Week in Edtech 12/08/2023: Gemini, 2U's Challenges, Testing Overhauls, Veering into Politics and More! Feat. Gavin Cooney of Learnosity and Jared Grusd of Ethiqly

sponsoring it. So it will be a great celebration, a chance to let loose. It's a Friday afternoon. We're really looking forward to it. And we've also got some really exciting events coming up in 2024 that we're excited to announce. So stay tuned at our end of year extravaganza where we will outline some of that. Fantastic

interesting partnerships. And then right up for our Christmas episode, we talked this is one of my favorite conversations all year with Kai Frazier, who is doing Metaverse technology with her company, chi X are for all sorts of different kinds of students in all sorts of really wild ways that I've just I've never seen XR and VR use this way. So a couple of really cool conversations coming up by the end of the year, and then a whole bunch more starting in

January. So you know, talking to all of these innovators is just one of the highlights of my year. And I can't wait to go and review some of our predictions from this year and see how well we did with it. But looking back at the beginning of GPT as of January last year, and seeing how it panned out compared to what we thought would happen. I think I thought a lot of worse things would happen. And they did so be pleasantly surprised. Yeah,

fascinating. It's basically a gentleman talking to the AI and showing a series of pictures on post it notes and essentially having a fun engaging gamified conversation with the AI. Wow, like blow your mind. I will just tell you that here in the Valley at least. There's a big yawn. I don't think people are seeing a leap forward. Like we fell when chat GPT launch or even GPT four and let's also be realistic, like GPT four was not so much of

a leap from 3.5. But the fact that 3.5 had just introduced the concept and then four was like meaningfully better. Was like bam bam like open AI is crushing it and Right now what we have with Gemini is Gemini Nano, which can operate on your phone, it's kind of the, like limited level, it's more of a GPT 3.5 killer, which means not killer at all. And then there's supposed to be essentially the Pro and ultra versions coming out which, you know, in different testing environments

outperformed Chet GPT. For I think the challenge is one that actually many theorists have articulated that there is a diminishing return element here, like, how large does your language model actually need to be to deliver value, and you get to some ASOM total value, where larger and larger models or more and more capabilities don't actually add that much

incremental value. And so my take on the Gemini launch is really, what's profound and important is how is to see how well they incorporate Gemini into the suite of Google's products services. Google is the largest edtech company, as you've heard many times on the show. And that team at Google is chronically under resourced, underfunded, under support it it's a rounding error for Google, you know, the ad business has been their business

for a long time. But you know, cloud business, Chromebooks and kind of operating Android, like there's so many things that have really buried the education business in the stack. And, you know, this is the real opportunity for Google to actually find a way to elevate the education work, I think there's so many great use cases in that spot. And it takes advantage of Google's reputation for being whether it's earned or not a safer alternative to open

AI. So, you know, I'm really pushing for and looking for, how do they bring this to the G Suite? How do they bring this to Google Classroom? How do they bring this to Chromebooks? You know, 80% of schools are using Chromebooks as their one to one device. So how can they integrate it into all of this? That's really what's going to determine the impact on edtech. And ultimately, I think for Gemini, it's all going to be about integration with tools. It's never really going to get

the pop over open. Yeah, it's

pretty specific ways. I would say things like mid journey, and Dali did as well, partially, because we had just never seen anything like it before, the ability to do this kind of high fidelity, generative magic in response to you know, whatever you could think of, and then compared to that, it's going to be hard to have another, you know, see change, like blow your mind moment, this soon after, that said, the things that got me excited about the release.

And the things that get me excited about Gemini and Google in general for this one is what you already said, which is that Google is already in our world, everywhere, right? I mean, it runs our Gmail, it runs half of our apps and our spreadsheets at our messages and our it's ever. And I think that there's no world in which they won't start incorporating it into the G Suite at so many different levels for individuals and for especially probably for enterprise. And that does get me

excited. And as you say, it's potentially exciting. The spillover of that could be exciting for education, right? The fact that Chromebooks could use that the fact that, you know, it's able to potentially, you know, you click buttons in your Google Drive, and say, hey, you know, have I ever thought anything about this topic? And you can look through your entire Google Drive and say, yeah, actually, you, you wrote a note to yourself about this in 2013. And here's what it looks like.

Like, that's pretty exciting. The other thing that's exciting about it, and I don't know, I have not done enough homework on this yet, but I just I don't fully know how to read it. But it sounds like compared to the open eyes of the world, that Gemini was trained on super hyper multimodal content and designed to be multimodal, and I think that is something that maybe makes it potentially very,

very exciting for education. And I'll tell you why that is, we have been thinking so far until recently, you know, these tools were they would work in one modality and maybe there'd be a little bit of an overlap, like, you could ask chat up to you to write you a song and then take that song and put it into a music generator, you know, lyrics and put them into a music generator, and it could make music and then you can take that music and put it into a, you know, a visual generator, and it

can make a video for you. And like, that's pretty cool, right?

