Edtech Insiders
Edtech Insiders
Sizzle AI and the Rise of the AI Study Companion with Jerome Pesenti
Jerome Pesenti is the founder of Sizzle AI, an AI-for-learning company whose mission is to make learning amazing for everyone. Jerome has worked in AI for the past 25 years, including as VP of AI at Meta, co-CEO of BenevolentAI, VP at IBM Watson, and co-founder of Vivisimo (sold to IBM).
💡 5 Things You’ll Learn in This Episode:
- How Sizzle AI uses bite-sized, adaptive learning to help students study more effectively.
- Why Jerome left big tech to tackle educational challenges with AI.
- The evolving role of homework and assessments in the age of generative AI.
- How AI can personalize learning paths and study plans for individual learners.
- The future vision for AI as a lifelong learning companion across school, work, and personal interests.
✨ Episode Highlights:
[00:02:21] Sizzle AI turns study guides into personalized learning paths.
[00:03:10] Jerome shares why he left Meta to focus on AI for learning.
[00:05:45] Test prep mode boosts readiness with smart study plans.
[00:06:51] Students love Sizzle for making hard content manageable.
[00:08:18] Gamification features are coming soon to enhance engagement.
[00:09:16] Why Sizzle chose a direct-to-learner approach over schools.
[00:11:00] AI shows steps, not just answers, to promote real learning.
[00:14:36] AI can assess student readiness, not just complete tasks.
[00:17:29] Future of testing: both AI-assisted and AI-free formats.
[00:20:14] Sizzle already supports learning beyond the classroom.
[00:23:04] The role of personality in AI study companions.
[00:25:03] Using data to drive learning science at scale.
😎 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] Jerome Pesenti: What we try to do in our case, if we try to not give you the answer right away, we try to show you the steps, have you think through the steps as you go through it? But ultimately, teachers have to know if the students want to answer, they can use GPT. They can use Gemini, right? They cannot figure out a way to get it.
So you have to rethink homework in that new context. But ultimately, I think it's like everything. It's like the calculator in the past. These tools can be well used and can make us all more efficient. They can make us lazy. You have to kind of change your curriculum to make sure you adapt to that.
[00:00:36] Alex Sarlin: Welcome to EdTech Insiders, the top podcast covering the education technology industry from funding rounds to impact to ai developments across early childhood K 12 higher ed and work. You'll find it all here at
[00:00:50] Ben Kornell: EdTech Insiders. Remember to subscribe to the pod, check out our newsletter, and also our event calendar.
And to go deeper, check out EdTech Insiders Plus where you can get premium content access to our WhatsApp channel, early access to events and back channel insights from Alex and Ben. Hope you enjoyed today's pod.
[00:01:16] Alex Sarlin: Jerome Pesenti is the founder of Sizzle AI. An AI for learning company whose mission is to make learning amazing for everyone. Jerome has worked in AI for the past 25 years, including as VP of AI at Meta, heading up the company's AI research and applied AI teams and as co CEO of benevolent ai, a British technology company using AI to accelerate scientific discovery.
Before that, Jerome was VP at IBM Watson, where he led the creation and development of Watson's AI cloud services. Jerome joined IBM after it acquired Vivisimo, a tech firm, specializing in language processing and search for the enterprise, which he co-founded in 2000. Jerome Pesenti welcomed to EdTech insiders.
Thanks for having me, Alex. I'm really excited to speak to you. So we've covered Sizzle as you've been raising money and building out these really exciting AI products. But for those who may not be familiar with it, let's start by just talking about what is Sizzle AI and what brought you to it? What made you decide to go into learning after all of this amazing tech experience in all sorts of different fields?
[00:02:21] Jerome Pesenti: So Sizzle today is an app that you can find in the App store or the Play Store. You can download it and it really is meant to help learning either for homework or when you prepare for a test. And I would say the most distinctive functionality is that you have a test tomorrow. I. You get a study guide from your teacher.
You take the study guide, you fit it to sizzle, and it will give you a course of the things to do to get where you're ready. And it feels a bit like Duolingo. It's like bite-size little activities that will get you to the level that's gonna be better for acing your test tomorrow.
[00:02:55] Alex Sarlin: You've had such a big background in ai.
You've built AI at IBM Watson and at Meta. What made you decide that AI for Learning was a really great use case for it and the ability to help people study and learn with ai? What brought you to that idea? I.
[00:03:10] Jerome Pesenti: I was, you know, almost five years at Meta. And obviously Meta has amazing AI and I had an amazing team there doing things like optimizing how reels work or ads work.
