cycle of venture capital and ed tech seems to be boosted.

By the capabilities and qualities of these new models that increasingly have reasoning and logic. And as you say, could boost the learning and personalization potential. So as much as sometimes in tech tech, when things advance and evolve, it crushes ed tech companies. In this case, I actually think it's going to enable a small company with a small team to do so much more than they could do in the past.

So it does seem like a really strong trend. 

[00:14:59] Alex Sarlin: I totally agree. The only thing I would add to that last part is that I've been trying in my conversations with different companies, different products. And, you know, we've been doing our best to make sense of this fast changing space to try to sort of create a mental litmus test for exactly that.

You know, is this the type of tool that is trying to compete with a you. You know, small gap in what the current models do that they're probably going to fix in the next launch. Or is this something that can actually benefit from the very fast evolution of the underlying models? When video comes out, is your company going to get better or is it going to get Wiped out or very hardcore competitor or a bunch of new competitors because you're filling in a gap.

And I think it's a really important question, even though it's hard to define, you know, I've been trying to figure out exactly how to define it. It's a little bit abstract because we don't know how these things are going to develop, but you can sort of feel it in some ways. When you see when a company is doing something, ed tech company is doing something that sort of like pushes what a model puts out a little further, you know, Claire's out, who of course is becoming a real, real guru in the space talks about there being this sort of like Could you get the same results with a few really good prompts out of one of the main models?

And if you can do it with just a few really good prompts, then you're probably at risk. You're probably sort of in that zone of maybe being threatened by the core capabilities of the models. And I think it's something I would encourage everybody in ed tech to sort of try to put that litmus test against yourself.

And you're seeing more sophisticated players in ed tech. Well, yeah, we met math. We saw Saudi. platform raised 4 million this week called Alguru, a Polish platform about coding raised over 8 million euros. There is still funding happening, especially in the workforce space, but there is still funding happening.

But I think a lot of people are still sort of just trying to make sense of this crazy moment where All of the big tech companies in the world, pretty much all of them. And we didn't even mention places like Alibaba, which is doing this or Elon Musk's company, where he's also trying to get back in the AI game in a big way after leaving open AI.

So like, there's a lot of people trying to do this, to do AI models. And the question is just what exactly is the role of ed tech players in those models? Is it about You know, having amazing data sets that can be used through the anthropic type augmentation generation, where you can say, look, take this tool, but then look at this huge, incredible data set.

And it all fits in your context window. And you can actually make sense of it and use that to tailor the recommendations or to tailor the lesson plans or to tailor the activities that you put out. That's really interesting. And that would evolve and get better and better with time. with the underlying tech.

It's a really exciting moment. And we'd never even talk about things like gaming or audio here, which are also moving very quickly. A lot of music and audio tools coming out and DeepMind just unveiled a new tool that's basically can create world models, like physics based world models through AI. You can imagine that that could change the ability to create simulations, immersives.

Games that are educational. So I think everybody in the space should sort of plan on a future that's coming very quickly, where you can do things that are just almost unbelievable soon. And don't feel limited or feel like, you know, the tiny gaps in the model are business opportunities yet. You got to sort of expect them to get better and better fast.

[00:18:14] Ben Kornell: Yeah. I think there's also some interesting implications for how schools and school systems and university and university systems need to evolve. We, you know, moving to another headline in K 12, the Texas Education Authority, TEA, they ended up announcing a pretty monumental change in how they do curriculum development.

Curriculum development and curriculum procurement is a hot topic in our space always. It's kind of the biggest albatross in, in trying to grow your company. There's these four year to eight year windows when states adopts core curricula. And so this is why for the most part, ed tech startups have played in the supplemental space and not in the core space.

So Texas just announced that they are going to be launching a new process and that process adds 40 per student to buy curriculum and curricular materials. Now here's the catch, that extra 40 is for curricula that's been developed by TEA themselves. So homegrown curriculum and content. This is a major shot at, you know, national content players.

And could have real ramifications for large organizations like Curriculum Associates, Amplify, Houghton Mifflin Hardcore, et cetera. But it does open up a window for rolling approval of core curriculum. So while it is a play, I think primarily to, for people to buy homegrown Texas curriculum, there is a way in which the process now will become an annual process where.

Any outside or third party curriculum could be approved for the state of Texas. And state of Texas is one of the three biggest markets, Florida, Texas, and California. So it could be a huge, huge opening for edtech companies that want to pursue that. Even if you can't win the whole state, now you might have your core curriculum approved by TEA, and you could at least sell to some segment of the state.

The excitement, I think, around this is not just in the legislation itself, but in the potential for this to set a trend in the landscape where states are now moving, would follow Texas lead, especially red states, and Move into more like a regular adoption cycles. What we're hearing from the ed tech insider community is there's fear around quality.

If you're opening this up and you're not having these rigorous approval processes, it's not clear how quality standards are applied. And in some ways new processes can be vague on what the selection criteria is in the first place. That said, we just literally did a segment on how fast AI is changing and how much potential there is.

This notion that curricula doesn't change every four to eight years, you know, like that's the rate of change that has to go out the window. So I'm, you know, cautiously optimistic that this could signal new purchasing behaviors and patterns. And we also know that just because the state changes their adoption cycle, ultimately adoption happens at the school level.

So there will be checks and balances in the system. What do you think about this landscape change? 

[00:21:40] Alex Sarlin: It's 

double edged sword. I mean, I agree with you. I think one of the most exciting things about it is that it allows for more continuous improvement. It allows schools and the state itself to change more rapidly, not be as sort of continuously in lockstep with some of the major publishers, the Curriculum Associates, the SAVAS, the HMH, as they have been in the past.

It opens things up a lot. At the same time, my sense is, especially because this is coming from Texas, which has a sort of a very aggressively political agenda in many ways when it comes to education. It's mentioned in this Ed Week article that, you know, the state board greenlit all the products developed by Texas education officials, including controversial reading lessons with Bible stories.

And you're like, right, that is part of the situation here. So on one hand, it opens up the ability for not only ed tech publishers, but also ed tech curriculum creation companies, which is a burgeoning space, right? A place where a company can sell to Texas, the ability to create really good, engaging, high quality.