The fact that you can sort of chain these things together, but for a company with Google's technical expertise, to make a model based on multimodal training set that is designed from day one, not that it can do it yet on day one, but it's designed from day one, to be about multimode Well, the urgent talking about haptic feedback and things like that on it, that feels like that is laying the groundwork for potentially that type of, you know, blow your mind moment that we did have

with chat GBT. Because in a world where you can put it in the most silly possible way where you can ask it to look into your Google Drive, pull out every you know, graph that you've ever made. And every song you've ever, you know, saved to your Google music and make something out of it. That's multimodal, like, that starts to be something we've just never

seen before at all. You know, even like the silly thing you're talking about where people draw on post, its and it can read from your post, it's or Google has a thing where you hum into it and can turn your humming into a real song. I'm like, that is where it starts to get kind of exciting.

GBT. But Omni modal would be a world where your query allows you to go through all of the different modes, and produce an output that is in multiple modes, like, please go into my Google Photos and create a photo album for every single year since my son has been alive, and put a music soundtrack to it. You know, that kind of requests is possible, not because of Geminis technical superiority, per se. More because Google is

sitting on all of that data. And their Omni modal system is built to build connections across those rather than kind of discrete outputs. This reminds me a little bit of what we talked about last week where the kind of speech translation was going speech to speech and it

never went syntax, yes. Where there's like one, it was really powerful because of the modality but to the fact that it didn't have to go to text to then go back to speech, or, you know, there was no translation layer, to the text, the natural language processing layer that we were used to for a year, I think this idea of vision to vision speech to speech, but now adding it all together, you know, Omni modal to Omni modal. That's really exciting. Exactly. I

kind of creation journey. So you know, I think this is one of those where we're going to have to keep watching Gemini. But if you're listening Google people by the way, I'm trademarking

become the inherent player. Oh, and if you miss it, open AI is going to be the inherent player.

you need from anywhere? And I think there's something really interesting about what you're saying, in that context to imagine a future because Google, of all the companies Google is probably the most at threat their core business, it's probably the most at threat by AI. Because if they didn't do an AI play, and it's only open AI and anthropic and you know, meta and folks like that, then there's a serious threat to just the concept of using search to find information. You don't need

to do it anymore. But if you think about what Google has already put into their search, they have all the maps, they have all the images, all the videos, they're all indexed like a world in which in the future, instead of searching, instead of, you know, a text box where you type what you need, and it looks for it, it could be any kind of input you ask in any way you want. Even just, you know, a button somewhere next to a big question mark next to something

or just say what you want? Or who knows or haptics, swing your finger in a certain way. And he goes, Yeah, I can now make you any image, any video. And he texts explanation, I can personalize it to you. There's something pretty exciting about that. And it's certainly not

here yet. But I don't think it's such a crazy thing to imagine, you know, instead of searching for, I hope, a teacher out there made a great YouTube video explaining, you know, the theory of, you know, capillary action, you just say, What the heck is capillary action, and it makes things for you. That's pretty exciting. Yeah.

standards. In higher ed, the first article that we have is from Politico where we're actually seeing people from both the left and the right, really questioning state test scores and the validity and it's coming from two different perspectives. One is, with the pandemic and learning loss, we're just seeing incredibly depressed scores worldwide. Why are we spending so much time testing we need

more time teaching. And second is, these kinds of standardized tests have become this barrier to getting your high school diploma, when there's many other measures that could tell us

whether kids are competent. And from the equity lens, there's a sense of, you know, many of the standardized test set bias, and from a workforce standpoint, many of these tests are irrelevant to the skills that learners need for the future, throw into the pot, this kind of questioning of grades in higher ed, there was a recent report in New York Times that basically, you know, 95% of Yale students receive a or a minuses for every class on every grade. And it's like, you get an A, you're going

out. And, you know, I'm really curious, because part of me laughs at the story, because, you know, this has been actually a big joke for a long time, you know, the testing regimes have not really been effective, or in effect, I don't know, for six or seven years in most states, but to when you think about the conditions to transform education, one of the leading conditions to create that transformation is assessment. Because once you change what you measure, you change what matters

in the system. And if it was all about reading, and writing and math test scores, that's what the system is designed to deliver. But when you start opening the aperture and saying, actually, it's about real world skills, and demonstration of competencies, and and all of that, not only do you open the assessment playbook, but you open the pedagogical playbook to things like project based learning, and so on, all the while, we could never reliably

assess those things. Now we can because we have Gemini, and we have open AI. And we have all of these products that now can leverage unstructured data and turn it into structured data. We have all these systems that can do real time assessment, so that you don't need summative, you know, four or five weeks long testing regimes. Now you can actually real time test, leveraging AI, human lube, and know where your kids are and help them grow and progress. So of course, I'm very excited

about it. I think the counterpoint is, is this just another example of us lowering our standards so we can hit graduation rates. But I'm curious, your thoughts? Well, that's from a K 12 and higher ed standpoint,

difficult. And we've created these massive complex systems around them, where we've sort of created all these assumptions that you have to teach these subjects and in this order, and, you know, different in different states and blah, blah. And I have felt for a long time, like standardized tests, were one of the only things in the whole system that sort of shot out any kind of external accountability signal to whether We have any