And we were really good at making people use the software a lot more, you know, and stay on the platform, stay hooked to it. And what I always wanted to do is how can we use that power to keep people engaged for something that's clearly valuable for them? Yeah. And learning is something people feel strongly about, you know, they feel like brings them value.
So I wanted to use these techniques. When you bring them bite-size content, how do you optimize that bite-size content to make them learn, not just watch videos.
[00:03:49] Alex Sarlin: And does sizzle tend to focus on K 12 students, students in in high school or secondary school, or do you see a lot of college or do you see learners sort of across the spectrum?
[00:03:58] Jerome Pesenti: I think right now it's targeted toward independent learners, and it's mostly, I would say 15 to 25. So high school and college. Yep. And the idea is how do you make a learning more bite size, easier to consume for an audience? Gen Z, gen alpha, who has been used to consuming things in small increments, you know?
[00:04:16] Alex Sarlin: So what does a bite-sized experience on sizzle look like? If somebody was studying for, let's say, a physics test and they uploaded their study guide and it made a whole course for them, what would one bite-size unit of that look like? Would it get pushed to them for the app? Yeah, I'd love to hear when the system now is actually
[00:04:32] Jerome Pesenti: based on learning science.
So when you upload your study guide, you will parse it and figure out all the key knowledge components in it. And the skills that you need to acquire to basically master what's in that study guide. So it creates a curriculum for you. It tells you, okay, here are the skills that you need to acquire. And when you click on one of these skills, it gives you little bite-sized exercise.
Think of similar to what you'll get on dual lingo for language. But they are obviously for physics where. It lets you kind of learn the concept, like fill in the blank, multiple choice, and it tries to give you basically practice each of the concept you need to master with multiple different exercise and it's a stream so you can keep going at it.
You can skip whatever you want and we'll keep generating new ones.
[00:05:15] Alex Sarlin: Yeah, and so people, they could do little bits at a time, but they can stack them. They could do a hundred of them the night before the test or they could do the, that's right, bit by bit over course. We try to
[00:05:23] Jerome Pesenti: encourage you. We have also a feature that gives you a study plan, so you know, if not just a crammer, you can come three or four days before and you say, okay, you have four days.
You have an hour every day. This is what you could do this day. This is what you could do the thing after this go down there. But some principle, a stream, a little bite-sized exercise you go through and then you see your proficiency go up in each of the skills.
[00:05:45] Alex Sarlin: Yeah, that, that's really interesting. So you have been in AI for a long time.
Is this idea partially made in conjunction with the rise of generative AI and LLMs, or is this something you've known that AI could do for years because of your experience with Watson? Sort of take information, break it into its knowledge components, and then organize new content around them? It's actually both.
[00:06:06] Jerome Pesenti: So I knew from my time at Meta that you could optimize a bite-size experience, and that's the thing that we want to bring to learning. But there's no question that the LLM. Make creating content learning much easier. So what's nice? The ability to generate almost an infinite stream of exercise that can be more adapted to you.
And then we use the what we knew to kind of optimize an experience for each user, you know?
[00:06:33] Alex Sarlin: Yeah. So I'm sure you've heard from lots of users about their experiences there. I'd love to just hear you talk through, is there a story or two that stand out to you about individual? Learners who have reached out to you about what Sizzle has done for them?
Or do you hear sort of trends throughout your user base about how it changes the way they study and think?
[00:06:51] Jerome Pesenti: Yeah, I mean, what's fun is to read the reviews of the products. So like I follow every week we are on rotation, you know, one of us on rotation, on support to answer this week is actually mine. And I've been blown away by the reviews that our users, I really advise you go on the play store, on the app store and see what users say.
And what they say is sometimes they go to a class and it's complicated and they get complicated material and it's long. It's a whole chapter of a book. And what it does is the system like divides it in little bite size and it's more fun. It's on your phone, they're used to it, so it's easier to digest it.
Easier to use. It gets them to do something, and it ultimately gives them a better grade in less time than they had before.
[00:07:29] Alex Sarlin: Yeah, which is exactly what people are often looking for from their learning experiences. But you mentioned Duolingo a couple of times and Duolingo, it cuts things into bite-sized pieces.
It also does all sorts of. Psychological game-based mechanics. They use streaks and coins and maps that unlock and all these, these different things from gaming. Do you do any of that currently in sizzle or do you expect to do that to even increase the engagement factor and the sort of fun factor even more?