Pedagogically sound curriculum, suddenly there's a much stronger reason for them to do that because it's all, there's actually money behind it. That's all exciting. The idea that this is coming from this sort of loaded political space and that there may be a little bit of a race to the localized, you know, I don't know how to put it.

I want to say bottom. It's not necessarily bottom, but like a little bit of a feeling of, Oh, well we can just do it ourselves and we don't trust the federal government and we don't trust that we may or may not trust. companies that are headquartered on the, on the coasts. And so there's a little bit of a sort of an isolationist bent to this of like, we don't need your curriculum anymore.

We're just going to make it ourselves. And we are, we're going to put Bible stories in it if we want. That doesn't make me super happy as an education person. But at the same time, the loosening of regulation makes it more interesting for ed tech companies to play in this space. So I have a similar sort of reaction.

And it seems. Exciting on some level. I think the bureaucracy around curriculum adoption and procurement has been one of the has been terrible for the ed tech space. Let's just put it, you know, bluntly there at the same time. I don't think this is coming from the most noble of purposes. I don't know. What do you think, Ben?

[00:23:54] Ben Kornell: Yeah, I think what we're entering, whether noble or not, I think we're entering a period of deregulation. 

[00:24:02] Alex Sarlin: Yeah, fair. 

[00:24:02] Ben Kornell: And so, I mean, I think the impetus behind a lot of the movements coming out of Texas. Are just really around choice and you, you know, not having the government make the choice for you. What that does is puts a burden on the people at the parental level or at the school level around making good choices and, you know, doing so in a way that's transparent and effective for everyone.

I think a lot of questions that we've gotten is how do the federal elections Education and the answer ultimately is not a lot because the U. S. Department of Ed doesn't have as much of a impact, especially in K 12 as it does in the U. S. In other sectors of the economy, but the trickle down effect is that the politics of deregulation have a lot of momentum and, you know, at the state level, that's where we're going to see these things needed out in terms of higher ed.

You know, I think the question is really around the state systems and Florida seems to be the one that has been moving things the fastest, the furthest on state systems. And that has had really mixed results. The University of Florida recently was deranked as one of the top public universities for student outcomes dipping, and they had a president turnover.

So I think as we move to a place where things are a little bit more deregulated and people are pushing on the different pieces of status quo, we're also going to have a higher degree of thrash and that's not something we've been used to in our space. You know, as innovators, I think we're a little bit more accustomed to that and like willing to lean into that as an opportunity, but I think for parents, families, and communities, that can be really, really challenging.

[00:25:55] Alex Sarlin: And I think it creates yet another potential opportunity for companies to help. Families navigate this complexity of space because I think it's with the increase in choice. I like the way you're putting that Bennett sort of, you know, take any kind of politics out of it. There's probably going to be increased local control and choice.

And if that's true. then that gives a lot of the onus of choosing curriculum, choosing providers, choosing tools, all the way down to the, you know, school educator and home level. We've seen some of that happening with homeschooling and micro schooling and pod schooling over the last few years. We, we, there, a report came out this week that said that public school enrollment has dipped two and a half percent, which is a lot from 2019 to, you know, to 2023 last year.

So we've seen a sort of exodus. It's not huge, but you know, that's 140 kids from public schools and, you know, 45 kids. And I think, you know, where are all those kids going? Well, parents are having to make more choices about their kid's education than they maybe ever had before. And potentially ed tech companies can either sell directly to them or create tools to help them navigate this space, which I think is really interesting.

[00:27:05] Ben Kornell: Yeah. So in terms, as we step back from this and just look at the. broader world of ed tech. I do think that there's a flurry of end of year reports and funding and M and a and financing. One that caught my mind is that emerge, which is a friend of the pod. They're a fund in the UK and Europe. Generally, they raised a 56 million pound fund too.

And we also know that bright eye is in the process of raising another fund. So it does seem that European ed tech. Is alive and well and growing in momentum. And then we also saw some really interesting funding rounds for workforce and higher ed element for 51 had 175 million strategic investment from PSG to accelerate AI adoption in higher ed admissions and student success.

So, you know, pretty big round and sauna labs, which would, I would qualify as workforce learning. So Joel over there is a friend and they raised 55 million. What they do is they use AI to essentially create a repository of all the knowledge within your company and create learning lessons and learning modules.

That take advantage of your company's learning IP and train and level up your staff. What caught your attention? 

[00:28:27] Alex Sarlin: Yeah, no, those are big round. We also saw Speak raise a Series C 78 million funding. Speak is a leader in language learning, basically, and the ability to do conversations across languages.

different languages in a very sophisticated and slick way. And they've sort of been out front in that space for a while. Like 78 million is pretty serious in this, in this day and age. Another one that stood out to me is Stepful. Stepful is a REACH Capital and Y Combinator company or backed company that has been thinking about healthcare training.

So it's a workforce upskilling play, but really specific to the healthcare industry, which is a very hot industry with a lot of shortages and they raised 30, 000. 1 million. They announced it over the last couple of weeks to figure out, you know, exactly how to scale that and go bigger and bigger. I mean, these round sizes and the numbers are more than we thought.

The emerge one is really interesting. We are big fans of the emerge folks, and I think they think very deeply. I highly recommend reading their investment thesis and sort of looking at their portfolio because I think they think really Futuristically, and they do something a little bit unusual in investing in that they try to, they come in early, but they come in with pretty sizable checks early.

That's their, at least, that's what they're trying to do, which means that they have to take some bigger chances, which means they try to have a really clear theses about exactly where the space is going, especially with tech and AI. So I think they're just really interesting investors to keep an eye on.

Yeah. Those are the ones that stood out for me. To me, it looks like there's also one called, um, outsmart, which is X duolingo folks who they just raised 13 million for startup. That's about college access, which is really interesting. Cause you know, we're, we're in a very interesting moment about, you know, sort of shifting from a college for all perspective to a, you know, what exactly do we want for college and how do we, you know, there's just a lot of back and forth about it, but college is still something with a lot of need and a lot of loans and a lot of.