idea what we're doing. And the cynical side of me, says that the fact that all of these big states are saying, You know what, these tests, I don't think they really measure enough. I don't really care at the exact moment when we're hitting like, you know, you lows. We're 50 year lows on those tests. Feels a little coincidental to me. That's the cynical side of me. I'm like, Huh, okay. Yeah, I guess,

know, instead of it being standardized old school testing, if we can figure out how to turn on structured data into, you know, what they sometimes call self assessment or embedded assessment, that's always been the dream, or even gamified assessment, where you're doing something where we can tell how well you know something, but you don't feel like you're in the midst of a high stakes high anxiety test that is obviously better than what we have. Now, that said, it's not like that

exists quite yet. It's certainly possible. But it doesn't exist quite yet. And the fact that you know, New York and Florida and the states at Louisiana, are starting to say, you know, what, these tests requirements, I'm not sure. It's not like they're saying, because of this amazing technology, or this amazing company, or this amazing set of assessments that just came out. We don't need these bubble sheets anymore, because we can just learn what students know from XYZ. There is not yet an x,

y, z, it doesn't exist. So I know, I'm usually not the cynical one. But like, the cynicism is winning out a little bit on this one, because even though I am bullish on the fact that we could have authentic assessment and we can build these, the solution, you're naming Ben, it's not there yet. And it's not like Florida is saying, You know what, let's not

do a bubble test. Instead, we're going to do this really amazing, high tech, beautiful solution powered by MLMs, where we can listen to students conversations with Teach effects and use the data to tell us exactly how they score on that doesn't exist. So until it does, I think there's I'm a little cynical, no, is it wrong?

bar around competency? And I think we're, I think most of the critics are having issues are around the sorting hat of like, you are in group one, you are in group two, you are in group three. And it seems kind of random, that it would be based on, you know, these 30 questions that only measure a very limited set of knowledge and facts

clinical standard of care? What are the new sets of measurements, and even some of those actually create indexes based on those clinical care standards. So like in nephrology, I happen to randomly know about that, but they take you know, eight to 12 metrics, put them together into a score of one to 100. And those component parts actually get updated by the science so that your score today isn't exactly comparable to your score might

have been 20 years ago. But it's this idea of it's not about absolute score, it's around directionally improving the health of patients. I think there's something like that to be done here. And given the failure of Common Core and given the controversy around No Child

Left Behind. You know, my guess is that the US will end up with four or five very similar looking assessment systems that optimize around different things, and that there will be consortiums of states that that go into those different assessment consortiums. But literally, you know, AI, it's the Wild West, we're starting to enter the wild west in assessment.

the schools. It's never about the students, if a school wants to say, we are an effective school, and you there's no data to point to, to show that you're an effective school. I just don't think you can say it. So to me, that's the saving grace of standardized tests. But there's lots of negative aspects of standardized tests, but getting rid of it, arguably, um, there's a quote in this political article that I that resonated with me from Louisiana actually about was like government sanctioned

mediocrity. So they called it the State Superintendent of cities, Jana says it's a government sanctioned excuse for mediocrity. I can't disagree with

have to be balancing that. Well, speaking of accountability, we also have some politics, this week's pod. As you all know, I'm on a school board. So often, the politics is in K 12, and school board elections and so on. But this time, we get a reprieve. It's now at the University President's office, we had the presence of MIT, Penn and Harvard, called into Congress and taking the task on their stances around hate speech and the protests around Israel Hamas

war. And I will tell you, you know, and we don't have the clips to play for you here. If you listen to it, regardless of whatever side you are on the issue. The tapes basically show bureaucrats hiding behind lawyer speak. And it is a really crushing and incredibly embarrassing show for some preeminent institutions, to have leaders who really are hiding behind like, precise legal definitions of freedom of speech of harassment of hate speech.

When the questions from the Republican congressional woman, what's her name's to pay? At least the phonic Stefanik? Her questions are very straightforward. It's the classic like, give me a yes or no answer. And then the other person is like, I will repeat the thing my lawyer told me to say it was just so tragic. And I know this is an issue that really hits a chord with

everyone. But if you're at a school or university, and someone is calling death to you, and your family, and all the people, you know, it's like such a clear like, yes, that's not allowed kind of answer. And so I think it's playing into these themes that we've seen around universities losing their moral and ethical standard, and the perception that they are moral and ethical institutions that really hold up the best of our society. They're really losing

some momentum. They're curious, Alex, you're so close to the space. How are you doing? Oh, gosh,

businesses, right? They are collecting humongous outsize tuitions from students and families, they are running humongous amounts of different activities and initiatives under their umbrellas. And, you know, they're big organizations, and they've gotten bigger and bigger as they've gotten more and more,

you know, administrators. And I think that what is a little freaky to me about these testimonies is that you've gotten to a place where the university presidents clearly feel caught between, you know, several rocks and several hard places, you know, one is that their students have very strong opinions about a lot of the things they're talking about here. And they don't want to be canceled or protested against from the student body. And the students have a strong standpoint and sort of nothing