[00:07:54] Jerome Pesenti: Yeah, I think we have some of it like vertical, the core game dynamics of having lessons that you uplevel each time. Yeah. In order to see the type of exercise and the type of experience like. Very mobile centric and very easy to consume. Yep. Uh, we haven't implemented all the gamification, and that's something we're working on like streaks and needle boards, these things that work really well for Dogo and we intend to implement that shortly.
[00:08:18] Alex Sarlin: Yeah. Sizzle is obviously a direct to consumer. And you're working directly with learners who have their own motivations to learn or sometimes, you know, upcoming exam. We talk to a lot of people on this show who are working with AI that they wanna put into classrooms for educators. Mm-hmm. Or for educators and students.
And there's always something they have to grapple with. Educators not knowing, some being incredibly excited by ai, some being very threatened or unhappy about it. You sort of bypass a lot of that, which I think is smart. I'd love to hear you talk a little bit about. Your experience. You've worked in consumer technology a lot.
Tell me a little bit about how you think as a B2C, your end user is your buyer and is a student. Do you ever think about some of the broader issues in AI and education? Like people worry about academic integrity, people worry about people outsourcing their thinking to an ai. Ai, does that affect how you design, or is that sort of all noise because you're just trying to serve the end user?
Who is the learner themselves in this
[00:09:16] Jerome Pesenti: case? I do think these problems are important. Right. And I do think the new generation of AI is kind of thrown a ran in the educational system. Yeah, it's good and bad. That means it will change things and it's creating some headaches, but for the most, I think it's gonna be very good.
And what we're trying to do with Sizzle is. How do you leverage that to get people to do some real learning and more effective learning by going with B2C? We don't have to worry about all this. Really. We need to create a tool that students will wanna use, and I don't have to convince school districts or others, which are complicated, right?
The problem when you try to solve to school districts, you may not have the best technology win, right? The technology has the best entries or know how to sell better, or your customers is not the end user of the system. So. We didn't wanna do that. We wanted to go direct to the consumer. It doesn't mean down the line we'll wanna work with schools, you know, and make it easy for teachers to leverage the platform.
Yep. But right now focus is direct to consumer.
[00:10:14] Alex Sarlin: Gotcha. Yeah. No, makes sense. And one of the features that Sizzle offers is the ability to upload a question or photograph a question and get. Support with it. And the site talks about getting step-by-step sort of breakdowns of what you can do for any individual question.
And I feel like this is a really key moment, Fred Tech, because it's one of these moments where students may, at that time, if you know it's late at night and they're trying to get through a question and try to. Prepare. They might just want the answer, but we all know that the learning experience there isn't the answer.
It's the process. It's learning how to actually do it. It's feedback on what you know and what you know how to do. How do you do that step-by-step breakdown, and how do you sort of try to ensure that you're not helping students bypass some of the work they have to do, which I'm sure some of them want to, even if.
It's not good for them.
[00:11:00] Jerome Pesenti: And one of the problem, see with this new generous ai, right, is that it can make students not do the. Obviously, I think it's a news generation of tools that we're gonna use in the workplace, so I think teachers need to assume that the students are gonna use these tools for the more specific question for you, what we try to do in our case, if we try to not give you the answer right away, we try to show you the steps, have you think through the steps as you go through it, but ultimately teachers have to know if the students want to answer.
They can use chat pt, they can use Gemini, right? They cannot figure out a way to get it. So you have to rethink homework in that new context. But ultimately, I think it's like everything. It's like the calculator In the past, these tools can be well used and can make us all more efficient. They can make us lazy.
You have to kind of change your curriculum to make sure you adapt to that.
[00:11:50] Alex Sarlin: Yes, I really agree with that. I've heard some people say that basically, you know, if a homework question can be completed by a AI bot just immediately, then it's probably not a good homework question. Assume it'll
[00:12:00] Jerome Pesenti: assume it'll, yeah,
[00:12:01] Alex Sarlin: yeah.
'cause as you say, there's no shortage of tools that are able to do that right now, including the foundation models. So that makes a lot of sense. And yeah, we talk about that calculator metaphor on here as well. So you're in a space that I think is really exciting in the AI ed tech world right now because it's very fast growth.
You have a lot of. Learners out there learning all sorts of things. You focus a lot on STEM topics, on science, technology, engineering, math, and people are doing it. They wanna be learning, they wanna be studying it in a way that supports their grades and supports their mental health and supports their ability to get things done quickly.
So you have sort of a really built in excited audience. It is a really hot part of the AI ed tech space that I think a lot of. People who listen to this show, who often think from the school perspective, just don't realize how big it is. I don't wanna ask you for any specific stats, but I'm curious, has sizzle grown faster or less fast than you expected?