Gaps. So this is a company called outsmart education. I think it's worth keeping an eye on if only because you have some very high level people, including Gina Gotthelf, who is one of the most visible growth folks at Duolingo. And that's, uh, obviously, you know, a big deal. 

[00:30:38] Ben Kornell: Yeah. I mean, overall, Alex, I would just say.

A good sign here is that this is not just your EdTech VCs jumping in here. MountSmart's a good example of this. They have Coastal Ventures, Latitude, Lightspeed Ventures involved. We also saw Tichy with Goodwater Capital as the lead. General Catalyst led the Amigo race. Just, and we mentioned sauna, NEA and Menlo ventures were in a pretty sure Jennifer Carolyn's husband is over at Menlo ventures.

So you can see that like the, the funding ecosystem is actually coming back a little bit to ed tech. I would say the trend is more in the workforce and higher ed learning specifically upskilling. And there is this sense that AI. Is creating this need for more constant re skilling of people. So to me, this is actually a great close to the year.

If any of you read the Oppenheimer report from our friend, Matthew Johnson, it was looking pretty, pretty bad and now looking much better on the MNA front. It feels like we keep predicting this is going to be the year of M& A, and then the results are still not quite popping. But on the M& A side, we did see that there was a lot of activity from Imagine Learning.

They acquired Pango Education and Early Bird. And, you know, a lot of like singles and doubles in terms of the acquisitions. So maybe that's something that will heat up in 2025. We'll have to see in the predictions episode. 

[00:32:13] Alex Sarlin: Agreed. It's an exciting moment. I mean, Jorge, the head of smart learning is the ex chief product officer of Duolingo.

And I think it's one of these moments when you sort of look back at the, at the span, the little rollercoaster that has been ed tech for the last 15 years. That's a great. Person to be, if you want to raise money because you are, you were part of something that was sort of undeniably successful, never crashed, has only continued to really grow and has become pretty much a household name around the world.

It's pretty rare ed tech company to do that. So I think there's an interesting, there may be some sort of second generation. 

[00:32:49] Ben Kornell: There's going to be a duolingo mafia. We know exactly. Well, I guess the question is, are they going to be in education or are they going to. Go into consumer tech, or are they going to go to AI?

Cause depending on which investor you talk to, they'll say it's ed tech. They'll say it's AI. They'll say it's consumer. Even if you look at it as a consumer app, it's still top five, most successful. Consumer app over a 10 year period. So, you know, it kind of defies categorization a 

[00:33:17] Alex Sarlin: hundred percent. It's going to be a really interesting moment.

And I think, you know, one of the things that's been surprising for me all the year is that you've seen people like, you know, Ilya Sutskover from open AI starting, you know, a stall. Something that at least they, that in some ways they call an ed tech company, or you've seen some people play in ed tech or learn LM that we've talked a lot about, you know, Google create a whole model specifically around learning, you know, learning and education is such a natural fit for AI that I think a lot of ways, some of those, you know, do a lingo mafia people or people who are leaving some of the biggest ed tech companies and starting new things.

For one thing, they may be mission driven, I hope they are and that they stay in ed tech for that reason, but it's also like AI and education are like, you know, a pretty nice combination, especially if we can figure out a few, you know, kinks along the way. One last thing I want to bring up really quick before we talk to our guest is this idea.

This was in Claire's newsletter, and I think it's worth pointing out just because it's really, I think it's something we've been sort of waiting for, and we're starting to see it really happen. Now there's a new AI benchmark called AI Luminate. And basically, it scores models based on their ability to avoid generating harmful content across a wide variety of different definitions of harm, and it's a sort of secret what they are, and it's trying to be, as you can imagine, a more free floating, not within one of the models, measure of Safety, basically.

Now, that is something we've desperately needed in education and AI, and I'm excited that at least it's starting to happen. I don't know if education is necessarily their, their main bailiwick of what they're trying to do, but there are lots of places in which harmful content is harmful, including social media, including, you know, Lots of places, therapy, but education is a big one.

So I'm hoping that this is, you know, that some of the fear that we've had about hallucination and bias and just in appropriateness and that, you know, you have to have, if you are an educator right now, if you're a principal, if you're on a school board, like you would be negligent to not think that this is possible.

Hopefully we're starting to get some numbers around it and that will accelerate growth next year as well. 

[00:35:26] Ben Kornell: Yeah. Wonderful. Well, such a great note for us to end on. And now let's head over to our interview. 

[00:35:33] Alex Sarlin: For our deep dive this week in the Week in EdTech, we have Alex Lindley, co CEO and co founder of Capfinity, one of the world leaders in skills based hiring.

Welcome to the podcast, Alex. 

[00:35:45] Alex Linley: Hi, Alex. Great to be with you. 

[00:35:47] Alex Sarlin: Yeah. So first off, for those who may not yet be familiar with Capfinity, tell us the origin of the company, how you came to think about skills as such an important part of hiring and the Capfinity story from beginning till now, just the overview.

[00:36:01] Alex Linley: Yeah, sure. Thank you. So we will celebrate our second decade next year. We started in 2005 and myself and my co founder, Nikki Garche, are both psychologists by background, but from slightly different angles. So Nikki was, and still is, an IO psychologist working with organizations. And I historically, 20 years ago, was an academic and I was doing a lot of into positive psychology.

Both Nikki and I have backgrounds and an interest in positive psychology. And my research particularly was into people's strengths and how to measure strengths, what happens when people use their strengths, how can those strengths be developed? And one of the things that was. Quite a standout and consistent finding was that when people use their strengths, a lot of good things happen.

So they're happier, they're more fulfilled, they're more confident, they have higher esteem, they're more likely to achieve their goals, they're less stressed. It's not exactly world peace, but there was a lot of very positive, quite broad based good things that were outcomes from People acting in ways that use their strengths and so Nikki and I through a series of conversations thought this is really something that would seem to have applications in practice.

And that really was the founding genesis for capital 

[00:37:30] Alex Sarlin: and skills based hiring is something that you hear about a lot. Now, but I don't think it was something you heard about a lot a decade or especially two decades ago. Tell us about, you know, when you first jumped into the talent management talent acquisition field with this strength based approach and this idea of being able to evaluate each individual's strengths and higher based on that rather than traditional signals.