to lose. They also don't want to be sued. They just why you see some of these legal issues. And they also don't want to just sort of end up on the wrong side of a political divide of a really angry, very polarized political system that we're in the US right now. Just sort of related to the student side, but it's a lot but different. So they sort of have nothing to gain right now, just as corporate CEOs often have nothing to gain by taking any kind of sort of moral risk of

any kind. And I personally am not, I do not like Elise Stefanik. As a politician, I've never had one moment of interest or admiration for anything she's done just That's my personal politics. But it was pretty weird to see conversations between her and you know, the presidents of Harvard and Penn and MIT, in which you sort of can't help but side with her, or at least I hate to say that, but you're like, it is pretty weird that they can't answer this

question. But they obviously feel like there is no benefit in doing any kind of public intellectual stance and everything to lose by saying the wrong combination of words and having either their own students or the trustees or somebody who has looked breathing down their neck, you know, throw them out at the same time, the head of Penn came out of this with the Democratic governor of her state, and the senators. Basically, everybody's saying she did a really bad job, and

they're really angry. And you already saw the board chair of Warden ask people to rescind, basically calling for her resignation. I mean, like, they can't win, but even knowing they can. I

you'd be so celebrated. And let's say they run you off campus for whatever, you know, student groups, trustees, whatever, you know, bold leadership, transcends like the positional role, I'm not sure that it would be so much upside for any of these votes. And there was only downside in the answer in the course that they took in those answers. Like, you know, what they were really hearing was, no one will pay attention to this, like you don't make any news. Yes,

female. And, like, I could imagine so many ways in which we'd want to celebrate the careers of each one, one of them is actually known for free speech. And so it was really interesting to me, I don't have a conclusion on it. But how the press is covering it, or how the questions came across? Would it be different? If it were you no white males? My answering those questions? Or would the lawyers have handled it differently? Had they had different gender or different power dynamics?

Interesting. I just think it's important to ask, and especially when you have most of the people calling for these people to resign or criticizing them being white males in the media stories that we covered the New York Times, in particular, it's all men criticizing the female university presidents have to wonder what's going on, you know, is there an element here?

now. Yeah, great point. Yeah. Which is part of why there's so much polarization is part of why the difference we've recorded on this show how the difference between the two political parties in the US like, it's like twice as many people in the Democratic party believe in college as an institution. It's really, really very different.

So even though a lot of the politicians, these male politicians, as you're saying that are criticizing the university presidents are mostly Democrats, not all but like, I think that there's a really weird dynamic in which these university presidents are coming out and I think they're trying their best to sort of walk a tightrope and not alienate the college age population, the entire college, like literally, and I don't know if there's any

winning there. I'm not sure if anything they said II might have sounded at the time, like a moral stance, and the next day instead of just headlines being that, you know, Bob Casey of Pennsylvania senator Pennsylvania is angry at at McGill. It would be that You know, the entire Penn campus walked out, and five professors and heads of departments resigned. That might have been the reverse headline really?

Palestinians. That's the headline news. When you read the actual article itself, I'm not so sure that it's as blatantly pro Palestinian as it is, like, let's understand other cultures and other parts of the world. So there could be a little bit of inflaming the situation through the press. But in microwaves, what we saw play out in Congress with the university presidents, we're seeing that tightrope being walked by educators and basically every context every day, yep,

Likud party and Netanyahu. And the reason why I think that was such a crazy shift, and you're seeing it play out, literally, in almost all of these headlines is that it tore a very, very tight alliance that had been true for basically the entire 20th century of the, you know, American left, and Jews, basically, I mean, for lack of a better way to put it, it was like, those two ideas were extremely overlapping, for most of you know, the entire history of Israel and most of the post

war era in the US. And when Israel sort of went into this really right wing place and became very close friends with the American right, and it created the sort of this underlying tension that people were paying attention to it certainly saw, but people who weren't paying that much attention to it didn't sort of really think that much about it. But all the surprise that you're seeing from all of these

senators are all of these. I mean, you saw Sheryl Sandberg put together a whole testimony this week about the sexual violence during these attacks. And I think you have this whole generation of older, it just totally generational, you know, who just do not realize how fully that Israel has lost the young American left, like they're just literally lumped in the same basket as the worst, you know, quote, colonizers and oppressive states in the world.

And I think everybody older than 30 missed that memo completely. And they're seeing it happen here. And it's just blowing people's minds how far apart people are. And when you see these Oakland educators, you know, hello, you know, Palestinians supporting things in various ways. It's like, yes, because Oakland is one of the national centers of the progressive left and the progressive left knows where it stands on this. It's just, it's

really scary. Frankly, from my personal perspective, I'm sort of right on the I think, you know, you and I are born one day apart. So we're right in this middle, this middle generation here. But like, I find it really freaky. And I think it's a little sad that people have not realized until these crazy events just now how far apart those two sort of mindsets have shifted.

