.png)
Edtech Insiders
Edtech Insiders
Week in Edtech 7/9/25: OpenAI’s $28M K-12 Push, Meta’s Talent War, Higher Ground Bankruptcy, Teaching Lab + Relay Merger, & More! Feat. Josh Reibel, Dreamscape Learn; Dr. Jennifer Cruz, Pendergast Elementary; & Thomas Thompson & Thomas Hummel, Eduaide.ai
Join hosts Alex Sarlin and Ben Kornell as they explore the latest developments in education technology, from AI showdowns to immersive learning pilots and funding updates.
✨ Episode Highlights:
[00:02:00] OpenAI announces $28M K-12 AI training initiative with AFT, UFT, and Microsoft
[00:04:34] Gallup-Walton poll reveals 68% of teachers lack AI training despite high usage
[00:07:11] OpenAI takes a bottom-up approach as teachers criticize Google’s AI rollout
[00:23:41] Meta forms Superintelligence Labs, aggressively hiring AI talent across the industry
[00:36:17] Senate blocks AI regulation ban, states retain power to legislate AI in education
[00:37:31] Higher Ground Education files for Chapter 11 bankruptcy, prompting edtech funding debate
[00:42:05] Honor Education raises $38M to enhance asynchronous social learning
[00:42:45] Teaching Lab merges with Relay Graduate School of Education to scale AI-aligned PD
Plus, special guests:
[00:44:17] Josh Reibel, CEO of Dreamscape Learn and Dr. Jennifer Cruz, Superintendent of the Pendergast Elementary School District discuss VR and immersive learning with Dreamscape Learn
[01:46:15] Eduaide.Ai founders Thomas Thompson and Thomas Hummel on empowering teacherpreneurs with AI
😎 Stay updated with Edtech Insiders!
- Follow our Podcast on:
- Sign up for the Edtech Insiders newsletter.
- Follow Edtech Insiders on LinkedIn!
🎉 Presenting Sponsor/s:
This season of Edtech Insiders is brought to you by Starbridge. Every year, K-12 districts and higher ed institutions spend over half a trillion dollars—but most sales teams miss the signals. Starbridge tracks early signs like board minutes, budget drafts, and strategic plans, then helps you turn them into personalized outreach—fast. Win the deal before it hits the RFP stage. That’s how top edtech teams stay ahead.
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] Ben Kornell: I'm not sure that the people who are in this deal necessarily know what the end game objective here is, but I would guess they have a series of objectives and one objective is if AI gets banned in schools, that's bad for us and let's partner with the teacher's union who, let's face it, have the most power in the system today to make or break a technology happening or not happening.
[00:00:29] Alex Sarlin: I think in this particular case, this is just a wacky. Moment because Google is the incumbent here. And Google I think likes to see itself as the newbie, the fast moving, and I think it is, it's still in many ways, even though it's a huge company, but in this particular moment, clearly open AI and philanthropic are coming in saying, like you said, we're gonna do something for teachers.
We're gonna be like. Totally bottoms up. And oh, Google Classroom might be giving you these worksheets and you might get insulted by them, but we're gonna work with you to teach you how to do it, and that's a pretty smart play.
Welcome to EdTech Insiders, the top podcast covering the education technology industry funding rounds to impact to AI developments across early childhood K 12 higher ed and work. You'll find it all here at
[00:01:17] Ben Kornell: EdTech Insiders. Remember to subscribe to the pod, check out our newsletter and offer 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.
Hello, EdTech Insider listeners. It's another week in Ad Tech. I'm Ben Cornell alongside Alex Arlin, and we are excited to jump in to all that's happening in education, ed tech, ai, you name it. Before we do, Alex, what do we have coming up on the pod?
[00:02:00] Alex Sarlin: Yeah. Well, in today's episode, we have some fantastic guests.
We talked to Josh Reibel from Dreamscape Learn, and Dr. Jennifer Cruz from Pendergast Elementary, the superintendent of Pendergast Elementary in Phoenix, talking about using VR in the classroom that's in this episode, as well as the two Thomases, Thomas Thompson and Thomas Hummel, the co-founders of edu aid.ai about their teacherpreneur, and they're really trying to embed instructional design and cognitive science into their AI tooling.
Really cool. And then on the pod coming up, we have Sam Whitaker from Study Fetch, which is really a very interesting AI student suite that does a lot of interesting stuff. We have Tammy windup, the new CEO of Securely, and Vijay Kar coming up soon, who is from Photon vr, a big Indian VR and AI company that is coming to the States.
Always interesting things happening in the EdTech space. So, Ben, where's your mind? This week it feels like we saw Google go crazy last week. This streak. It's open ai.
[00:02:56] Ben Kornell: This is the sequel. Yeah, it's the Empire Strikes Back. Yeah, exactly. And in the Star Wars trilogy, I think Empire Strikes Back is the best one.
So maybe this will be the best one. OpenAI Lobbed a huge salvo back at Google with their announcement of their K 12 AI initiative, and they're taking a pretty different text. They're jumping in with the American Federation of Teachers, Randy Weingarten's Group. And going with an AI through teachers bottoms up kind of lens.
You know, I was reading a little bit more about this, and they're actually gonna build a physical location out in Manhattan that will be a national AI training hub for educators. And the deal was disclosed to be $28 million. So my assumption is that OpenAI is pouring $28 million into this initiative, both through the funding of the fiscal location in Manhattan, but also the work with schools.
When you saw this, uh, you know, of course we're friends with Leah and Sia and Lane and Chris Lehan and a lot of the folks over there at OpenAI, and they've long looked at education as this great opportunity. When we had Sam Altman on the pot, he was talking about a tutor for everyone. But over the last two years, they've been so engaged in higher ed.
Right. And in the ed tech space and workforce to a degree as well. Workforce learning that this kind of came as a surprise to me. I had thought they're seeding K 12 to Google. This very much changes the narrative. What do you think?
[00:04:34] Alex Sarlin: Well, let's give a couple of the details of what they're doing first, and then let's jump into analyzing it.
They're basically working with the A FT and the UFT, so two big teachers unions basically to provide, they're saying 400,000 educators in the US with access to open AI's educational tools. That's Open AI Academy and giving them API credits this national hub in New York City, like you said, and free workshops, curriculum, technical support.
They're doing this in conjunction with not only those teachers unions, but Microsoft and Anthropic. That's interesting. I was surprised to see Anthropic sort of tucked in there because Anthropic and open AI have somewhat different lanes in this way, so 400,000 teachers is one of every 10 teachers in the United States.
[00:05:15] Ben Kornell: Well, and and they also said that their free AI training curriculum is for all 1.8 million members of a FT. So they've got dedicated outreach to 400 and free resources for one point.
[00:05:27] Alex Sarlin: Exactly. So like you said, it's this bottoms up approach, and I think it's a approach in response to what they've been seeing in the K 12 space, which is interesting.
We're gonna talk about a Gallup poll that came out this week that basically said that teachers are really enthused about ai, right? I mean, 60% of them are using it based on this new Gallup poll with from the Walton Family Foundation In Gallup, nearly two thirds of teachers have used it weekly. Users are saving about six hours a week.
About 40% of educators are using it once a month, and only 30% you could consider a high or low are opposed to it. So without getting into all the numbers here, I think they talked to a lot of teachers. They looked around at the space and they said, look, teachers want to be using AI tools. They see the potential there, but they're also feeling like they're not getting training.
The thing that jumped out from this poll is that 68% of teachers surveyed said they did not receive training. On how to use their AI tools. And I think that's the key stat here. When you think about what open AI and philanthropic are doing, they're saying, look, it's not happening fast enough. There's not the school policies, the professional development organizations, it's not happening fast enough to keep up with the tech.
Their teachers are not getting training to be able to do the, I think, incredible things that you can do with AI and teaching. So they're jumping into that vacuum and saying, we're not gonna wait on regular PD providers. We're not gonna hope that there's sort of a grassroots approach. We're gonna jump right into this, make a huge amount of free resources, and then get them into educators hands with as much.
And they're giving, you said $28 million. The announcement was $10 million over the next five years from open ai. But I think that's probably a fraction of what it is. So their bet here clearly is the way to get K 12 education, to embrace AI is to work directly with educators. And it's probably a really smart bet.
It makes sense to me.
[00:07:11] Ben Kornell: Yeah, this is a really interesting one on multiple levels. On Tuesday, July 8th, there was a press release, and the press release has different things than what OpenAI said on social media and so on. Interesting. In the press release, at the top, you have a FT, Microsoft, OpenAI, anthropic, and UFT, all organizations with their press release.
Contacts. So the number of parties that had to come together and get on the same page is massive. In the press release it says 23 million and other places it says 10 million. Yeah. Some places it says 400,000 other places it says 1.8 million. Yeah. The reality is this is a big lift. So one, but compared to open ai, is this OpenAI?
Is this OpenAI or is this Microsoft actually making a salvo against Google? 'cause Microsoft is enterprise and Microsoft actually has a lot of Kate's well of skin in the game.
[00:08:10] Alex Sarlin: Yeah. These aren't big numbers for OpenAI. I mean, they have crazy amounts of money to spend on. Great.
[00:08:15] Ben Kornell: Great. So maybe back to what we always say, it's a rounding error for Google.
It's a rounding error for open ai. It's, but what they're clearly seeing is there's a user acquisition opportunity and for open AI to acquire users going with a more of a consumer approach rather than enterprise. You know, rather than going with Google Workspace, which is district purchased and your CTO or IT professionals need to sign on and all this stuff, which is more of a top down Google move, open AI is kind of coming in a bottoms up.
Let's get it in the hands of teachers. And then the third layer of this is for years we've been saying. Teachers' unions are standing in the way of innovation. Yeah. How can they be agents of change? I have innate pessimism about teachers' union's ability to drive innovation. Having been a teacher's union member myself, many times I've been on strike and everything.
But just seeing how slow the union has to go, because you have to keep all of your members kind of in lock step. But maybe this is a time where Randy Weingarten's like, look, if we aren't part of changing it Yep. It's going to go against us. Let's dive in. So I think it is just a fascinating, fascinating story.
Yep. And I think just like the Star Wars trilogy grew into this whole epic franchise. It's clear that this press release is not the end of the story. It's just the beginning.
[00:09:41] Alex Sarlin: Yeah. That stood out to me too, though. In fact, you have Randy Weingarten and Michael Mulgrew, the president of the UFT, both part of the announcement.