How was it received then? And then how is it received now? 

[00:37:58] Alex Linley: Yeah, sure. So I think definitely What I would say now is skills based hiring has become far more widespread in terms of its interest, its application, the appetite to use it. And I think that has been driven by much more of a recognition of how job roles are evolving.

So probably the simplest way to describe it is once upon a time, a job was a pretty defined, pretty consistent set of skills that would be relatively unchanging. I think there's increasingly a recognition now that can a job even be defined? Or is it better to think of different collections of skills that might need to be used at different times?

And that certainly plays through into a lot of the appetite for skills based hiring. When we started out 20 years ago, I think There was definitely an appetite to look at what people felt was missing in talent acquisition at that point, which was really to do with the question of energy and motivation.

So we were pretty good at working out, can you do something, but we didn't really have a good answer to, would you love to do it? And so one of the key findings with strengths and the way we think about defining strengths and skills is that combination of being able to do it, but also being motivated to do it.

It's that combination of can do and love to. And that certainly resonated with people and got sort of some of our early supporters and early 

[00:39:31] Alex Sarlin: Yeah, you mentioned that jobs have been changing. That's been true for a long time, but it's especially true. Now, people talk about the shrinking half life of, you know, jobs.

So in that world, you've sort of created and designed a system of understanding strengths. That's not directly related to any job role or even any industry, but this sort of 80 skill system. And, you know, these skills are so interesting. This idea of customer champion or customer. bounce back or opportunity spotter.

They're really exciting to even think about these different skills. Tell us how you came up, you know, you're both psychologists. Tell us how you came up with this universe of skills and sort of how it plays out for people when they're assessing their own strengths. 

[00:40:11] Alex Linley: Yeah, thank you. So I think the first thing to say is we draw a distinction between behavioral skills, cognitive skills, and technical skills.

In simple terms, technical skills are what we might all traditionally think of as skills. So can I code in C sharp? Can I build a macro in Excel? Can I speak French? These would be what we typically think of as skills generically, but we would refer to as technical skills. Cognitive skills, Things like numerical ability, critical thinking, the types of things that you might get developed at school.

And then behavioral skills, we think of as being much more on the human end. So we detest the term, but traditionally these might have been referred to as soft skills. So things like, could be collaboration, or listener, or bounce back you mentioned, which is about doing even better when you've had a setback than when you were doing before or customer champion, which is kind of what it says on the tin, being an advocate for customers.

And the way that we identified and ultimately arrived at this set of 80 was simply through observation to begin with. We spent a lot of time in the early years, what we call strength spotting. So literally, as we went about our daily lives, just watching what people were doing and thinking, is there something unique about that?

Is that a particular strength, particular skill that Either has got a name. So listener is, is a word that we would all recognize, but it's the strength or skill of being able to listen really well. Collaboration, we again would recognize, but then there were other things that were slightly more novel. So bounce back is probably recognized, but certainly not as clearly, I think, as we find it in terms of that motivation to use a setback to do even better.

than you were doing before. So using it to bounce back as the, as the name suggests. And those observations were ultimately sifted and validated through looking at what was relevant to life in general, but also the world of work, which was a lot of our focus and to making sure that things were suitably distinct from each other and therefore would lead to different outcomes.

Now, certainly there are. relationships, there are correlations between the different skills, but they also are distinct. So they lend themselves to being bundled in different ways for a particular role or a particular industry. We're always very careful to say that it's not a cookie cutter model where you have to have these things to be successful in this industry.

But if you know nothing else, we can certainly give you a very good place to start. 

[00:42:55] Alex Sarlin: Yeah, they're really amazing to look at relationship deepener adherence. This is an interesting way you love to follow processes operating firmly within the rules. That is not a skill of mine, but something that I have definitely observed many people being very, very, very excellent at in various, and I totally see how that could correspond to certain types of roles in the world or industries.

So when I think about this sort of skills based hiring, this skills based model, it is so exciting. It feels like a completely different and frankly, much better way to think about oneself. It's positive. It is feels much more humane. And I love the distinction that you just made between behavioral, cognitive and technical skills.

This is something I've been wrestling with. We all list our skills on resumes and that's Technical skills, but that is so different than what we mean when we talk about skills based hiring. And so it's been a little bit of a loaded term. My question for you is how might we get this type of approach into the education system so that, you know, younger people, whether they're higher ed or K 12 are starting to think about themselves and their own authenticity or their credibility or their diligence.

I mean, this feels just like a totally different way to envision what education is. 

[00:44:07] Alex Linley: Yes, I think it's a great question. I think I am encouraged because I do significant moves in this direction in certainly higher education. I'm less familiar with K 12 and the high school environment, but certainly in higher ed, I think the skills agenda is very strongly alive and well.

And we are working with a number of university partners with our skills discovery product, which is the assessment tool and the framework for the 80 skills. So they are actually looking at two things. One is looking at how can this serve their students as individuals and how can their students think about what skills they have, what skills they might want to develop.

But secondly, and this is really quite intriguing, using that to be able to take a view as a school, college, university. What does this tell us about our cohort overall? And that's exactly what we see in the use of skills discovery in organizations. What does it tell us about our employee base overall?

And that is particularly relevant if a firm is undergoing transformation, thinking about how they evolve their strategy for the future. It gives them a baseline of where are we starting from at macro level, but also what could we transition to right down to an individual level. So if we needed people to move from customer service to be more technical, who have we got who, who might lend themselves to more of that technical bent.

[00:45:36] Alex Sarlin: It's a really interesting concept of looking at individual skills, but also aggregated skills and a whole cohort, whether it's in a university or a company or a school or an afterschool program. I mean, you know, almost any context sort of understanding the makeup of the people in it through this lens would be really informative.

So it's really a vision of. What education could be that gets me incredibly excited. And I think one thing that you've thought about for a long time and is also sort of taken fire over the last few years is the concept of, you know, when we think about skills based hiring, it's an alternative to credential based hiring, an alternative to the type of hiring we've traditionally done where you say, Okay.