of people. And they don't feel also like places that encourage free speech for another group of significant number of people. And so if we were, you know, another podcast, we'd be talking about how universities are under constant pressure, because their ROI and their tuition and all the kind of business side of things are really creating a lot of pressure in that business model. It also is worth pointing out that these cultural tensions are creating pressure on the

socio cultural model. And those things combined together could be the kind of one two punch that brings down the kind of 50 years of building industrial colleges and universities. It could bring down that system. And I think we've also covered it it's not a one size fits all Not everyone's being brought down. But this kind of sentiment, further erodes the trust and confidence in higher education. So, you know, I know we're running low on time, and we want to get to our great

guests today. So we're going to transition to a more hopeful story about a leader who really has been that catalytic and transformational figure. And I will just say, you know, for every set of tensions and moments of like destruction in our system, there's also so many moments for optimism and hope. And so I'd love to give you the mic and just talk a little bit

about Paul Leblanc. And, sure, his work at Southern New Hampshire and what that has meant to the space and a little bit of, you know, 20 years ago, when he started to where we are today, the impact that he's had, yeah,

of a very small state. And what Paul LeBlanc did, really, I wouldn't say fully single handedly, and he certainly wouldn't say this, but he was really a leader in the space. He swam very far against the traditional current of university thinking for a long

time. And what he basically did is he took a university that is a regional university, Southern New Hampshire and say, we are going to revamp the way we think about education so that we are going to have fewer majors, we're going to go competency based so that people can actually prove what they know, and take tests more frequently, if not, basically, whenever they want. We're going to create a subscription model for tuition, where people pay for the amount of time they're in the school.

And when they're registered as a student, they can actually take as many classes as they want, and as many tests and get those competency based. I mean, he questioned so many aspects of what you know, we consider a traditional college environment. And he did it in the name of all of these non traditional students. And he said, of course, we're going to do it

online. That's why there's an edtech story, right, he said, we're going to be able to deliver education to people far afield of Southern New Hampshire. By doing it online, he really, really changed. Basically what people's conception of a university could be, and was ahead of the game on pretty much each one of these decisions, the idea of lowering tuition and capping it and keeping it from spiraling out of control is obviously exactly

what students need. The idea of focusing on non traditional learners who often who are older, I believe the average age of a Southern New Hampshire University student is 30s, you know, somewhere in the 30s, way ahead of the field on that. And competency based, which, you know, there's a big competency based movement for a while it has not quite taken off the way that many fans have it hoped it would, it still may come back, but just the idea of being able to sort of very pace and

recognize prior learning. I mean, these are some of the ideas that now are considered, you know, they're still considered revolutionary, but they're considered like the things that everybody knows should happen back then they were considered absolutely out of nowhere. And it turns this regional university into an

incredibly big success. And there's a handful of what they've called, you know, mega universities, in med surg and other higher ed publications, that are places that have really sort of just scaled beyond what you could imagine a university could serve. And Southern New Hampshire is one of the biggest

of them. And not only Southern New Hampshire, one of the biggest of them, it's one of the biggest of them without falling prey to some of the sort of predatory practices that we've seen with for profit universities, which were the first generation of these sort of mega universities. So he basically put a flag in a new land of what a university could and should do. And it was wildly

successful doing it. And, you know, as with many things in university life, I will feel like we've been picking on universities a lot recently, but he basically he and his team and his group really showed away I mean, you probably all see in Southern New Hampshire University commercials on television the same way you see commercials for W GU, in a similar model or for some of the for profits. They just completely changed what it means

to serve students. And I'm pretty confident he will go down as one of maybe 20 People in the last 100 years who truly, truly changed how universities were perceived. Probably more than several of the college presidents were just mentioned. You know, in the last segment, I had the chance to see Paul LeBlanc speak at Coursera. We invited him as one of the speakers there because Coursera was dedicated to sort of access and lowering the cost of education, of high quality

education. And he came and spoke there and talked about the Southern New Hampshire story and it's just like I mean, they were changing, you know, old textile mills and things in Southern New Hampshire into these incredibly modern, innovative centers, they were hiring huge numbers of marketing and sales people as well as the Student Success advisors. They weren't because of their model, they have a huge like a ton of people who are out there helping students succeed, because their students are in

their 30s. And they're almost all working. And they have so many opportunities to fall off the path. And so they invested heavily in that. And it's like an alternative universe of what universities could be. And hopefully, everybody, not everybody, but many people have sort of come around to see it as a really viable future for what many more universities can and

should be. So He is a legend in the higher ed space, we should see if we can get him on the show to talk about this, because hearing from Him is no substitute for my little testimonial here. But it is sad that he's leaving. But his legacy in higher ed as like, as a concept can't be overstated.

less. And that by being a bold leader, yeah, you may end up you know, like superintendents that I know that I most admire, they acknowledge that they're always on the edge of getting fired. Because if you're really pushing to do what's right for kids and educators and transforming the system, that's an acceptable risk. But I think that the degree to which someone like Paula bonk at Southern New Hampshire University, he didn't have to go climb some ladder and find some, you know, prestigious

perch to make the change. He made the change where he was. And he led with integrity. And by the way, I also had to say he led with values, not necessarily with the business model, the business model came, because they said, how do we actually reach the needs of professionals and learners and so on. So some may bemoan the kind of Southern New Hampshire era that we're in where colleges and universities are online and people of all

ages. And maybe there's a nostalgia for, like, the quaint campus life of universities in the past, that was always a limited access model. And really what he was driving from was a full access model. And I think it's even more relevant today than it was 20 years ago. So just when you're listening, like, sometimes you lose your job, like, go be bold, go make the change happen from where you are, and long term like good things will happen. I always when I go back to Stanford GSB.