It's not just that they're sort of in it, like they're sort of quiet sign on. They are actively championing and, and going outta their way and saying, AI is coming. AI is gonna change what education looks like. It's gonna change how we all think and how we learn and what we teach. They're saying it full throated, which I would not have expected, especially a couple of years ago.
I think when I first hit the space, there was some caution. I mean, I, on this podcast, I'm often get very paranoid about various forces coming together to support the status quo in education. I just feel like it happens again and again. Over the history of education, AI seems to have sort of broken the event horizon on this, right?
I don't think. Anybody at this point is like, AI is just not gonna be a thing. It's just gonna be a small fad. Three years from now, nobody's gonna be talking about it again. Like I don't think that is there anymore. That doesn't mean there won't be lots of pushback. That doesn't mean there won't be lots of hype cycles and up and down and situations, but it does feel like there's a consensus now that AI is not only changing education, it's changing society itself, and that if education doesn't reflect those changes in society, it's actually going to do a disservice to its professionals, to the teachers and to the students it's teaching.
It feels like there's now something of a consensus there, and frankly, I'm thrilled to hear that that's a consensus. I've been worried about the cliff, the hype cycle, going all the way to the trough and sort of falling out with this because. Before we even get to the potential. So I'm excited about it. I mean, a question for you, Ben.
I, I feel like your top down bottoms up analysis of this is really spot on. When we saw Google Classroom do those 50 announcements last week and just basically say Google Classroom is now a suite of AI tools, among other things that we're incorporating everything into us, into what we're doing. I said just last week that Google is sort of trying to draw a line saying we are already baked into K 12 learning.
We are there, Chromebooks are there, classroom is there like stay out. It's not totally clear. Obviously this training is going to train on open AI tools and anthropic tools, but it's also trying to train on sort of AI generally. What do you think is the strategy here? I mean, do you think open AI and philanthropic are doing this because they're like, we really want teachers to be using.
Are exact tool suites or are they trying to say, we just want AI to be baked into society that's good for all of us. Like is this a competitive move to Google or is this sort of a, trying to just raise the floor on AI usage in general for the next 20 years?
[00:12:10] Ben Kornell: Yeah, it's so, so interesting because I'm not sure that the people who are in this deal necessarily know what the end game objective here is, but I would guess they have a series of objectives.
And one objective is if AI gets banned in schools, that's bad for us. Yes. And let's partner with the teachers union who, let's face it, have the most power in the system today. Yep. To make or break a technology happening or not happening. Interestingly, there's been a huge backlash this last week. On some of the Google tools in like teacher channels, some of that, yeah, a lot of like haters out there showing what the worksheets are that they produced using the Google stuff.
And I think it just shows how risky it is to roll something out in a very, very big way and not have everything fully, fully tuned. Fully, fully baked. And I have a lot of confidence that Google will be nimble in, in like improving the different use cases. But it is interesting that you might, the teacher backlash is what we're seeing with Google.
Yeah. And open AI and anthropic or we're all in with the teachers. We're in with the teachers. I think the other thing here is OpenAI has to believe that if AI wins, OpenAI is gonna win. Like at this point, they're so baked in that whether they have 80% market share or 40% market share, or somewhere in between like.
They need the growth of AI to be successful and they're gonna be fine. And then for Anthropic, they're in a such a confusing place 'cause they have relationships with Microsoft who's also got a relationship with open ai. Their primary mode is B2B, working with companies they're working without.
[00:14:06] Alex Sarlin: So they're talking about Siri being connected to Anthropics models now too.
Yeah.
[00:14:10] Ben Kornell: And you and I, we happen to be fans of Claude. Yeah. And great Drew Ben and some of the team over at Drio, over at Anthropic. They do think about the moral implications in education quite a bit in a way that I respect a lot. So yeah, I think to go back to your original question, it probably is baseline, hey, $23 million for all of us coming together.
Not a big amount of money. Right. And we need teachers on the same page. Maybe Google should be thanking them for this. The second order would be, let's make sure that our training tools do train people on anthropic or on open AI so that they feel like it's easy to go to chat GPT or it's easy to go to open ai.
One thing that is a big debate, uh, how this will play out, that's more of a user cases. Will AI tooling and education be one where I just go to chat GBT and I get what I need? Or it's super important that it's baked into my learning management system and it's like embedded in all of my tools and for learners.
So far what we've seen is most people get like 95% of what they want just through the general LM Chat. GPT, Claude or just Gemini? Open Source.
[00:15:31] Alex Sarlin: Yeah. What do
[00:15:31] Ben Kornell: you think?
[00:15:32] Alex Sarlin: I think that they don't know what they're missing with that, though. I think that we're so early, like Chad GT and Claude and Gemini's, regular consumer facing sort of core products are very powerful and I think people are like, wow, it can do so many things for me.
It can generate images, it can make slides, it can do essays, it can do all, you can do assignments, lesson plans. You can do all these things. We've seen this, of course, ed tech, whole layer of middleware, of all of these tools that are trying to make it education specific, the outputs of some of these tools.
But I think what people are not. Really realizing is coming. We already see it coming with the baked in systems, but is that, I mean, it's literally like, yeah, somebody mentioned last week. I think a really great example of this is like the internet. If AI is the internet, which I always think it is, then being able to ask and get a response is Google, it matters.
It's a very easy use case, right? It's search. It's a very core use case, but it is not the whole thing, right? The internet enabled things far deeper, way more embedded, way closer. I mean, this is how we. Do most of our life now, and it's not through this sort of give and take based high level, just, oh, I asked for something and it gave me an answer.
That's what these tools do, and they're powerful just as Google is and was when Search first launched and all the search engines. But it is gonna be baked in ways that I don't think we can even imagine yet. It's gonna do all the data analysis, right? It's gonna do so much work in terms of differentiation.
It's gonna do so much work in terms of translation. It's just gonna be built into everything. It's gonna do so much work behind the scenes in terms of workflow, right? I mean, you can imagine a world in which, you know, if teachers need to get permission slips for something to happen for students, they don't do that.
The AI just does that, looks at the calendar and says, here's what's coming up. I'm gonna need permission slip for this, I'm gonna write it, I'm gonna send it, I'm gonna collect it, I'm gonna send it everybody. This is something nobody else has to even worry about. This stuff is going to happen soon. So I don't think that sort of.
Give and take. Search layer is gonna be dominant as the primary use for very long. I just think it's the one that everybody can just jump into and use. And it's what chat GPT was when it launched. So we are all used to seeing it that way, but I think it's just scratching the surface. But I, I mean, Microsoft I think is a mysterious player in this and mm-hmm.
I agree with you. I mean, there's been these plays over the years where companies put out all this free training. As a way to sort of build their brand as to a way to get people to use their tools early. We've seen in education companies like Figma and Adobe and Canva create education departments basically to be able to sort of embed these tools into education so that you can imagine The business case for this is if you learn, use Canva from when you were in seventh grade, well you may be using it as a professional, you may be asking your company to use it, and those things are all true.
I think in this particular case, this is just a wacky moment because Google is the incumbent here and Google I think likes to see itself as the newbie, the fast moving, and I think it is, it's still in many ways, even though it's a huge company, but in this particular moment, clearly OpenAI and Anthropic are coming in saying, like you said, we're gonna do something for teachers.
We're gonna be like totally bottoms up and Oh, Google Classroom might be giving you these worksheets and you might get insulted by them, but we're gonna work with you to teach you how to do it. And that's a pretty smart play. I don't think it's gonna last that long. Personally, I still think that being embedded has a huge advantage.
And until OpenAI or Anthropic has some kind of infrastructure for K 12, it's gonna be really hard for them to be actually truly what powers K 12 ai. I still think Google has a 10 year advantage on this. I might be proven wrong, but
[00:19:01] Ben Kornell: I tend to agree. And it may just end up not being an either or. It may be an and.
And I think like most of our entrepreneurs, what we're finding is that there's the embedded AI workflows and then there's supplemental AI workflows. Exactly, exactly. And to the degree that these companies don't need to make a business case for why they would do this. It's a user acquisition case. It's a defensive case.
Then maybe everybody can win. The real question I have, Alex, for us and our audience is what does this do to EdTech entrepreneurs who are trying to do this kind of work in this space? No, they didn't say We're partnering with AI for Education and Amanda Bicker staff note, they didn't say We're partnering with Teach for AI or with Alex and there's groups that have nonprofits and for-profits that have made their business doing this kind of work with educators as service.
Then there's groups that are trying to start up companies that may compete with these free services. Does this accelerate? Does this decelerate our a tech innovation ecosystem? What do you think?
[00:20:13] Alex Sarlin: It's a great question. I mean, I noticed a few weeks ago in the sort of White House, big announcement about 60 companies offering AI tools to everyone.
Curriculum to everyone. When you look down that list of companies, you had huge name tech companies and you did have the magic schools on there. You did have a number of AI native companies and sort of startups like the AI for Education. There were folks on there, so it was a sort of funny mix of those groups.
That is true that you do not see this in this announcement. That said, I think that the Open AI and Anthropic and Google folks all do realize that, I mean, what Google did last week. We, as we mentioned in the pod, threatens a lot of EdTech business models, no question. But I think with OpenAI and Anthropic, I would be surprised if these big libraries of training, these OpenAI Academy Anthropic put out a lot of tools for educators too.
If they aren't realizing the potential benefit of working with these go-between nonprofits or the middleware software solutions that are powered on top of them. I don't know. I'd be surprised if they truly box them out. I think just for right now. People aren't aware of them that much. I mean, magic School and Brisk and School ai, there's a few companies that I think have really truly broken through the consciousness of like many, many, many educators.
And they're used pretty widely. But I think of the 400, 500, you know, 600 AI tools that we all cover, they're all still new and they're all sort of straining for market share and they're all trying to get into the mix. And I think that if you're open ai, you don't have that much brand advantage. If you already have the unions and you have Microsoft in your announcement, I don't think you're gonna include an AI for education or a magic school because I don't think it adds any brand value yet.
But I do think behind the scenes they're gonna be working with them. That's, I guess what I'm trying to say. So I do think that in these giant trading academies, they're gonna make space for some of the ed tech middleware companies to be able to present how you can use it, because I don't think they wanna build all that functionality themselves, especially open AI and philanthropic.
I think Google seems to want to. And open ai. I don't think that they feel like that. The only way that they can work in educational system is through Shachi, BT, and Claude. I think they know that there's all of these other more sophisticated ways to use them, and that's through these middleware or these training solutions like AI for Education and IEDU.