Oh, I look at someone's resume and I'm looking at their experience and I'm looking at their education. And if they have a degree from the right kind of place in my mind, and if they have experience at a place that has some kind of similarity to what we're doing, then maybe they're the right hire. And those are such weak signals in many ways.

And it's caused such a strange, they're weak signals and they're very hard to get. They're very expensive and very difficult, especially for career starters to get education and experience and being able to talk about them. This is such a different way to manage careers. Tell us about that alternative credentialing concept or the idea of what might the world look like if people truly cared more about skills than traditional credentialing signals.

[00:46:56] Alex Linley: I think the first thing to recognize is that the credentialing signals that you, you talk about in terms of experience, in terms of educational qualification, expensive to get for the individual, but they're very cheap to review for the employer. Great. So part of the challenge is it costs the employer nothing to look at where you went to college and what job you worked in previously.

Whereas an assessment, even if it's very low in the dollar stakes, is still a cost. Now the distinction, and certainly what we've seen in the UK over the last decade, Is quite a significant shift in major employers to move away from pure credential based hiring and indeed credential based screening to really trying to remove those credentialing barriers to people and giving much wider access and wider opportunity to people to be able to demonstrate their skills through the assessment.

And that's proven to be extremely effective at broadening talent polls. Filling harder to fill roles that, that you always thought it was difficult because you couldn't find people with these particular criteria, but it becomes a lot more possible with skills based hiring, I think as we go forward and as we see the half life of jobs reduce ever more and jobs become more bundles of skills that might be deployed in different ways as distinct from sort of the longevity of jobs that we've seen in the past, I think that's going to become more and more prevalent.

And. While I don't think that education will, will change overnight, I don't think there's anything wrong with earning a degree and having that as a marker, I do think that it will become one of multiple pathways, one of multiple ways to be able to demonstrate the capabilities that you have. 

[00:48:48] Alex Sarlin: I couldn't agree more.

It's an exciting vision. I think we're finally maybe getting there. It's so it's amazing. Capfinity has built so many different ways to assess ways to strengthen skills to train to, and you know, you have, you have video interviews, you have job simulations, your virtual reality of mobile assessments you've had, you know, because you've been working in the skills based hiring space for much longer than most people, you've had a chance to really expand.

My question for you is, Everybody just got a little bit of a jet pack with AI to be able to create new content, create new simulations or experiences. How do you hope to use AI to expand Cap Infinity's offerings even further? 

[00:49:31] Alex Linley: Yeah, thanks. I think you've already touched on one of the areas where AI can be a great support to idea generation and content generation.

We're certainly already using it in that way to turbocharge some of the work that we're doing there. We're also seeing really interesting opportunities where clients are interested in how can we assess for what I might generically call AI skills. So who has the skills to be able to use AI in the most appropriate way.

And that's really quite intriguing because AI is evolving quite quickly. The way that you might use it and work with it is evolving quite quickly. But guess what? Behavioral skills like curiosity, adaptability, technology focus, will consistently help you to think about and use that AI in the, in the right way.

The other area that we're seeing a lot of interest is In what you might consider is people misusing AI to misrepresent themselves through a recruitment process. And so we are, have been doing a lot of work over the last number of years to mitigate that, to detect that, to deter that, to ultimately to prevent that.

And that's certainly an area that I, I see as, um, as growth as we move forward. See more and more emphasis on the integrity of responses and the integrity of candidates and applicants. And that's where AI very much has a place as well. So to some extent, there is both an opportunity on the integrity side, but also an opportunity on The evolution of new AI skills that we're very much involved with as well.

[00:51:11] Alex Sarlin: Fantastic points. Sometimes I think we see, you know, situations in education or hiring as like arms races, right? It's like the applicants or the students have a set of tools that they're getting better and better that they can use to sort of get a leg up. And then the instructors or the hiring managers have a set of tools that they could use to get a leg up.

And I mean, I've seen that a lot. But I think that one of the things I really love about what you're doing at Capfinity is I think there's a chance to, even in an integrity setting, sort of turn that on its head and say, no, we are looking for the same thing. You know, we're looking for people who are great incubators or who have great look at your time optimizers.

That one seems related to. I write if people who are great time optimizers, always looking for ways to get something done more quickly, more efficiently and find faster ways. Well, that's probably a great set of people to do an AI training about AI productivity and supercharge the whole org. And there's, there's a way to sort of align organizational needs and individual needs around these skills, which takes us out of that sort of, you know, arms race mentality.

I'm really excited. So. What are your thoughts about the future of skills based hiring? Is it going to continue to explode in the way it is? And what will that look like? 

[00:52:20] Alex Linley: I think if we look at the trends that are driving the labor market broadly, I, I do think all the indications are that skills based hiring is going to continue on this trajectory.

And the, the two reasons for that are there is always going to be the so called war for talent, and there will always be roles that are harder to fill. I think. Those roles that are harder to fill become easier to fill if we have a more flexible, more granular understanding of the skills that we need.

And secondly, an AI will have a big part to play as a driver of this. Um, as jobs change at an ever increasing rate and the meaning of a job becomes quite different to what it was when we were kids and growing up, jobs will effectively be rethought of as bundles of skills that can be deployed in different ways.

And I think as soon as you start to move to that model, Skills based hiring becomes even more relevant because it doesn't make sense to recruit a, let's say, a customer service agent if, if actually what you need are a combination of skills in empathy, customer champion, and listener that may be deployed in these different situations as things evolve.

[00:53:41] Alex Sarlin: Amazing vision. This is Alex Linley, CEO and co founder of Capfinity, moving the bar for a long time in skills based hiring. And I think, you know, moving it in exactly the right direction. Thank you so much for being with us here on EdTech Insiders. 

[00:53:56] Alex Linley: Pleasure, Alex. Thanks for having me. 

[00:53:58] Ben Kornell: Hi, everybody. We are so excited to be joined by Henrik Appert.