I always tell them like, actually, the more risk I've taken has always been the least risky path. Because at a certain point, no one wants somebody who's gone every single step of the ladder. They want somebody who's bringing in divergent views from a wealth of diverse experiences. And so I just think this is a great career lesson to

Alright, everyone, we're excited to welcome Gavin Cooney, our guest today we can ed tech, CEO of Learnosity. Welcome, Gavin.

education. Last year, we delivered 17 point 9 billion questions online. So these are kind of massive numbers. And that's by design where the idea is aware of behind the scenes. We're providing these tools to let other people build products.

But November, last year, the open AI came out with Chuck GPT. And there was just a leap forward. And there seems to be an AI companies or products that were in the before times. And that kind of came along after the GPT. And we decided to go kind of big on this area, the first product we've just launched in the last week or so. And that's called author aid.

And as the name implies, it's there to help authors create content, that 150 customers are creating content in our platform and storing with us, you know, in their own databases that we manage for them. And we have a 337 million questions in there. So it's like a third of a billion questions in there. So it wasn't such an obvious area for us to focus on to make that authoring experience so much better, so much more efficient.

And really, the whole idea here is that authors can create much more content, but also much better content. So we brought in stuff around creating feedback for every distractor a feedback for every incorrect answer, and correct answer these learning opportunities within every single question that was always possible to author but was really financially impossible, not really viable to create all that feedback and all the constructor rationale in every single question that can be

created via AI. So not only are people creating more content, they can create better content as well. And, you know, we've been testing this with a bunch of our customers, the feedback so far is that, at least anecdotally, people are creating about 10 times the content and the same time. I mean, think about that creativity it takes to I want to create a question on the assassination of JFK. Okay, fine. What do we think of

where do I do? How do I do? And what are the options, what a viable option, as an alternative, and so on, it just takes all of that and automates it, and gives a very good first draft of a question for an author to continue with and edit themselves. The other thing it does is it takes feedback. So let's say in my example, you say, when was JFK assassinated? And the four answers are 1963 6465. And 1978 78? Just seems

like the wrong answer there. So you can kind of put in that kind of feedback, remove the 1978 option and give me something more likely, I will go do that. Give me the same question as a 10 year old reading level, give me something more analytical, give me something at this depth of knowledge level for this Bloom's level. So it's incredibly powerful. I'm incredibly excited about it. We only launched last week.

sometimes general knowledge. One thing we do is allow our customers paste in kind of chapters of text, from their learning materials, transcriptions of videos, that kind of stuff, and can create multiple questions from that. I see a paste in the chapter about plastics. And I'll give you a bunch of quick questions about plastics. And that's really, really effective for creating great questions straight off the bat. You also do things like you can paste in the curriculum

standards. So I mean, in my demos, I have like the Florida curriculum standards sitting in a tab. So I go in they paste in curriculum, a standard standard around opportunity cost. And there it goes, a creative, perfect question. For that curriculum standard, so it's really a tool to allow the experts, the authors out there who are already out there creating day in and day out and creating great content, and let them do that much, much quicker.

And let them use their own judgment and their own years of experience to decide if these are valid questions, and so on. Of course, then it can go through a process where you send that out to 1000 students and see if see if it's still statistically valid as well, for more some of the use cases.

experience. It's great to kind of under underpin that so much to that. And you know, we focused on formative assessment, mostly. And the difference between formative assessment and summative assessment is pretty obvious. But the way I like to think about it is a chef in a kitchen tasting the food, that's formative assessments, and the customer eating the meal. That's somewhat of assessment. Obviously, we do both. But the formative is so important. It is part of it is part of learning.

And I think AI brings some

threshold. Where do you think it heads over the next four to five years and what's on your mind and roadmap as you think about the potential for AI, intersecting with learnosity is core strength in assessment.

really innovative. And what we're working towards, I'll give you a little bit of sleep is the preview of those things. Firstly, is we've built an engineer that will check for diversity and inclusion in items. So we've put in all the guidelines we could find we wrote our own set, we found a bunch of stuff on the internet of different guidelines on what can you can use what you can't use. You paste in any question, and the AI will come back and give you a critique of the question based on those

guidelines. So one question I've been demoing is the question is, a nurse has to put in a saline drip. What should she start with? And it goes back and says, Look, you're unnecessarily gendering the nurse. And why does the nurse have to be female? This question is flagged for human review. Obviously, you put in something outwardly bad, it will say this is a bad question. Or if it's okay, and it doesn't, it doesn't step in any oil coals, it will say that's fine. And just use our

correct language. If you're talking about a child with Down syndrome, you see it that way. You don't say a Down's child, just all these little things to try to make it more inclusive, obviously, then we're looking at how to do representation. So an item about a John or Gavin is totally fine. So long as it's coupled with a bunch of female names and a bunch of non English names. That's, that's all fine.