So I don't know, it might not be out there yet, but if I were any of those groups, I wouldn't be that threatened by this because I don't think they're trying to be boxed out. I think they're trying to basically say, no, we care about K 12. We wanna make a big splash and make everybody know we care about K 12.
And then when the rubber meets the road, I think they're gonna partner with all these companies.
[00:22:52] Ben Kornell: Yeah, there's safety in numbers. There's safety in partnerships. Right? And let's remember also, channel is our biggest challenge in education. Through these partnerships, you're immediately getting a channel to 1.8 million educators.
I think it's important for us to just acknowledge that through grassroots Growth Magic School, the last number I saw, they had 4 million educators globally. Yeah. So it is hard to reach that many educators and with one stroke of a pen, like there's free resources in the hands of 1.8 million, that's for 23 million as the down payment on that.
That is a great move. Stepping back into the broader like AI space. Are there any other headlines that are kind of catching your attention?
[00:23:41] Alex Sarlin: Yes. The thing I think we need to talk about is the drama that is happening in the AI space around meta. I think it's really worth just taking a moment. This was the headline of Claire Zas AI and Education newsletter this week.
And basically, you know, meta the artist, formerly known as Facebook, as we all know, is basically pulling out crazy stops. They are changing the comp structure for what AI looks like. They're paying tens of millions, even hundreds of millions of dollars collectively to basically try to pull top AI talent from across the ecosystem.
And there are these lists of the people they have hired. They got the head of Apple Intelligence, which totally pulls out the rug from Apple. They're pulling people from OpenAI, they're pulling people from Anthropic. They're basically trying to pull everybody they can into Meta to create what they're now calling the meta Super Intelligence labs.
And they're doing it with just. Gads of money. And I think what you're seeing here is that Meta is very afraid of being left behind here. They've had AI labs for many years, but they are just, you know, we don't mention them, right? They've had this llama open source model, but they have not been as much part of the conversation about cutting edge AI as they want to be.
And Mark Zuckerberg and Meta are just going really tripling down on saying, no matter what happens, we are going to be a major part of the, the AI ecosystem moving forward. The question is, what are they actually gonna do with the Super Intelligence labs? And I don't think anybody really knows. The headlines were meta, as they have been for a while, tend to be pretty skeptical.
There's headlines about how they're using AI to try to get data or try to raise engagement, and there's a lot of skepticism.
[00:25:16] Ben Kornell: Well, bay Area politics, there's a way in which this is a national story, but this is also a hyper-local story. There is a dichotomy now between missionaries and Merc. That everyone is basically assigning mercenary to those who go for the money and go for meta and missionaries for those who stay at the other AI firms.
But let's be clear, people have made ridiculous sums of money just being part of the missionary work of OpenAI and Anthropic and so on, and I don't know what their liquidity looks like on their stock options and so on. But there's been a mercenary element of this from the chump with people essentially I, I would say DeepMind was kind of the genesis group that fractured and then that seeded and then open ai, like in nonprofit open AI days.
That group fractured. So there's kind of this root of all of that. I think that the challenge is that there's a flawed business strategy here, which is. The open source is not a great way to make money, and it is not creating a developer ecosystem that meta can really monetize. And the flawed business strategy also is like having the best LLM is a defensible lead.
It is not fast Following is quite effective. And yet you can see the strategies that OpenAI and Anthropic and Google have done and they're solid business strategies because they've embedded into business use cases where regardless of whether they're, you know, 5% better or worse than another thing, they're powering their business and then they have lots of channels to monetize.
They can monetize through cloud, they can monetize through other B2B partnerships. In some ways, the AI is something they can throw in there. So you know, one of the questions is like, has Zuck lost the magic touch? And I feel like this is reading in local, you know, and again, I'm down in San Diego now, but I'm still hearing from a lot of my friends up there.
It's reading as desperate move. Yeah. And the people who got laid off at Meta have a sense of relief where if people getting laid off at Google or Microsoft have been incredibly sad about that meta, it's like there's already kind of a cutthroat mercenary feel there. I think that it's a little bit more like, is this, they're gonna be the moment where meta finally buckles under all the bad bets on vr, and now they're making bad bets on ai.
Will they go down or can the Facebook engine, the ad engine, and can the Instagram and WhatsApp plays? I mean, they're gonna bring WhatsApp ads now. Can that save them? And the irony of that would be essentially met as a holding company of. Social media properties, not an AI company anymore.
[00:28:31] Alex Sarlin: I think you're onto something incredibly important there.
So you've gotta think that channel is everything. And Meta has had a number of very successful channels that have really dominated for many years. Obviously Facebook and Instagram, and then WhatsApp being the key there. But they do have others. They have the Meta VR channel. They're doubling down on that, continuing to, maybe not doubling down, but they're continuing to grow it.
My guess here, I may be wrong, but my guess here is that they realize that they don't have the clearest channel for ai. They can incorporate AI into Instagram, and they will, they can, and they are, and they will do a lot more. They're incorporated into WhatsApp and Facebook. But what they're missing is a consumer product that can compete in a, in a normal way.
I mean, compete head to head with the big AI players right now, you know? And I think that, my guess is that when they're bringing together this killer group, they realize they need new channels. If they have incredibly sophisticated AI builders who can create incredibly strong engines, they're gonna have to then put similar amounts of money and energy into building new channels that actually deliver that.
Because, you know, if you have the most brilliant reasoning model in the world created by all of the people who created the, the open AI reasoning models and, and then some, and you put it into Instagram, that makes no sense, right? I mean, Instagram doesn't need to reason like that. That's just silly. I would guess that they have some new channels in mind.
They're also thinking about super intelligence. So they are trying to push towards this sort of quote unquote A GI or however you wanna think about it. This idea of like, how might this intelligence actually really get to even, you know, new levels that we haven't imagined yet. But I think the new channels have to come, and I don't know what they look like.
I don't know if they're thinking that this is gonna occur in vr. I think maybe they've already. Feel like that chip has, has sailed a little bit and, and that wouldn't be where they'd double down. But I wouldn't be surprised if we start to see in the next six months some pro, some new products coming out of meta rather than being acquired like Instagram and Open and WhatsApp.
Some new products that are like, okay, here's a new thing brought to you by Meta that is, uses AI in a way that you sort of can't deny. The same way that Chat BT came out and said, you can't deny that this is useful, no matter how you feel about it. It answers everything. Formats, everything. It translates everything and it does all these things.
I think you're gonna start to see new channels because I think you're exactly right. Channel is everything. And I think they know that. And all of this investment in the backend, on the actual models has to be married with investment on the front end. And that can't be an Instagram and WhatsApp. It just can't.
Even though those are, that's where everybody is right now. They need new places where people are, where the AI is actually happening, the level that they want it to be. So I don't know what that looks like. I certainly dunno what that looks like in education, whether they're even thinking about it in that way, but I can imagine them thinking about it in other fields and putting together some tools that are just undeniable where the AI is so powerful and it does, it adds so much value so quickly that you just sort of suddenly they're part of the conversation, which is I think what, what Zuckerberg actually wants here.
[00:31:42] Ben Kornell: I think you give them a lot of credit and they've probably earned that benefit of the doubt. And I think the A GI as the game changer. You know, one thing to be clear is like we're looking at AI and the use cases currently and over the next two years, as you start thinking about an a GI like sentient AI world.
[00:32:04] Alex Sarlin: Yes.
[00:32:05] Ben Kornell: That is a world where having your own is worth a lot. And you know, mark Zuckerberg is one of the few, like ultra billionaires who still has full control of his company. And if I'm sitting in Mark Zuckerberg's shoes, you have infinite money. You, there's no way you can spend it down during your lifetime.
It just grows too fast. What am I most worried about? What am I thinking about most? What am I like watching my peers do? It's all about a GI
[00:32:38] Alex Sarlin: like how would they deliver the a GI? What does it do for people?
[00:32:42] Ben Kornell: You make a great, great point. You're basically taking logic. I've used. Before to describe other people with a competitive advantage.
And you're right. Exactly. You're right.
[00:32:50] Alex Sarlin: But I mean, the former GitHub, CEO, is part of the Super Intelligence labs. I mean, they, there are channels I don't think we've thought about yet. Maybe I'm giving them too much credit, but I, I'll recommend to them that if they're not thinking about product channels to put this powerful ai, they're definitely building into action.
They're missing out. But no, I'm with you. I'm with you. Don't get me wrong. I, I think your skepticism is warranted. It's definitely warranted. It's just, I don't know. There's a lot of money, there's a lot of talent and I still think there's. To your point in a world where a GI or, or super, whatever you wanna call, it's a true deep intelligence where you could, you know, ask something to say like, okay, get me a job in this.
And it goes, gotcha. And it comes back and it's like, you now have the job. Right? I mean, imagine that, that kind of world, you need new channels for that because right now it's, everything's hodgepodge together, right? You use Zapier or make to have Claude talk to, to LinkedIn and pull data from this and then send an email through your Gmail and blah blah.
I mean, that's not gonna work for very long. So you, you will need some kind of sort of integrated channel for a GI to work through and maybe that's what they're gonna work on. Maybe it's an agent platform. Yeah.
[00:33:53] Ben Kornell: Yeah. I mean I think there's also like growing skepticism about agent AI and how relevant and real it is, which I've, I scoff at.
But there's a couple people from ServiceNow and other companies that have basically said the nuance really is that human of the loop is really important with the state of current ai. And so the main takeaway is just like keeping human in loop is important. But yeah, there's just a bunch of chatter about like AI agents not being good enough yet.
[00:34:28] Alex Sarlin: I still think yet is the key word here. I mean, my take on a lot of this stuff, and I think about this in education, but also just tech generally, when the value becomes undeniable of something, it takes off. And I think that was true of Facebook. It was true of Google when it started. It was certainly true of the iPhone.
It was true of Google Chrome. By the way, perplexity is now launching a browser to take on Google Chrome. Speaking of channel, right? That's that perplexity. Another thing that doesn't have much channel, it's creating their, entering from a standing start with incredibly powerful ai. They're creating their own browser.
That's clearly their attempt. Maybe Meta will have its own browser too, I dunno, but. When the tech is just undeniable, it takes off. And I think that Meta has not had, that. They've been building these LAMA models, these open source models that fuel the whole open source movement and, and it's been a very particular approach, but they haven't made AI products that have broken through.