He's the founder and CEO of MagmaMath, an internationally recognized and award winning math learning program. He's an appreciated speaker, writer, and panelist on topics of education, innovation, and entrepreneurship, featured in top tier publications and on national news outlets. Most recently, Magma just announced its 40 million Series A round, wow, with five elms capital and was awarded the Bill and Melinda Gates Foundation grant to advance math teaching for underserved communities.

Henrik, so glad to have you on 

[00:54:34] Henrik Appert: EdTech Insiders. Welcome. Thank you so much. It's great speaking to you again, Ben, and happy to be with you here as well, Alex. 

[00:54:42] Ben Kornell: So let's just start with Magma's origin story. What inspired you to create MagmaMath and what differentiates the platform in solving challenges in math education?

[00:54:51] Henrik Appert: So it really came from a identifying a huge need in math education. I'm from Sweden. So just saying that from the get go, if my accent is a bit weird, you know where that comes from. And in Sweden, one of the main data points that we look at are PISA surveys. So PISA is this international way of measuring academic results in the OECD countries.

And it showed that math results are really tanking, unfortunately. The thing is with math is that it's the single best predictor for future academic and professional success. So if we're serious about giving all students the start in life that they have the right to, and we believe that they're capable of, we need to get math education done right.

So we got really curious about, well, like, well, how is math being taught and what's working and what isn't working? And a couple of things stood out to us. And so one thing is that whenever we're learning something new in math, we base that new knowledge on our current understanding of the subject at hand.

Which means that in order for math lessons to be meaningful and efficient, teachers need to understand, well, how are my students currently thinking about and understanding the math? Now, that's very easy to say and much harder to achieve when you have 20, 30 kids in the classroom that are scattered over a pretty broad spectrum of current thinking.

And at the same time, we realize that, you know, during a math lesson, there's actually so much data that is arising around how they're thinking. But the problem is that all of this data is kind of stuck in the student workbooks. And teachers, they don't have the time to go around and collect it all and grade it manually and then summarize those like data points and then draw some conclusions.

So we just thought like, well, surely there has to be a much better way where we can capture that data and actually give valuable time back so that teachers can spend more of their time in like value adding activities together with their students. That was kind of like the starting point, like that need and that realization that the data is there.

How can we just make it more accessible and actionable to teachers? 

[00:56:47] Alex Sarlin: I mean, it's a really important insight and I think one that speaks to the heart of math pedagogy, which is really fantastic, and you're taking a very AI centered approach to helping metacognition and sort of surface student thinking.

But when I think of magma math, I'm like, wow, it's an AI approach to math. It's probably brand new. But you actually started in 2015 and you, you're still at the forefront of AI in education. Tell us about that journey and how you evolved alongside with the tech. 

[00:57:15] Henrik Appert: Yeah, so. Our company has Always been built with the realization that AI is going to have a dramatic impact on the future of like really any industry and especially the education space.

I was actually prior to running Magma, I founded a different company also in the AI space. So, you know, I was on OpenAI's website for the first time, I think in 2014 or something like that. Now, I should say that, like, so AI is very much like integrated in our solution, but we don't believe that. AI, there's no like intrinsic value just in using AI.

At the end of the day, it has to do with like, how are we actually solving real problems for our users, for our customers? And then if AI happens to be the technology that is best suited to solve that problem, then that's the tool we should be using. So we've always, you know, believed in the powers of AI, but The key in the beginning is to actually understand how are teachers and students interacting in meaningful ways in math lessons, and then how can technology enhance those things.

I'll say one more thing on that AI piece, which I think maybe sets us apart. So I think that, you know, if you're living in a world of AI, it all comes down to how much structured data do you have? And what I mean by structured data is not just like, you know, a bunch of data points, but you actually know what type of data this is, Who's the student?

What problem is it? How have they solved previous problems? And so what sets Magma apart from other solutions is that we're not just capturing right and wrong answers in Magma, but we're actually capturing every single pen stroke of how they are arriving at the answer. So this means that we can actually build much more deeper understanding of students thought of how they're arriving at the answer.

And then of course, like when OpenAI released Chat GPT and just, you know, completely redrew the landscape. We were like ready for that because we had the data, we had it structured, and we could, you know, just keep on building on what we've already been laying the foundation for. 

[00:59:16] Ben Kornell: Well, and you have it longitudinally too, so you can see.

It's not just even structuring, like they write wrong versus like why, but then also seeing how does that progression proceed over time. And so once you have enough of that, you know, progress data, then you can have the predictive model. You know, we just were talking earlier in our new segment about the duolingo effects in education and this idea of creating these learning graphs.

that predict, you know, based on a student's misunderstanding or the reason they got this wrong, what do they need to do next? And I think that it's a fascinating. Point, which dovetails to the fact that it is an incredible asset to have been around for 10 years, because if you were starting today, you wouldn't have any of that data.

[01:00:00] Henrik Appert: Exactly. So another point that you're really making there, Ben, is you're talking about that graph. I think it's so interesting if we potentially can start seeing that, you know, students who think about math in a certain way, have more success. later on in a perhaps related but not directly the same topic.

And maybe we can then actually identify, Oh, well, it was when that student unlocked a certain thing that they got those things. Then we can make that insight we can make available both for teachers, but also for students. So I think it's really like, well, how can students become much more of a learning resource to one another, not just, you know, in their direct proximity, like in their classroom, which I also think, by the way, there's a huge.

value for students to be in the same classroom, to be talking math with one another, to be debating, to be hearing others talk about it. But I also think that there's a possibility for students to be benefiting from the work of others, whilst of course still protecting students privacy. There is no conflict of interest there.

[01:00:59] Alex Sarlin: So one thing that magma is doing that is incredibly interesting and I think really speaks to a whole movement now in the tech and AI space is really thinking about the thinking behind the thinking what we call, you know, metacognition and what the magma product actually, as you say, it records every pen stroke.

It's not just the answer that's coming out, but all of the work, all of the thinking, all the mistakes that have happened on the way. Right. So that's a really interesting approach. It's one that, you know, math teachers have tried to do for generations. You show your work, right? Please, you know, show all your calculations.

But now we have a whole new way to do it. Tell us about your thinking about metacognition and why it's important to capture all of the data inside a student's head and everything they're doing rather than just answers. 