So the idea of going through an item bank of say, 100,000 questions and analyzing them and say, you know, 98,000 of these 100 questions has a male English protagonist, what's with that, and looking at the representation within that, that's something that AI can do in seconds, which is just incredible.

already. Actually, there's another customer who does a bunch more than that, including that figure. So it's maybe 150 million essays a year, which is mental, and how much time teachers are spending grading those audio questions as well. We do about 7 million of those. So all the dictation of that, right now teachers are listening to 7 million audios. And here's kind of an Uber scenario here. What I mean by that is, I use

more Ubers. And more taxis than I did before it was Uber, because the very fact that I have Uber means I don't need to rent the car, I can decide I can use an Uber or taxi at any point. So I used to go to San Francisco, loving his rent to Kerala have the pocket and drive around and so on. Now, I can thrive there, I can get an Uber go to wherever I need to be safely all the time I do more

taxis because Uber exists. Well, I think teachers will do more essays, because essay scoring exists, I think teachers will do more audio because audio dictation and eventually audio scoring exists. So I think it has a massive impact on what we're doing there. A few real things were examining is multistep mathematics and how we give feedback to a student of

where they made a mistake. And why there are wrong with their calculation lovers, you can do that if you show them multiple staff, and you have AI in there. And one of the things that I'm particularly excited about is I spent, I've talked about essays just now but why is it an essay? Why is it not a conversation? So I built some stuff. I'm not sure if you saw it on my LinkedIn over the last week or two. JFK was assassinated 60 years ago this week. And I built a

question in here. That was a chat that asked a bunch of questions of the child around this general knowledge topic of JFK assassination. And if you think about if you ever tried to ask a teenager how their day in school was, and they'll say it was fine, and your first 15 follow up questions, how did you get there? What did you eat? What did you learn? You learn it and interesting and math and interesting, the geography, and

so on? Well, once you put in a rubric, and there you can have the AI ask those follow up questions in context with how the student is answering. And it could be a massive step forward. In my example, it would say, Well, what happened to Lee Harvey Oswald and I'd say he was shot on live TV by a fat little man with a hash. And they are Oh, you mean Jack Ruby. And they will give you that feedback. And it tells me that I didn't know Jack Ruby's name, but it will give me that feedback, which is

amazing. So I think there's some really, really interesting things coming down the line for how to be used. Yeah,

tools. And, you know, it sounds also like you're embracing human the loop elements where, you know, the author might actually be a person working at curriculum company, or it could be the person, the end user who's using that curriculum company's stuff. How do you think about leveraging the AI in concert with your partners who are also thinking about AI and assessment? How does that play into your approach? Look, from

education. And I didn't have I just physically did not have the money or the resources to go out to sell school to school, family, family, parent, a parent, whatever, it just wasn't gonna be possible for me. So we decided to do what I classify as a leveraged business model, this idea of selling something b2b, and let our customers and sell it on. I saw a very, very interesting article recently was somebody they were talking about, like, what how would you fix education with $100 million?

What would you invest in? It was just so far from what we had, we didn't have $100, let alone $100 million. So we needed to build something that was self sustaining, that was cashflow positive. And we'll just kind of would say a lot on its own wins. So that was something we did. And we just built these things. We figured if I could sell to one person was almost as hard to sell a million dollars of software and it was a sale 100 hours of software going into a

school or something. I just didn't have the resources to go with kind of door to door school to school to do this. So we built something so that those bigger companies in bigger deals. And now we have this big company with over 150 clients that are nasty API's. And that's a big company doing a very, very large impact in education. And what I'm most proud of with that, I mean, this all started by me wanting to make an impact in education. This was me offended by bad ad tech and

wanted to go and fix it. So we now have these 40 million users, which I never would have reached any other way. But what's more significant than that is the certain features that we have made standard. So if you go to, I don't know, some conference for teaching mathematics, and you're there and you're in, there's 10, products, 12 products are something that have been built on learnosity. And if you're proud number 11, say, Well, sure, I'd love you to come

and license learnosity. And I'll take your money all day long. And that's great. We're neutral power, anybody. But if you're not doing that, you need to build a product every bit as good as what you would have

built. If you had a handler acity, we've made certain things standard that now includes AI altering will include all of the other things I just mentioned, but so far, math scoring and different question types of teacher authoring and that are reporting matters, scalability, accessibility, those kinds of things. These are things that we made standard, and are no table stakes in the education

industry. So I feel I've personally raised the bar and the industry has had to rise with the tide of our innovation. And that's the leverage impact as well as leverage business models to leverage impact that we've had in the industry, which I can really be proud of.