My guess, and I'll put it on record here, right, is that, you know, we'll be talking in, maybe it'll be next February and we'll say, today meta launched, blah. Something that is not Instagram, it's not WhatsApp, it's not vr, something new to showcase their a g or whatever, their most recent intelligence model.
And it'll be something if they get it right, it'll be something where you're like, okay, I have to try that today, because that is just too, that's too crazy. We talked the last few weeks about how there was a regulatory ban in this giant bill going through Congress that was gonna keep states from being able to regulate AI for, it used to be 10 years, then it was gonna be five years.
Yeah. The Senate struck it down. That means. Regulation from the states is still gonna be possible. I think that is probably a very good thing for the education space. But yeah, so we can maybe close the book a little bit. And the bill passed. We should say that too. The bill did pass and that is now the law of the land here, for better or worse, but the AI ban against states, it didn't go through.
What do you think of that, Ben?
[00:36:17] Ben Kornell: Who knows if it's good or if it is bad? I mean, I think the idea that there would be no legislation and a moratorium on any legislation scared a lot of us who are concerned about AI safety for kids, whether it's good to have it at the states or better to have it at the federal level from an EdTech perspective, sometimes it's consistency as a law of the land is better for us.
So we may end up having like tastemaker states who basically say, here's the law for our state, and then that essentially becomes the defacto law of the land. Or we could see the feds jump in. So I think it's clear that it's still unclear.
[00:36:55] Alex Sarlin: On either side of the political aisle I think at this point.
Right, exactly.
[00:36:58] Ben Kornell: You know, we, Texas, Florida, and California basically are making policy right now for a lot of federal stuff.
[00:37:07] Alex Sarlin: Yeah, that makes sense. And I think it is a good thing. I mean the idea of boxing all states out entirely was pretty scary. Even for people like us who are bullish on AI and bullish on its ability to do good things, it's a lot to not allow people to do, you know, any kind of regulation.
So let's talk about some of the ethnic things. There was news from Higher Ground education this week. It's a company you have personal connections with. Do you wanna tell us about that?
[00:37:31] Ben Kornell: Yeah, I mean, obviously I have like some insider intel here that I can't really share outside. And I also don't wanna overstate, I don't really have that much insider intel in that I was with Higher Ground for a year after they acquired Altitude Learning.
But I think the kind of headline news is Higher Ground has applied for Chapter 11 bankruptcy. They are a Montessori network of schools that start with infants and go, were went all the way up to high school, but most of the locations were infant to like lower elementary and they had over a hundred locations across the country.
And when they are filing for chapter 11 bankruptcy, that doesn't necessarily mean the whole thing closes down. It just means that they're restructuring their finances and that any debt holders or obligations that they might have with landlords and all that kind of stuff all gets restructured. And you know, I think first, like I'm sad to hear that they're kind of going through all of this.
I think the really hard part has been the impact on families who have been relying on those schools. And then the schools are closed down and so they're like scrambling for childcare. And the founders, Ray Kern and Rebecca Ke. Left the company. And so there's the new leadership stepping in are actually people that have been there for a while trying to kind of save the company, save the business, turn the business around, I guess.
But there are a couple good articles we can link them in the show notes that basically describe the kind of financial challenge of growing a brick and mortar business in a venture capital model for education. And the kind of core challenge is that you've got to basically have this hockey stick growth, but at the same time, your capital structure is so front loaded with physical build outs and rents, and it takes you three to five years to fill out a single site.
And so I talked to tons of people who are interested in private schools or new schools or micro schools, and I think each model has its own nuance. But ultimately a venture capital model where on a seven year or 10 year timeline, you're expecting a hundred x results. That is a really hard fit for physical brick and mortar businesses.
Nord Anglia, Matt Tower, our friend pointed out, it's like over $1.5 billion business. So it can be done. It just has to be done on a slower timeframe with a capital structure that is different. And most of the capital structures that we see that work are either like wholly owned, homegrown. That's very rare.
Second is private equity backed where you're using debt, which just requires paying back your interest or franchise models. And the franchise models have been quite successful, especially in early childhood because you have these dis economies of scale as you grow, like your fifth location, your sixth location, your seventh location actually suck attention and resource away from your first couple of locations.
And so when you compare a growth model where you've standardized certain parts of it, but then you've got localized like owners who are really incentivized to drive local engagement, that can be a really winning model. So right now, the company I work at at Art of Problem Solving, we have digital and we have physical locations.
And I think there is this question of like, what is the steady as you go growth rate that. Organization can handle where you don't get over your skis in terms of expanding too fast or taking on too much in terms of lease obligations and schools themselves are the biggest capital investments because you've gotta build out all the ti, you've gotta have the certification as a, there's a huge lift to kind of getting it launched and a lot the available real estate is not great for that.
You know, ultimately, while I'm, I am absolutely sure there were mistakes made around, you know, lots of different vectors. Ultimately, I think what did in higher ground and their model was trying to chase a hundred x revenue, multiple or growth vector, when you know that it's very hard to compete with the gravity of physical brick and mortar sites.
[00:42:05] Alex Sarlin: It's a great analysis. A couple of other just headlines that we wanted to mention before we go to our guests. We saw Honor Education, which a really interesting company started by the ex-head of Apple University. We really, really great group secured a $38 million Series A. They're trying to figure out how to make asynchronous learning feel social and engaging, which if you get that right, you actually really do change the landscape of of EdTech and they're, they've got now got some money to really pursue it and they're doing it with very high quality design in the sort of Apple tradition.
We also saw Teaching Lab, which is a nonprofit that does a professional development, curriculum based professional development combining just right now with Relay Graduate School of Education and Sarah Johnson, who is the head of teaching lab and a big advocate of AI in education is now going to be the head of, of the combined organization.
That's really interesting. I think, you know, to your point earlier. That type of organization is a very natural partner for both the Google Classrooms of the world and you know, the Open AI and academies. And the philanthropic academies because these are groups that are really thinking very hard about how to help teachers use AI in really professional, really curriculum aligned ways.
And they know more about it frankly, than the frontier model providers do, even though they make their own materials a lot. So congratulations to Sarah Johnson and Keep, you should definitely keep an eye on Teaching Lab and rely on the combined org. I think that's it for us, for EdTech Insiders this week.
Do you wanna go to our guest, Ben?
[00:43:30] Ben Kornell: Sure. I'm excited to have Josh join. I was able to go down to the Dreamscape Learn Studios in Los Angeles and actually experience the full immersive alien zoo where I was also able to dissect an alien frog with like three hearts. So excited for Josh to come on, and also excited for our EdTech Insiders.
To continue the conversation, we have our WhatsApp channel, and that's part of our EdTech Insiders Plus membership. So if you get a chance, check out our substack. You can back slash subscribe and join in on the EdTech Insiders conversation. 'cause if it happens in EdTech, you'll hear about it here on EdTech Insiders.
Thanks everyone. And onto our guests.
[00:44:17] Alex Sarlin: For our deep dive this week, we are here with Josh Reibel, the CEO of Dreamscape Learn, and Dr. Jennifer Cruz, the superintendent of the Pendergast Elementary School District in Phoenix, Arizona. Lemme give a quick biographies of both them and we'll jump into the conversation.
So Josh Reibel is the CEO of Dreamscape Learn. It's an immersive education company formed from a partnership between Arizona State University and Dreamscape Immersive, a VR based entertainment company. Founded by Dreamworks Motion, co pictures co-founder Walter Parks Dreamscape Learned, delivers transformational curriculum experiences through the integration of cinematic storytelling, advanced pedagogy, and cutting edge immersive technologies.
I've done the Dreamscape Immersives myself there. Amazing. And Dr. Jennifer Cruz is currently serves as the superintendent of the Pendergast Elementary School District, located in Phoenix, Arizona. She has previously been a teacher principal, professional development consultant, and a chief academic officer, and her career in education reflects a deep commitment to equity, innovation, and student success.
Josh Reibel and Dr. Jennifer Cruz. Welcome to EdTech Insiders. Thanks, Alex. It's great to be here.
[00:45:23] Dr. Jennifer Cruz: Thank you so much. It's good to be here.
[00:45:25] Alex Sarlin: It's fantastic to talk to both of you. So Josh, let's start with you. Tell us a little bit about how Dreamscape learned works. It blends virtual reality with storytelling, and how has it work in the educational context, and what evidence have you seen that it actually moves the needle on student engagement, on outcomes on some of the things we care about most in education?
[00:45:43] Josh Reibel: Well, I'm glad you mentioned engagement in that sort of build up to the question because that's really what we're focused on principally more than anything else in that, look, we all know what kinds of things capture our attention and get us in a mode where we just can't wait to find out what happens next.
We're all addicted to episodic television on Netflix or Hulu or whatever, and yet curriculum in school, which is a mediated experience in many ways. Reflects none of those sort of values and design principles, and we sort of wonder why students tune out. And the point I'm making here is not that we should turn school into chocolate covered broccoli and make it fun and games as we'll explore in this conversation.
I hope what we're doing is really, really rigorous academically, but what I am calling attention to is how we turn school into an experience that students want to lean into that is as binge worthy as what they're addicted to in the more entertainment arenas of their lives. And so that's where the storytelling aspect of what we do comes into play.
The curriculum solutions that we provide are often, although not always, plotted narratively where they have episodes and there are twists and turns and suspense and characters and emotion. And what we have found, not surprisingly, is that when students become emotionally connected to. A narrative. They will do enormous amounts of very hard academic work in order to fulfill their role as protagonists in a story, as heroes in a journey.
And so one of the ways I like to describe the difference between business as usual and, and what we do is just by illustration of an example, for instance, in our chemistry product, which is about two thirds done and will come out next year, well, better way to put it, you all remember your first day of chemistry class, whether you did it in 10th grade or you did it in college, and I'm guessing it was something like read the first chapter of a textbook and do the problem set at the end, or you were lectured at in class and then they sent you off to read the first chapter of a textbook.
So day one of chemistry class with Dreamscape Learn is you come into our environment, put on your headset, and all of a sudden you find yourself piloting a turbo copter through the foothills of the Himalayas. And you land in a remote farming village and see that the local yakker is clearly ill. And you run some tests and discover that this very cute creature who you really don't want to be sick, seems to have arsenic poisoning and there's a river running from the mountains ahead down to the village below, and you pretty quickly realize, oh, that's gotta be what's conveying the arsenic.