[01:01:43] Henrik Appert: So, to me, really, like, math is so much more about how you arrive at the answer than what the correct answer is.

I remember being seven years old and doing math with my dad, and, you know, I had the privilege of having a dad who was pretty comfortable with mathematics, and I know that all students don't have that privilege, but I had the privilege of having that, and I was seven years old, I was doing math at home, and I was so proud that I got the answer right, but he said, well, you know, Henrik, this is all great.

But you have to show me how you arrived at the answer. And this is really true for whatever kind of like level of mathematics that we're doing. I think it's also very important that students get into the habit of showing your work, even when it is very basic mathematics that students are working with.

Even when you're actually able to do things in your head. And the reason why I think it's very important is that the science of learning tells us that we're only able to hold so many things in our head at the same time. And so we need to, you know, we have the working memory and we have our like, you know, long term memory, if you will.

And if we're able to only hold so many things in our working memory at the same time, we need to have the habit of like, you know, having this like milestones, if you will, in our solutions. And I think that when we are showing our work, that exactly what we're doing, we're able to like break the problem down into different parts.

And we can like revisit a piece of information and Collect it once we're like ready for it at another part of our solution. And I think unfortunately what happens for a lot of students is they don't realize the value of showing their work that like at the end of the day, the person that you're showing it for the most is yourself.

And I think that that also like when students are showing their work, it means that. Suddenly it can also be discussed with others. Now, I think another challenge that has existed previously is that, you know, we are motivated by what our teacher wants us to do, right. But only to a certain extent, and only if it's actually being celebrated.

And I think a challenge is if teachers are saying things like, you know, explain your reasoning, show me your work, but we never take the time to actually look at it or discuss it. Then it's kind of like, go dig a hole, but what are we supposed to do with a hole? Like if we're never going to do anything with it, it doesn't feel very motivating.

So we're trying to create like, how do we make it easier for teachers to have really, really good math lessons where students can not just like, you know, show their work, but they can also talk about their work and they could talk about their work in a safe way. They can share it anonymously and this creates much more.

You know, you're creating a why for students because students work are now being projected to the front of the classroom. They can see it. They can compare it and contrast it with other students work. Teachers can support that, like, you know, making connections between different types of strategies. So I think it's both.

enabling them to learn better, but it's also helping teachers to actually teach better. And at the end of the day, like so much of the learning happens, not when we're solving problems, but when we're thinking about it and when we're talking about it and when we're hearing other students reason. So I think it's just like, you know, there's no intrinsic value in the showing of the work.

It's more that like, that is resulting in students actually like internalizing the knowledge. And at the end of the day, that's what we want. 

[01:04:48] Ben Kornell: Yeah, I think to your point, these are challenges that have been held by teachers for decades, if not the entirety of pedagogical time, and this misalignment between student just wanting to get the answer right and move on, and teacher wanting to create cognitive friction so that you're learning metacognition is really that space where the answer lies.

And yet at scale, it's very, very hard to do this. You know, I was a middle school teacher, so I'd have like 150 students. So how can I foster deep metacognitive discussions? How can I look at the data and understand what are the patterns of breakdown in my classroom? Finally, the technology really has intersected the moment.

[01:05:34] Henrik Appert: Yeah, I think it's very exciting. I just want to say that at the end of the day, Like teachers are still, you know, in the driver's seat and they're the ones that, you know, need to put these practices into place. But always when we partner with, with districts, we want to understand like, well, what does good math education look like for them?

Now, fortunately, like this is where they're, they want to be going. So if we can make it easier for them, if we can lower that threshold, then we think that we can get more teachers and students having these meaningful discussions. 

[01:06:02] Ben Kornell: Yeah, I mean, I believe that math is the great first adopter of this because they're used to computational technology, just interrupting their space when calculators were introduced.

And that was probably the first time in the U S at least where teachers were all trained to say, show us your work as a contrast to calculators. So in some ways, there's a way in which that pedagogical profession is already aligned to this show your work, where I get excited is where we have magma English and we have magma history and we have magma all the others.

So, you know, whenever you're hiring for those product leader roles, do let us know. 

[01:06:40] Henrik Appert: I think I'll leave it out there for you. Like, if there are any entrepreneurs out there listening to this, like we're not going to go into those spaces anytime soon. I think if you're a startup, you need to know specifically what problem is it that you're trying to solve?

We are solely focused on math. We want to build the best solution in the world for supporting great math lessons. So there's plenty of like blue ocean out there for ambitious entrepreneurs out there to go after those subjects. 

[01:07:06] Ben Kornell: Yeah, this is where I think the, you know, when chat GPT debuted, English teachers had the first taste of what it must've felt like when the calculator came out, because now you have to be metacognitive around how you do your writing and reading.

So coming back to like actually stepping back and that was a great advice to founders. You know, you've done a couple really challenging things. One, you launched a company in Sweden and you've managed to cross the pond and introduce it successfully in the U S two, you've raised a funding round and three, you've managed to build something.

That not only endured technological disruption, but thrives. Can you tell us a little bit about what that journey from Sweden to here was like and any advice you'd have for entrepreneurs trying to navigate an entry to the U S market, and also any advice on how you navigated the funding market, given that venture capital has been down on ed tech here in the last two years, 

[01:08:06] Henrik Appert: Sure.

Yeah. So there's a couple of questions in there, so I'll try to remember them, but you know, if I miss something, I can only keep so many things, you know, in my working memory. So, I mean, I think there's a couple of reasons to why we've gotten to the place. We are today from the get go. I want to say like, yes, we've come some ways, but I still think that we have the majority of the work cut out for ourselves ahead of us.

We're still like not even serving 1 percent of the world's population. So we, we remain very ambitious, but it's true. We have, we have made some progress in the last decade or so in Sweden. We serve. We're used in more or less every single school district. We've shown a reduced fail rate on national standardized exams by 25 percent.

There came out a TIMS survey now recently that showed that Swedish students have significantly increased their abilities in mathematics since 2019, which pretty nicely correlates with how we've made headway there. And I think, so some of the reasons to our success has been like, first of all, just like an amazing team starts out with, you know, I have a, uh, an amazing co founder, Arvid Gilliam, you know, we've just been in the trenches together.