learn already have channel. So now, when you layer on AI tools on top of what you've already got, incidentally, you can make that available to 40 million customers. But I do think that we're often under leveraging those things. And many of the big players, you know, Pearson amplify and curriculum associates and Houghton Mifflin Harcourt, they don't want to build everything themselves, they do acknowledge that it's

built by partner. And there's many cases where partnering with somebody like learnosity, I don't know if any of those are your customers. But you know, partnering with somebody like learnosity, accelerates their development timeline, reduces their costs, and then often results in long term value. So

that model. It was actually it's very hard to draw something out on an audio podcast, but it was something that I thought very well articulated was intercom as an Irish company that do some marketing software, and they had a blog post about picking your business model, they showed a two dimensional matrix by two by

two. And on the y axis was kind of lifetime value of a customer on the x axis was the customer acquisition cost for the self service to a consultative sale, we were up in the enterprise space, I can fly to wherever San Francisco to go make a sale, because it's a long term deal, probably a bigger sale. But what happens to a lot of ad tech companies is they're in the bottom right hand quadrant, which is a consultative difficult sale for relatively

low lifetime value. Something that in a consumer market would be consumer, say like something the price of Netflix or something half price of Netflix, you can't afford to go door to door. So Netflix, it does not work. So no, it doesn't work in education either. And that's one of the difficulties with this market.

incredible retention. And I think a lot of people and this is also part of the learnosity story that you may not want to toot your own horn, but the degree to which those customers you've retained is also like a big part of your success is that it's never been a leaky bucket. And kind of each year, you've been able to kind of climb the ladder. When you have like four year olds as your customer, they're going to age out the kind of retention there is

super, super hard. So aside from private schools, which have, you know, 14,000 You know, $30,000 a year price tag, that's a pretty good LTV. You know, you get down into the four and $3 You're just you're dealing with impossibility. As always Gavin It's so inspiring to talk to you. I just love everything that you're doing at learnosity I know our listeners are gonna want to learn more. Is it learnosity.com Is that the best place for them to find out more? Yeah, Snubby

for them. And we see that there's been a systematic and disproportionate impact on certain student groups. And so we've really thought about creating a platform that not only uses AI to improve writing skills, but also to allow many different groups who felt left out of the system to feel excited and empowered to get involved in their writing education.

parts of society. And what we thought to do is provide the sort of ethos and sensibility of some of those great consumer product companies, and the power of breakthroughs and technologies such as AI, but not limited to AI, together in the education space, that we could create a platform that's hyper engaging, like the best consumer products, but also super useful to not just students, but also to teachers across the country. So

educational spaces. We don't believe that AI should displace learning, nor should it displace the relationship between teachers and students. Instead, we wanted to create an enabling platform that brings teachers and students closer together, that brings students closer to their learning journey and to have fun in the process. So we use AI to help but not to replace any aspect of the learning process.

writing. But in order to solve the problem from the student perspective, we began with what I think the most fundamental problem and writing is certainly the most intimidating problem, which is the blank page. Even if you're the best writer in the world, like staring at a blank page or a word processor is just really intimidating. Where do I begin? And even if I have a rubric, even if I know what I'm supposed to do, just getting going is really, really intimidating for most students.

And so what we set out to do is to create a platform that doesn't begin with the process of writing. It begins much earlier in the journey, and

begins a notetaking. It begins at ideating, a begins of brainstorming, it begins at outlining, it begins at like trying to piece together topic sentences, putting together paragraphs, and what we tried to do is build a suite of tools at products and platforms that help students do every phase of the journey so that by the time they get to the blank page, they can have a lot of confidence that what they're about to write will make sense and sort of reflect not just the content of their

ideas, but also their voice taking into account their own backgrounds.

to do? The most reassuring thing is that there's huge demand across school districts in the country to have a platform that can help facilitate writing in a responsible way that's ethical and equitable. So the core hypothesis of the company seems to be validated, people love using the product, something that we under anticipated are some of the more basic blocking and tackling, which is its inner connection with other learning platforms, how signup process

works. And so some of like the basic blocking and tackling of how platforms need to be integrated with existing technologies that are distributed in the schools, we knew we needed to do. But we realized now, it's actually a rate limiting factor. And so we over the course of the past month have been really building towards those Integrated Technologies, which is a great problem to have. Because in solving those problems, we can get to the core of the value we're trying to provide both

teachers and students. And we're like really delighted with the sort of impact that we're having in those 300 schools and just the excitement around what we're trying to achieve. Awesome.

all. Our approach is schools are here for good teachers are here

for good. How do we leverage technology to make the experience of teachers more delightful by freeing them up to focus on the thing they care about most, which is teaching, and also how to make writing and communicating something that students look forward to doing and I think technology is going to play a role in rebirthing, the value of writing, currently, we have a view that writing won't be a skill that's valued in the future, because there are such great AI tools that can

effectively take your ideas and write the paper for you. But that's actually missing the mark. Because while there's a lot of value to those kinds of products and tools, fundamentally speaking, I think the ability for human beings to express themselves to communicate, both orally and in written form is a uniquely

important quality. And I think by building a platform that sort of treats writing in the way many math teachers teach math, which is like show your work, because how you think how you process ideas, how you construct them, is the best way to express who you are not just as a thinker, but as a person, I think really, really matters. And so I think we will see a great research agency in the value of writing over that

period of time. And of course, AI needs to play a role in facilitating that process, too. Yeah,

math. So we're having this renaissance in part because for educators to deliver on writing instruction and assessment, we have a new technology that's enabling it. So I'm so excited that ethically is on that forefront, that cutting edge. If people want to find out more about ethically where can they go, that would be great

community. For those who want even more Edtech Insider subscribe to the free edtech insiders newsletter on substack.