That's a problem because there are a whole bunch of people living in that village downstream from here. We need to figure out how arsenic is getting in this water and how we're going to solve that. And therein begins a multi-week, multi episodic set of academic challenges to figure out what is the source of the arsenic and what do we have to understand about chemistry in order to eliminate it.
And I guess the last thing I'll say before I pause is, yes, we use immersive technology and VR to put students literally inside these experiences and inside the story, but for every 10 or 15 minutes they're in vr, they're spending two or three hours in a regular classroom environment using spreadsheets and books and pencils and all the usual tools of academic work to learn the math and science in the case of our STEM products that they need in order to advance the narrative and solve the problem.
And I would say the most exciting aspect of what. We've seen happen, and I'll get to the sort of outcome studies that have been done. So aside from that is the energy that students are bringing to the non VR work is every bit as intense as the Wow, that was incredible. As when they take off headset, and that's what matters is that the students want to be there hovering in front of a spreadsheet, modeling data, trying to figure.
Why have I not gotten this right? And if I don't get this right, that cute yak sanu is not gonna make it, and I need sanu to make it. I'll just say one last thing so I don't take up all the time. We're really only just getting going in K 12, but in higher ed where this all originated, there have been quite rigorous outcome studies done.
You know, hundreds of students demographically matched cohorts in the treatment and control groups, real randomized control trial type studies. And the two main things that we found that are really exciting are one students almost twice as likely to get an A in their science labs. I guess I say three things despite the fact that those labs in the dreamscape learned version are substantially more quantitatively demanding than what is typically done.
So it's not that it's just easy, but the most important finding, which goes to Dr. Cruz's career commitment to equity, is that the underserved populations, those that normally struggle most in their intro biology or chemistry, intro stem experiences in college. Are doing as well as the honors students in courses where historically there are massive achievement gaps between those populations.
And so we really feel like this is gap closing stuff, and it doesn't surprise me that much. If you make something emotionally compelling and interesting, people will do it.
[00:50:43] Alex Sarlin: Absolutely. And yeah. Dr. Chris, I wanna bring you in as well, but I mean, when I hear this immersive storytelling, epic narratives, this binge worthiness, the episodic nature of what you're talking about, it feels like, as you say, it brings in so many of the aspects of entertainment that especially young people love and subscribe to all the time and puts it into their experience, and it just changes the whole nature of what education looks like.
It reminds me, you know, the movies are, they have these montages where somebody is training for something or studying for the big test, and they always do it really quickly in the movie, and it's like, you know, it's just a few, you know, a minute with music over it. In this case, that's the heart of the work, right?
You have a 15 minute VR experience and then you're actually drilling and learning and making sense of it. The inspiration is there. You're inside the story and it and sort of bridges between the different story elements. So Dr. Cruz, tell us about what this looks like. You've led innovation at multiple levels, you've worked at school district level, you've done pilots.
How are you using Dreamscape learn and what does this all look like in classrooms?
[00:51:41] Dr. Jennifer Cruz: So we're really excited to bring Dreamscape learn into specifically the middle school level and it's actually really brain science friendly. So we learn by connecting our experiences with emotions and when we can have that emotional response, emotional connectivity to the work, then we wanna work harder, right?
So we were super lucky to have the opportunity to bring Dreamscape learn into one of our school sites. We actually have a fixed pod at Via the BA Elementary school to Computer Science Academy. And we were just really surprised with how quickly our students were learning content. We really felt like we needed to expand that.
So for scaling, one of the first things that we had to do was explain that. Why, like, why are we moving in this direction? Obviously, you know, we're a traditional K eight public school system, not exactly the institutions that are known for innovation or aggressively moving into an entertainment space, right?
To get kids connected to the state standards because we are very focused on our Arizona academic standards. So the first part is why are we doing this? You know, we had to really show the team student engagement levels. We saw improvement in student attendance, our science test scores, at least with the group of kids.
We did this with doubled this year over last year. There's still a long way to go for real improvement, but we're very excited about what it's looking like. So we started with why and having folks experience it. Students had the experience, parents had the experience, teachers, leaders. And then now we're working on the logistics.
I know it sounds really weird, but in an institution, if you don't make this user friendly, like if the end user has a rotten experience, right? Whether that end user is our faculty member, or if it's our student, then it's gonna be ultimately a no-go. I'll just, they'll just passively vote themselves out.
So our team has been working really hard on logistics, on communication, on training, but we've been very lucky to hire folks that have an innovation mindset to implement this work initially, and their excitement is contagious. So that helps a lot as well. And then we're gonna come right alongside folks and train them.
So we've got a fantastic team and we're focusing on the idea that really this is the way that people have always communicated. You know, we've always communicated through stories and we've always motivated ourselves to dig a little bit deeper in empathy. And this really helps both with empathy and engagement, and it's motivating kids to work harder.
And so our, our staff is really excited about it right now.
[00:54:12] Alex Sarlin: Yeah, it answers the sort of why question very quickly, right? I think that's always an engagement issue in schools. Why are we learning this? 'cause eventually you may use it if you become a chemist or be, if you go into, you know, engineering in a certain way.
But if the why is, 'cause I need to save this creature tomorrow and I need to put it together with my team and I'm immersed, I care about the story. You just, you can keep that momentum going really, really quickly and it's amazing to hear how it's working. Let's talk about the storytelling piece of this.
'cause I think the emotional engagement, as you're mentioning Dr. Cruz and the storytelling, that doesn't come easily. The logistics and buy-in is always key to any educational tech. So I think you're speaking, yeah, I think people would definitely recognize that. But the idea of making a story that people actually wanna see and that experience people actually wanna do.
Historically that's been hard in educational tech, right? Where people create sort of the AR trail version or the sort of the light story version. But these kind of really immersive story with characters with where you're really, really in it. It comes from this collaboration in dream learns that case between a SU and Hollywood studio.
So tell us about that model, Josh. Tell us about how Hollywood studio level storytelling can combine with educational technology to create products like this.
[00:55:20] Josh Reibel: So there are two aspects of that. One of which we've talked about a little bit already, and I'll start with, which is the storytelling, but there's also the cinema of what we do, and I'll come to that in a moment.
The storytelling is the hardest part. Everybody who asks me what's the bottleneck, what are the challenges, what makes it hard to build more of this stuff? It's that the creative side, not the visual creative side, the narrative creative side talent in that area is so scarce. Now, fortunately for us, not necessarily fortunate for, you know, a lot of people who live in la Hollywood is sort of crumbling, and so there's more access to experienced good writers than there might have been five, 10 years ago.
But it is really hard to find people who have the ability to write great narrative and have the sort of intellectual. Scope and curiosity to sit with a tenured chemistry professor and understand the domain enough to figure out how to weave that into the story properly. And so anyway, so we don't have an army of those people.
We have a troop and sort of figuring out where to find two, three more is one of the biggest challenges of the company. The other side though, the cinema side of what we do, I think sometimes gets overlooked and is really, really important. So lighting a soundtrack, all of that stuff and having that stuff done properly is what gets people to suspend disbelief and lean in.
Like so much of early VR was your sort of a very flat, ironically sort of disembodied head connected to hand controllers. In a cartoon world that doesn't feel real, it's sort of cool that it's all around you. And so the first time you put on a headset it's like, wow, this is in 3D, but you don't feel transported because the space that you are occupying.
Doesn't feel authentic. I'll give an example. We've worked on some humanity stuff that involves among other places a sort of 15 minute curated lecture and sort of guided tour of the is Sophia in Istanbul. There's an extraordinary place of worship that is at sometimes a history, been a mosque, it's sometimes been a church, really rich environment for learning a lot about various civilizations.
Figuring out how to take people through that in a way that really does make them feel like they have left. Phoenix in Dr. Cruz's case and are in this place is key. And that, yes, it's hard to find good production talent, but that's not nearly as hard as finding good narrative talent. But those production values take a real commitment, take real work, take real capital.
You have to believe that if we get this right, it's going to make a difference and it's gonna scale, and there's gonna be a return on that investment. And unfortunately, as, as you know, well, Alex, in our industry, innovators haven't always been willing or able to make a commitment to that level of production quality.
And I wouldn't say that there's a race to the bottom necessarily, but it's a challenging market where not that many providers have been willing to take that kind of risk. And so I think we are finding that being fussy about. All of that cinema is as important as the narrative aspect
[00:58:20] Alex Sarlin: of it. That's really interesting.
Yeah. That idea of production talent and narrative talent and bringing them together again, like a Hollywood studio and putting together all the pieces to make something that can truly transport learners. So it doesn't feel like this is an educational, and it
[00:58:33] Josh Reibel: really goes to the equity issue. It's really important.
Like, you know, there are a huge number of students in Dr. Cruz's district who have not been very far from their homes. Then there are lots of students in other districts that are going on safari with their parents all the time because they're wealthy and they're well traveled and well, and for us to be able to use this technology to really bring students to places they could never otherwise explore is so world expanding for students who live an otherwise kind of provincial life and an unfortunately, quite bounded life.
And that alone just opens up, you know, especially at the middle school level, students' eyes for like what's out there. Like what's possible, what could I be in life, where could I go? And no textbook can do that.
[00:59:13] Dr. Jennifer Cruz: And we find too, you know, to Josh's point, that our students have rejected a lot of the typical ed tech games because they find them so boring and so poorly executed in comparison to the games that they play, you know, when they're home or even some of the gaming that they can do, frankly, on their cell phones.
Phones. And so, you know, we have a more sophisticated end user than I think we have been designing for. And when you give the end user the product that they expect, as it turns out they enjoy it and they're willing to use it more, who knew, right? So we just have, and I think taken seriously young people's expectations.
And because we haven't taken their expectations seriously, we have seen a decline in performance, which is really, you know, to be expected. So I think that's really what's different about this work that we're doing, is that we really genuinely believe in the, in the end user experience, both for the teaching faculty.
And for the young person who's using the material.
[01:00:15] Alex Sarlin: Absolutely. I love that point about the sophistication of the end users. I mean, young people today, I would argue are the most sophisticated media consumers in the history of the world. I mean, they have access to every show, to every app, to every game, to everything.
They could absolutely pick and choose. And to, I think both of your points about equity, the access to those top level games and movies and things are not limited by money. Most of the access is actually equitable. There are some games that are very expensive, I guess, but not really that many, especially applications.
So people have so much experience with top tier entertainment experiences that their expectations are incredibly high. And it makes sense that education has to step up like, like what you're doing with Dreamscape Learn. So I wanna double down on this question about equity. It's come up a number of times here, you know, in terms of the outcome studies from Dr.