We haven't been, you know, too fancy or too good to do any work. It's all about rolling up your sleeves and doing what needs to be done, whatever it is that needs to be done. I think. We've had a couple of things that have been really, really important to us. And one has been to work incredibly closely to the district, to the teachers, to the students.

We visit classrooms every single week. And it's not just people in customer success. It's not just people in product. It's not just people in marketing. It's not just people in, in sales, like our entire organization. It's constantly visiting classrooms to see magma in action, both to understand like what's working and so that they can talk about it, but also like, what's not working and what's broken and what should be fixed.

So I think like staying really, really closely to customers and to users has been incredibly important. And we, and like, that's something that it's one of our like core values that we put teachers first. We think that teachers are this like incredibly underserved community and they do so much for society.

So if we can do something for them and they can have more success, that's at the end of the day, also going to end up like helping us be successful as well. So I think another thing where a lot of unfortunately ed tech companies struggle is in sales. I think that, and you know, there's a long reason for this, like education.

I think personally, I think it's very underfunded in contrast to like, you know, the amount of, of return that you can get into a great education system. You have the individual's entire life to kind of like recuperate that investment in forms of like higher salary. More skilled workforce, more taxes instead of people, you know, ending up needing society to provide for them.

If they become unemployed, if they, you know, there's often a correlation between that and like, if you have different types of like health issues and things like that. So I just think that it makes a ton of sense to invest in, in, in education, but I'm sure the people listening to this, I'm really like preaching to the choir here.

So like that does create a challenge for ed tech companies. That like money can be tight. And at the same time, you need to really prove that what you do are, is creating value for teachers, for students, for district administrators, whoever your like market is. I think we've always like really valued sales in our company.

And I think like when you're working with sales, it's important to like, first of all, like you need to show that your product is creating value, but you also like need to create like a partnership with the district. And say like, Hey, like this is a problem for you. And we think that we have a solution, something that can, can help you.

And we're going to be in the trenches with you. And like, we want to invest our time. And so it's important that you invest some of your dollars with us to also prove to the market that like, this is actually a problem that's worth solving. And I think that kind of like those two things, like working really closely with customers and users and understanding how the product is being used, and then also like being very business oriented and ensuring that that you can sell the product that is kind of in us, Resulted in that, you know, our revenues continue to grow very, very few, hardly any customers ever leave the platform.

So we get great retention. That's also in combination with us, like working really closely, that teachers feel really supported. And so that just gives us good business metrics. So when we speak to investors, they can see that. You know, this is really a business that is doing well. And I think if you're able to build a thriving business in the education space, it's a very exciting business opportunity.

Uh, you know, uh, education is going to be here 30 years from now. I'm certain of it. It's a very stable, non cyclical. Industry admits that like there, there are long sales cycles, so that's something that you need to like take into consideration, but it's a great business if you can get over the hump. And then I just want to say like, I mean, we haven't just picked any partner out there.

We feel so fortunate to, to partner with Five Elms Capital. They're previously investors in Aptogee, which is one of the companies that we've looked. Closely at how they've gone to market. And we were in touch with one another for over two years time, you know, they saw me move over to the U S they saw our very like fumbling first couple of steps in the U S and then they kind of saw things starting to pick up.

We had a discussion about a year ago and it. Wasn't quite the right time. And, and now we felt that it was. And so we were happy to kind of like enter into this, this partnership with them. And we're excited for, for the future because at the end of the day, it just means that we can spend more time and more resources in serving our, our students and our teachers and our districts.

[01:13:50] Ben Kornell: And I think, you know, ultimately, customer acquisition is the hardest thing. And, you know, I think just drawing on your previous answer to spending time with schools and understanding what's the problem you're solving? What are the needs in the classroom? What are the needs of administration and your districts?

And how do you support them? There is this effect too, I think of entrepreneurs who've successfully navigated this, helping others understand the growth path into the U S it's something we continue to see. And especially we see UK, Australia, Canada, and U S companies. Flowing freely, but now we're seeing many, many more European ed tech companies find traction here in the U S because they've kind of developed coherent collective strategies.

Alex, anything that you want to comment on before we, we take us out? 

[01:14:44] Alex Sarlin: I have a comment and a quick question for you, a little bit of a hot take question. So the comment is, that was an incredible answer and congratulations on saturating and really moving outcomes in. Sweden as well. I'm sure that that is also an incredible proof point for people who are, you know, who are looking at magma and thinking about what you're doing.

You say, well, every, every school district in the country we're from uses us and they've seen incredible growth internationally. That is, that's incredible. And I think that's just what a legacy to be able to leave. And then of course, export around the world. The question is, I know your second U. S. office, I think you have an office in New York and the office in Dallas.

Is that right? Yeah. 

[01:15:22] Henrik Appert: So Texas is one of our like states that are going incredibly well, but we're in over 30 states in the U. S. So it's really being picked up everywhere. We have a team in Dallas and we have a team in Austin. But I should also say for like, if we have more international listeners out there, that we actually have an office in London as well.

So the UK market, Scotland, England is going incredibly well as well. And then in our case, we have our tech in Warsaw, Poland. Where we have like an amazing team. And I really should like, you know, on that, I really want to emphasize like, this is really the work, not just of me and my, my amazing co founder, but it's just like an amazing team from like, you know, the first people that hired the joined the team until like the latest people.

So just like getting to work with a great team just makes it so much fun. 

[01:16:07] Ben Kornell: This has been such a great conversation with Henrik, CEO of Magma Math. Thanks so much for joining us today. If people want to find out more, where can they go? 

[01:16:16] Henrik Appert: They can go to magma math. com. They can go to magma math. co. uk or magma.

se if you're in Sweden. Wonderful. 

[01:16:24] Ben Kornell: Thanks so much for joining us. And if it happens in EdTech, you'll hear about it here on EdTech Insider. Thanks for having me. 

[01:16:31] Alex 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.

People on this episode

Head Shot

Alex Sarlin

Host
Head Shot

Ben Kornell

Host