Have learned in the past in terms of what you've been seeing at Pendergast. But tell us a little bit about how this type of technology does work across the equity spectrum. How does it work to be an equalizer and allow students of all backgrounds to really be able to jump into chemistry or other topics with immersion and excitement and smooth out some of the things we, we've traditionally seen in the educational system around inequality?
[01:01:26] Dr. Jennifer Cruz: Josh mentioned, you know, some of our students have limited access maybe to leave the city, the town, you know, the county that they're in. We have a number of students who haven't had the opportunity yet to leave the city of Phoenix. And so giving them access to new locations, to increasingly challenging academic vocabulary, you know, to put them in a position to make, uh, critical decisions, you know, really gives them that opportunity to Now put that again when we think about brain science, right?
To have that emotional interaction with complex material to be a decision maker. So now they have some agency, and then combining that with rigorous academic standards is something that's getting our kids more able to process the information in a way that they can use it in a novel setting. And that's really what needs to happen, right?
So we want, we wanna tie experiences to a novel experience. And then have that wisdom, have that learning from a previous experience that you can then apply. And that is something that unfortunately sometimes our students don't always have access to. Right? That opportunity to learn in a different time and space, to learn in a different context, to learn in a different location.
And so this does that for them. You know, for us, the other thing that's important is that we don't have a minimum bar set for your academic, your current academic performance. In order to engage with this material. Everyone will engage with this material. And so, you know, one of the things that's been, you know, folks keep asking us, well, how are these, you know, young people, you know, they're eighth grade students, they're accessing, you know, a college freshman biology course.
Unfortunately, we have a number of students who, if you were to give, you know, a reading test to, you would say, well, gosh, this, this person isn't even reading at the eighth grade level. How are they accessing this college biology vocabulary? And, you know, our response is they've always been able to, you just never gave them a reason to do it.
[01:03:21] Josh Reibel: Or a context like if language happens, it happens through, I literally.
Definition of some number of the words that come up in, in a dreamscape learn module, but it's all situational. So that's how you become initiated into a language or, and it could be language of biology, language or chemistry. And so, so many students who don't think of themselves as naturally being good at school don't do well simply because they don't think of themselves as being good at school.
And being good at school is a weird thing to be good at because of the way we do school. And so, but you know, there are visual learners, there are auditory learners, there are algorithmic learning, whatever. Everybody likes a good story. Everybody. There is not a human being that doesn't respond to a good story.
So once that captures you, there's not enough room in your brain for both fear and excitement. And so, you know, the fear goes away and then they try, and then it turns out that they can, and that creates a virtuous cycle.
[01:04:18] Alex Sarlin: I'm sure you see that, you know, Dr. Cruz on the ground where, you know, somebody reading a textbook chapter about some concept in biology or chemistry might not be very relevant and it might have, the reading level becomes a barrier to entry.
But doing it inside a story, doing it inside a immersive experience where you're trying to accomplish something, you have to learn the vocabulary in the context of that mission. And solving that problem just changes the context enormously.
[01:04:43] Dr. Jennifer Cruz: Exactly. Yeah. When you, you know, if you talk to a young person about the things that they're interested in, their vocabulary is right on target and is potentially far more sophisticated if you're asking them about a technology, you know, situation that excites them.
Right? So, you know, I have students who are talking to me about oxygen saturation levels. You know, they're talking about RNA, about DNA, and then other folks would say, oh, that, that student isn't very academically inclined and they're actually, we all are. You know, as long as we can put that in a context that's meaningful for us.
[01:05:13] Alex Sarlin: I have to ask you both a little bit about the AI aspect of this, and Josh, let me start with you here. You know, you mentioned the production talent, narrative talent, having to, you know, building these really immersive experiences at a sort of cinematic level. I'm curious where you see AI playing a role both in the creation of the content and perhaps in the
[01:05:30] Josh Reibel: delivery.
That's exactly the right way to ask the question because we use it in both development and as support for students or beginning to use it as support for students. So let's start with the product development side. I'd say we're in the middle of an evolution there, where today's tools allow us to do a lot in the sort of pre-production phase of what we do.
There are tools that allow us to automate, storyboarding and create, you know, an animatics that allow us to see, okay, so how's something gonna unfold before we put those expensive animator cycles against really rendering it in 3D and so on. And that's very helpful. There's like script extraction tools that'll add the tools that allow us to voice over something, cloning very particular celebrity's voice as the narrator and see how that plays and, and so all of that is great and does save us time and money, but we're not yet at.
Point with 3D tool, AI tools where we can say, render me a, a great Sistine Chapel and it will come out, you know? And Wow. And I no longer need an artist to do that. There's just, it's not refined enough yet. And so a lot of the actual animation, artwork, lighting, design, et cetera, does require human touch.
But that's okay. It means that like we're applying human ingenuity in the places where it really makes that last mile difference. Now, will there be a time in the future where even that stuff starts to get more automated? For sure. And we will continue to leverage those tools as they, as they evolve. The other side is in the delivery, where quite interesting thing came out of our partnership with a SU, where a lab there started building essentially an AI agent.
That is fully knowledgeable about the whole curriculum in A DSL curriculum. So for instance, in the biology course, they know the whole backstory. Our biology labs take place in a fantastical environment, what we call an alien zoo. The conceit is that sometime in the future, a space-based orbiting sanctuary for the endangered species of the galaxy has been created.
It was actually originally a movie idea that Steven Spielberg and our founder of Walter Parks had, and then they never made the movie. So Steven told us, go use that for vr. So the great thing about it is there are all these creatures and flora and fauna, and there's a whole ecosystem there to be stewards of and to save when it hits crisis.
We've created an agent that not just knows all of the facts of life science, the way, you know, chat, GPT could, but also knows all of the backstory of the alien zoo. And so it really feels like to students like, oh, this is my. Sort of mission guide who at any time I can ask a question and say like, so wait, when such and such creature collapsed and we did so and so, why do we do so and so?
And it'll remind them. And like if they get stuck in some modeling exercise, it's there by their side to guide them through that too. That is coming out of sort of last phase of RD. It will be in the product, I'd say within a matter of months, small number of months. And it'll be exciting to see whether, is that a tool that like helps, you know, students keep momentum in the course or is it just kind of a, a kind of nice play thing and, and we don't know, but everybody seems to be obsessed with the idea that AI tutors of one kind or another can help support students?
I think the more specifically that tutor is tuned to the particular activities and exercises that student has to do, the more likely it is to actually. Be valuable. And so that's, that's what we're trying to do.
[01:08:53] Dr. Jennifer Cruz: We think of that as like a learning concierge, right? Like you, you have a question, you need a recommendation, you know, and that's when AI can step in.
We've had some challenges with our youngest students who are using an ai, uh, reading tutor, and they don't particularly care for. It's way, it's, it's emotional tone with them. The young children prefer their teacher and the nice way that the teacher interacts with them and the more personalized way that teacher interacts with the young person.
And so it just kind of proves our point, you know, that people need people and we always will. And I think that to a certain extent, the younger the human, the more the caretaker should have a soul to make sure that we're making these important emotional connections with folks. But you really can't be AI as a concierge for experiences or ideas or questions as long as you understand what you're interacting with.
And you understand that there are limitations, you know, as a result of that. Um, and we think that we can teach young people that. So we're super excited, you know, to be implementing ai, but in reasonable ways and ethical ways, and making sure that we know who and what it is we're connecting.
[01:10:01] Alex Sarlin: Exciting visions all around.
I think the delivery piece, the idea of helping concierge, helping students be able to understand how to navigate where the activities they're doing, you know, with a sidekick that actually understands exactly how it all fits together is extremely useful and valuable way to do ai. And then of course, you know, as the AI for production continues to evolve, that's gonna be part of it.
And Dr. Cruz, very last question for you. I just want you to put on your sort of futurist hat here. You're doing these pilots, it's starting to build out this sort of virtual reality curriculum style of learning and, and you know, really being cutting edge with it five years from now. As vr, it continues to evolve as AI continues to evolve, as students continue to have more technology at their fingertips.
What is your dream of where this could go when you know, a few years from now when there's just even more capabilities within all of this technology?
[01:10:52] Dr. Jennifer Cruz: It really is my hope that this allows us to personalize student learning experiences in ways that we don't necessarily have completely mapped out right now.
You know, we would love for schools to be community hubs of learning. So not necessarily just for a specific age range of, of learner, but you know, maybe more broadly a spaces of learning and opportunity for folks. You know, we're, we're bridging the gap between public schools and businesses because we're really anxious for young people to find their purpose a little bit earlier.
You know, we're, we're excited about some of the virtual reality work that we've done with career and technical education training and, and I know that we're doing that pretty young. We're doing that with middle school students, but again, having a young person find out what they love and equally important, having them find out.
What they dislike, like what they genuinely dislike can actually be really helpful, you know, as they start to find their way in a complex world. So, you know, we're hoping again, to create these hybrid models, you know, where, where people are coming to schools. We know that we still need teachers. We will always need good teachers, but that there's this kind of play, this give and take between the learner.
The state standards, they need to learn the career and technical, you know, skills they need and, and educators, you know, and groups of educators working differently together. So we're super excited to be able to reinvent public school. We certainly think, uh, it's well past time and really grateful to the leadership of our school district that our governing board understands that we need to really move forward if we're gonna make sure that we're providing true opportunities for our students to compete and to be successful in the future, and that they're, they're willing to take a chance on some innovations on our way there.
[01:12:29] Alex Sarlin: I love that Dutch Jennifer Cruz is the superintendent of the Pendergast Elementary School District in Phoenix, Arizona, doing really innovative things with VR and technical education in immersive storytelling. And Josh Rebel is the CEO of Dreamscape Learn immersive Education company formed from a partnership between a SU and Dreamscape Immersive.
Thank you both so much for being here with us on EdTech Insiders.
[01:12:51] Josh Reibel: That
[01:12:51] Alex Sarlin: was a real pleasure, Alex. Thanks for having us.
[01:12:53] Dr. Jennifer Cruz: Yeah, thanks so much, Alex. It was great to talk.
[01:12:56] Alex Sarlin: For our deep dive for the week in EdTech this week, we're here with two amazing EdTech founders. They're both named Thomas, and they are the co-founders of Eduaide.ai.
So Thomas Thompson is a co-founder and CEO of Eduaide.ai. He studies how people learn and how teachers make instructional decisions. His work draws on cognitive science, instructional design, and the history of philosophy of education to develop tools to translate research into practical applications.
He's a former classroom teacher and a nonprofit director, and his focus has remained the same across all his roles. Building practical systems that support effective teaching and reflect the complexities of how people think, learn and develop over time. Very important. At this day and age, Thomas Hummel is edges.
Chief product officer and co-founder, but at heart, he's just another teacher working to make a difference. He believes the best way to truly understand the challenges and joys of teaching is to stay in the classroom. And that's why he spent the last nine years teaching middle school science where he's seen firsthand how curiosity and critical thinking can transform young minds.
Welcome to the podcast. We have amazing Teacherpreneur Thomas Thompson and Thomas Humel. Welcome to EdTech Insiders.
[01:14:05] Thomas Hummel: Thank you so much for having us, Alex. Thanks, Alex. That intro is amazing. I feel great about myself.
[01:14:12] Alex Sarlin: Well, you really are. You're both educators. You keep education at the core of your product, and I think that's sort of key here.
So Thomas Thompson, let's start with you. For people who are listening who aren't yet familiar with EDU eight eight.ai, tell us what pain points in the classroom led the two of you to jump in and become AI entrepreneurs trying to support education through ai. Certainly, I
[01:14:32] Thomas Thompson: mean, we created Edgewood as teachers.
It started in the classroom, so we're not technologists by training. So the pain points that you're referring to were very much immediate and personal to us. We taught at a rural school district on the eastern shore of Maryland, and we saw high rates of turnover year over year. What accounts for that turnover?
Well, there's quite a few things, right? We talk about burnout, motivation, other retention mechanisms, but ultimately the question is how do you build excellence in a field that struggles to retain talent? So we wanted to create a platform that would, yes, do all the things that you hear. AI platforms talking about doing right, retaining teachers through reducing burnout and reducing workload.
But that's just the first piece. The second piece really, and this is the lasting piece for us, how do you translate research to practice and education? Teachers often know what effective instruction looks like, but it's hard to really put like a fine point on it. It's like beyond saying space, your practice and interleave concepts, it's like, well, what does it mean to translate that practically into the classroom?
So there's a lot of translation work. It's a high cost. First you look at the research, okay. And then you try to figure out what that would mean in your context. And then you have to create the materials. And how do you know the materials are in alignment with that research? And then how do you implement it with fidelity?
These are all open questions and there's professional development in many things, but those things did not seem as effective to Thomas Humble and I, you know, we were sitting in professional developments wondering, is this going to circle back to the classroom? They're saying we should use such and such method, but they're just lecturing us as are they telling us that's effective?
Their actions would certainly seem to imply that. So it's like how do you embed professional development in the very act of planning? Yes. And we wanted to build a tool where you're now thinking about your instructional decisions very consciously, but in a way where you're not feeling so bogged down in the minutia of the translation costs.
So really those three high level things, it's research to practice gap burnout and motivation, and then excellence and kind of a leaky talent pipeline.
[01:16:31] Alex Sarlin: Yeah, we just saw a Gallup poll this week that said that teachers are saving an average of six hours a week using ai. Mm-hmm. That was incredibly exciting and encouraging for those of us who care about teacher burnout and reducing workload.
But I love where you're going that it's beyond administrative workload. It's about actually getting learning science into practice in the classroom, which is really hard to do, but AI can really help 'cause it makes sense of all of this data. Thomas Hummel, let's talk about your, you've been a practicing teacher for nine years, you're the chief product officer.
What does it look like to get that learning science and professional development baked into the product, sort of at the point of use?
[01:17:06] Thomas Hummel: Yeah, so there's like the use level of that where I blow up my team at like six o'clock in the morning when I'm in the building and I'm trying to plan my lesson for the day a little bit or make it better and the technology may not be filling the gap that I need.
When my time is crunched, just like all teachers' time is crunched on their planning period or whenever they come into the building. So there's like part of that where we're able to try to stay on top of what teachers actually need. Because if our product falls short, I'm using it every day to enhance my lessons in certain ways.
So that's a great asset to our team. And I would say the areas where we're finding that this has the best use cases is not to just recreate the wheel, but to supplement the needs of your students. And as the teacher, I know what my students need better than anybody, better than the administrator, better than some sort of technology that's scoring them.
And so I have the ability to fill the gaps of my curriculum that is just kind of cookie cutter. It's okay. Mm-hmm. Right. It gets me the basis. But what I really use the AI for is to supplement certain things. So like if I have a reading, I can scaffold a high, medium, a low, I can do a variety of questions, I can add context to things.
And so it's just like. Everything that you were asked to do as a teacher, right? As far as differentiation, personalization. Yep. All these expectations that. We're never realistic without AI kind of become realistic and it's making me a better teacher. So, you know, it's kind of fun actually. It's making teaching fun again.
[01:18:38] Alex Sarlin: That's my hope for what AI can do. You talk about retention and burnout and it is going down a little bit these days, which is fantastic from the sort of high levels of the pandemic. But teaching is a profession that I think people go into because they wanna make a difference. They want it to be fun, they want to share in the fun with their students.
It's not meant to be a a slog for anybody, but then when you're putting together all of the learning science, the differentiation, the curriculum and trying to match it to your students' needs, it's just a lot of balls to keep in the air. And I think AI can really help with that. Give us a sort of walkthrough of a day in the life of a teacher in the classroom using edu aid.ai.
What are some of the tools that they might use that would sort of showcase using ai? Well, to supplement a curriculum with additional resources?
[01:19:21] Thomas Hummel: Yeah, so like, it just depends where you're at in your professional journey. As a first year teacher, maybe it is writing lesson plans and maybe it is doing the things that you have to prove that you understand as far as pedagogy and the learning sciences, or even just like classroom management.
All these things that you have to account for before they actually take place. Maybe that is where you could use edu. Wade at the beginning of your teaching journey. For me, I have 180 lessons that I already do, and I do 'em every year, right? And so I don't know that, right? So for each teacher across their practice, they're gonna use it for different things.
I'm using mine to make everything higher quality. In terms of like really pushing my students, but I'm also using it to personalize things like my students love to fish, so like as many times as I can change questions to like about fishing and like the Chesapeake Bay and stuff like that, that makes sense.
The amount of just buy-in that I get from my students where I can correlate what they have. To where they're at in the reality that is just like everything for me. And I mean, it's just funny. It's like you can literally just go in there, find a couple different fish, find a bunch of different contexts for it, and now every kid, the kids who hate school are so engaged and it's been a total game changer in that regard.
So, I don't know, I think I answered your question there. I'm not sure.
[01:20:38] Alex Sarlin: Yeah, I mean I think that sort of interest based personalization, sort of closing the last mile gap of why should I care? Just immediately bringing that relevance to the classroom if students would rather be phishing than make them understand how what they're doing is connected to their phishing journey.
And of course, you know, phishing is a stand in for any interest for any Exactly. Or group of students. Yeah, we know that. So, Thomas Thompson, you mentioned the professional development piece and how sometimes people are trying to teach teaching methods that are research based, but they're using teaching methods that are pretty old school like lectures.
If you could redesign teacher prep from scratch and with sort of AI at the core of it, which, you know, we just saw open AI put $10 million into this this week. What do you think the priorities would be? How will teaching evolve and how will PD evolve in the age of ai?
[01:21:25] Thomas Thompson: Certainly, I mean. One piece that's worth pointing out is we have to like tear down the walls if some methods are off limits.
Most methods have some marginal effectiveness depending on the contextual factors. So like a lecture may be effective. Sometimes direct instruction, fully guided instruction is useful on the first introduction of a topic to novice learners where they need a lot of guidance, those kinds of things.
However, I'm not sure how much they came through in my teacher prep program like. I'm getting ahead of myself. So the, the question was about redesigning teacher prep programs for scratch and what that might look like in an ideal world.
[01:21:57] Ben Kornell: Sure.
[01:21:58] Thomas Thompson: I was very lucky that I had a mentor teacher in high school.
That teacher kind of pulled me into the world of teaching, made me aware of education as an activity and teaching as a craft. I shattered this teacher for my senior year. I made his copies. I stole all of his learning materials in a big binder. I still have it. I mean, it guides like the graphic organizers tool on Edge eight.
I mean, half of those little toggles you could use on those. They're all kind of ripped off from Mr. Claire's different instructional graphic organizers. But I would start there. You wanna start early? We're like, how do you attract talent to education? Well, students start in public schools or in some form of schooling.
Providing some sort of on-ramp through job shadowing or a mentor teacher or some sort of elective course, which is what we had at my school where I was able to do that. You want to identify talent very early and then allow that talent, those students, to steal the mission and purpose of teaching, right?
We can identify those dispositions of a good teacher, curiosity, empathy, clarity, and you can begin nurturing those earlier instead of later down in the pipeline when you got students trying to figure out what they want to do in undergrad or something like this. Then teacher prep programs. I mean, we should really center these around cognitive science as kind of the base discipline that's going to inform teaching practice.
I think a lot of teaching practice is built out of, there's a lot of myths floating around in education and fads. I mean, education is one of the faddish professions out there, and we like to repackage and repurpose old things and call them new. I mean, what's the difference between discovery learning, inquiry-based learning and self-guided learning?
I mean, they're all kind of the same thing, but we have different names that mean slightly different things. So you wanna get on the same page, you gotta rectify the terms. You gotta get everyone talking about the same thing. So when I talk about retrieval practice, someone understands, oh, I get what he's meaning, or spacing or modeling or worked examples, or what does it mean when I say I wanna offer feedback?
Right? We all have pictures in our head of what these things mean, but these pictures are not the same thing. And from context to context, you can have a wide range of quality, of course. Then you want to move into. Like really a practice driven curriculum. Get your pre-service teachers into the classroom earlier.
They should be observing teachers through most steps of the process. They should have one-on-ones with a professional teacher, and then of course you have to have a strong base in the content area that you're teaching. I mean, we throw around things like critical thinking. Critical thinking doesn't exist in a vacuum.
Critical thinking is. You can't connect the dots unless you have dots to connect. So you do need like discreet factual knowledge at hand to pull from, to like make connections. You can't just reason about a topic that you're not familiar with. So if I'm a history teacher, I should have a solid grounding in historical study as well as actual study regarding practice and education and then a firm understanding of how humans learn.
That's a tall order. I'm not sure that's something you can do in a four year undergrad. So maybe something like a fellowship approach, but this is an ideal world, like kind of getting away and onboard from reality. So if I had to narrow it down to three things, it's start early, identify talent, like wow, they're in high school.
Nurture those through like mentorships, job shadowing, have a firm grounding in cognitive science and an understanding of how people think, learn and develop. Then have a firm grounding in your content area.
[01:25:05] Alex Sarlin: Yeah. Fantastic. You know what I really love about that answer? You're combining the human relationship element of it.
The idea of entering job shadowing, you know, this almost apprenticeship that you're teaching mentor learning from expert humans, and then combining that with understanding, finding ways to get some of the complexities of classroom management, of learning science of. All the great instructional design concepts you drop there, right?
The retrieval practice and the modeling, the work, example, cognitive science. This stuff is hard to put all the pieces together, but AI can support with that piece of it. So it's like, and shadowing through humans, being connected to people in a really deep way, building human relationships and then supplementing that training with AI's ability to pull together, you know, all this research.
Mm-hmm. All these concepts and help you spin up. You can imagine that first year teacher making lesson plans and then the edge aid saying, oh hey, you might wanna add more retrieval practice here 'cause there's a reason why you could, because that's gonna improve their retention of the material. And you need factual basis of knowledge before you do this critical thinking.
I mean, that's a really exciting vision of sort of putting together the human and the ai. We talk about them as opposites. I think it's such a false dichotomy. Thomas Hummel, I wanna bring you into this. So, you know, as you are thinking about the Edge Aid product and how you evolve it, what do you think is the hardest problem right now for AI in education?
What are these pieces where you're like, oh, as soon as we push on this, we're gonna unlock all this new functionality?
[01:26:25] Thomas Hummel: You know, to be honest, I don't even know if it's a product thing, but it's a bigger ecosystem thing, and it's like. I'm not sure that teachers are always getting the tools that they need put in front of them because of X, Y, and Z factors.
Like I just think that if there was an objective view of the space on what these tools are and what they do, I think teachers are professionals. You know, they're not dumb. They have the intellect and the professional wisdom to choose what they need and to get it going with it. And I don't know if that's really how the overall space works.
I think that one thing that I've seen is that. If there seems to be like this narrative that if you can reach these teachers first, that they go with this thing. And I think that as we go down this AI path, I'm not sure that that is necessarily a great start for this AI and teacher Foundation is like the people who are the loudest and the companies that got the most funding are now in front of all these teachers.
And I'm not sure that that's actually good for the ecosystem as a whole. I, I think it would be nice if we had a governing body or an objective player saying, Hey, these ones are good because of X, Y, Z. These ones are good because of X, Y, Z. But we're trusting, we're not really giving the teachers the professional ability to choose what they need in that setting.
They're getting tools pushed in front of them saying that this is gonna solve all your problems. And I think that that just leads to more frustration down the line. And you know, as a teacher, I'm just like, it just doesn't really make sense to me seeing it from that side. If I could
[01:28:03] Thomas Thompson: jump in on that real quick.
I mean, I, I think you can boil that down to really like two main bottlenecks, which is verification. Like how do you ensure that the material that you're getting and going to put in the classroom, not only is of like, is ground in fact, which is tough to do when you're working with a, uh, a statistical model of language that is kind of working on a predictive system as opposed to something very deterministic and grounded, like in in truth air quotes that your audience can't see.
And then the other piece would be prompting, right? Yes. Even though it is a statistical model of language and like things like hallucinations, I'm not sure we can so easily like solve for that considering the core architecture of the system. But prompting is, is the other main bottleneck. Like there's a lot of great affordances that you can.
You can achieve and you can, you can access with an ai, but you have to be able to, it's a mirror, right? This statistical model is based off of the input that you provided. So if the input that you provided is kind of playing, I'm gonna use the word here, the language game of high quality instruction, you essentially get the AI to play that language game as well.
It has no concept of what high quality instruction is or what fact is. Yeah. But you can, I don't wanna say trick it 'cause you're, it doesn't have intentionality or, or mind or or brain or anything like this. But you can essentially put it in that frame so that the words and uh, sentences that it will sample from will reflect what high quality instruction looks like.
Uh, there's a lot of great organizations trying to do this already. I mean CZIA Chance Zuckerberg Initiative, they have these evaluator tools. I mean, they're essentially building out these great rubrics for verification using LLMs, which I think shows those two affordances at play. Verification and prompting.
You could do a lot with prompting for verification as it turns out.
[01:29:40] Thomas Hummel: Yeah, I mean, we mentioned that Gallup Poll, right? And you said that teachers are saving six hours. But for the last two years, everybody said it's gonna save them 10. And so now what are we gonna like? That's the foundation that everybody has been living off of, is that the AI's gonna save you 10 hours a week or plus and all these things.
But the reality is, is that the data came back at 40% less than that, and there's no repercussions for that. So it's like, what is next? If we come back 40% worse on student outcomes than predicted with these AI tools, God forbid that happens. You know? And we should treat teachers the same. It's not fair to teachers, honestly.
[01:30:16] Alex Sarlin: So many great points being made. And one of the other findings in the Gallup Paul that was interesting is that, and it's relevant to what you're both saying, 68% of the teachers surveyed said they are not receiving training on how to use AI tools. So I think to both of your points about how if there isn't training there and then these AI tools are just sort of out in the market.
They're all marketing to teachers directly. The loudest voices, the highest funding. Uh, a lot of the tools that are getting sort of head starts are ones that are just the most visible or have the most word of mouth and then suddenly teachers are using those. If there was more of an ecosystem of, Hey, let's actually evaluate different tools against each other.
Let's be able to show what they do, what they specialize in, I think that would empower the educators to be able to make more and more informed choices and get to the time savings they actually needs. I thought six hours was great, but you're right, it's under performance compared to the promise. It's a great point.
And then of course, the eventually to the student outcomes you need, which is exactly where we all want to go.
[01:31:09] Thomas Hummel: Yeah, no, and I'm not trying to be a hater on the time and the space, but here's the reality is that for the last 15 years, all the technology and all the promises that we've been telling that students and now it's gonna fix everything, has actually gone the complete opposite way.
And for us to have a technology like AI that's going to revolutionize everything, and we're gonna do the same thing. It's crazy. I mean, it's crazy to me. Let's
[01:31:31] Thomas Thompson: make one a little like jaded and cynical. You've been promised full personalization for the last 15 years, since it's always right around the corner.
It was MOOCs that were going to be full personalization or you know, and even way back when there were adaptive tutors that were considered to bring about full personalization prior to kind of large language models. So,
[01:31:49] Thomas Hummel: but what does that mean for me as a teacher, that means I'm going to learn a new system.
I'm gonna put my heart and soul into it because there's all these promises, it's gonna help my students. And if it doesn't reach that I'm not as good of a teacher as I could have been, if I would've spent my resources elsewhere and I'm not helping the students to the degree that I should be if I would've put my resources elsewhere.
And that's wrong. What I'm
[01:32:12] Alex Sarlin: loving hearing for both of you who have, you know, a foot in the classroom and a foot in the entrepreneurial space, I think you're speaking both languages and being very clear about, you know, we can't over promise with this technology. It's very easy to, the potential is very high, and I think everybody feels that.
But if the potential isn't living up to the reality, if it becomes yet another hype cycle, yet another unfulfilled promise for technology and education, we all lose out, especially the students. Exactly. No, I'm, I'm a hundred percent with you. The other thing I think we really share here is this desire to get the learning science.
And I love Thomas Thompson. You talk about the language of high quality instruction, how there can be this little, you can sort of talk back and forth with an LLM and both make it feel like you're doing something that will have great outcomes because it's wrapped in the language of high quality instruction.
But until we get to a place where LLMs and you know, the teachers who are using them all together are really understanding how the instruction actually works, which, you know, we're all working towards it. There is a little bit of a surface level piece that is happening here when these tools make lesson plans or make quizzes or make assessments or make, do feedback.
And if nobody is evaluating the quality of it, yeah. I don't know if we can actually fully get there. I think that's the next phase for ai. I'm really excited about it.
[01:33:25] Thomas Thompson: Yeah. I mean asking an AI to, because you have a teacher who is not so familiar with maybe effective methods or has a picture of them in their head, but can't quite express them and says, you know, I need a lecture on such and such topic.
That's not as great as saying like, I'm planning a unit of instruction on such and such topic and I need to develop mini lecture, fully guide instruction, however you wanna phrase it. I need to include questions for elaborative interrogation. So I need to ask follow up questions to my students. I need to ask self explanation questions.
Then I should have them, uh, have a brief review, maybe a low stakes practice quiz at the end. Like all of those pieces, you have to keep that in mind and you have to provide that to the LLM in order for that to come back. And then you have to specify like what those quality controls are. It's a lot of work.
I mean, and that's kind of where this middleware of edu aid, magic schools, school, ais, d it's, that's where we emerge in that we can essentially do that layer for the teacher. So it's like, if we're gonna do that layer, you better do a damn good job of that layer then because like. If you cannot ensure that you are bringing sound pedagogy to the tools, and you can ensure the teachers have an area to bring their contextual factors, their content, if you can bring those two together, you can create something powerful.
But if either end of that is lacking, the returns will be the. They'll not be as good as they, they ideally could have been
[01:34:42] Alex Sarlin: amazing and win the Gold award for using the term elaborative interrogation. I think you're the first person to ever use it on this podcast. So like one of my absolute favorite instructional design methods in mm-hmm.
I've used it and taught it for many years and nobody ever talks about it. And I think your point is incredibly important that currently. That is required in the middleware because it's not baked into the LLMs and because teachers are getting very little training on AI tools and the ability to do that kind of really complex prompting that you would need to get something high quality.
I wish we had more time. I really admire your approach to bringing that educator perspective. You say this can't yet be another hype cycle that doesn't pan out, that instructional design perspective of how do we make sure that that cognitive science and instructional design is at the core of this, and then of course the the actual technology.
How do we get in front of the most teachers and make sure that they are getting the value they need out of this? It's a really exciting perspective, so I have to wrap it up there. But this is Thomas Thompson and Thomas Humel, the co-founders of Eduaide.ai. Thank you both so much for being here with us.
Thank you, Alex. Thank you, Alex. Thanks for listening to this episode of EdTech Insiders. If you like the podcast, remember to rate it and share it with others in the EdTech community. For those who want even more, EdTech Insider, subscribe to the Free EdTech Insiders Newsletter on substack.