And sort of how have you seen it spread around the world? I.
[00:12:57] Jerome Pesenti: So, I mean, we've been growing faster than we expected initially. I think it was fairly easy to give a homework help tool to students like the use case. They like that. Doing my homework helped me do this. I don't have anybody to talk, and it's very useful.
I think when we had to spend a lot more time, and I think we finally kind of nailed it, is something for test prep where it's not just. Making it easier for you to do your homework, but really helping you do proactive work. So in this case, we don't replace. Your work, we just gives you something that makes you more effective, but you need to have the intent and that thing.
This is where it becomes more interesting, right? When we really facilitate and we get them to prep in a way that they wouldn't do without sizzle. So it's really like clearly, clearly positive. I think homework help is. Because ultimately there are a bunch of tools out there that give you the answer and you can create a homework help tool that's not gonna give you the answer ever.
Even if we can like make it harder for you, we have to give it so it's easy to do this. It's harder to get the people to really learn, but once you get it, then you give them what they're gonna learn on the test. And give them practice they like. They really like the tool.
[00:14:07] Alex Sarlin: Yeah. So it sounds like you have two different use cases that we mentioned.
The homework help one, but it really feels like this test prep and practice where you can generate endless practice problems, help people create study guides, study plans with actual dates and times and support them and being able to prepare efficiently and effectively. That sort of takes it to a different space.
It takes it a little bit out of the direct students having to war with themselves and trying to figure out what tool to use that was is gonna either give them the answer or help them learn. It's pure learning.
[00:14:36] Jerome Pesenti: Yeah. And ultimately, you know, the vision, right? Is that they need to be able to do, I mean, when teachers give them their homework.
They need to be able to do that homework on their own. Right. So that's the goal. If you can't do your homework on your own, you won't be able to pass the test. So you need to get them facilitate that step to get them there. Yep. And obviously if you just give them the answer, it's not gonna work. They're not gonna get there.
[00:14:57] Alex Sarlin: Yep. Let's deep dive on that second piece here, because I think that's a really increasingly interesting dynamic that's happening in the education space in the age of ai, which is that. People always say, you know, any kind of plagiarism or cheating only cheats yourself. That's sort of the, the famous line with anything like this in the world of ai.
I, I, you know, I agree with you that these tools are accessible everywhere at any time for almost any student, and this is just part of the toolkit that everybody now has at home. So the question becomes, what? Are you gonna be able to do on your own when you walk away from those tools? And especially like you said, in a test context, back in a school monitored environment.
That's right. So there's something really interesting happening where students go out into the world and have access to these incredibly powerful intelligence tools, and then they have to go into a little bit of a, you know, hermetically sealed environment. It could be a classroom or a test center and say, okay.
How much of this thinking, how much of these skills, how much of these algorithms, you know, have I actually brought into my head? I'd love to hear you talk about that dynamic. 'cause I think it's one that all of EdTech we're trying to figure out what should happen in each space and where, you know, should tests look the same as they've always looked, or should tests look different in a world where you're trying to assess a different type of skill?
Just look
[00:16:10] Jerome Pesenti: at the A a T. Not that I'm a huge fan of SAT, but it's a model of a test. Yeah. And they have a new part that's Swiss calculator and part that's without calculator. Right. My assumption is that our testing will go there. I do believe that in the future there'll always be tests without ai. And I think there'll be some tests that are AI assisted.
Yep. And then will require a different kind of performance. And if you know, you don't prep yourself well, you're gonna fail even maybe both of them actually. Right. So it's not gonna be about just showing your work when you are doing your homework. Like, who cares about this? And it's gonna. It's not gonna be writing essays at home that make 10 pages.
That's not gonna help you. But if you don't know that you're making progress. You know? And this is where tools can help you assess your readiness in a no-nonsense way. Then you know, you're not gonna be ready. And I think students will figure that out. And I think the the system will converge with something that's gonna be along those lines.
Right. And if students are smart, they'll figure it out. You know,
[00:17:06] Alex Sarlin: I like that phrase, assess their readiness, that the AI can be used by students to assess. How ready they are to achieve a certain, either pass a test or perform, do a performance task, and they can use it to practice. They can use it to train, but it can also tell them how close they are to being able to do it.
If they were to go to a less AI enhanced environment, or one with none at all. I mean, imagine today what's
[00:17:29] Jerome Pesenti: happening, right? So to do that, teachers give students homework. Okay? Which I guess more and more they're gonna generate through platform or ai, whatever. And then the students use AI to solve them and pretend that they did the work.
You know? And then the teachers actually have to evaluate sometime that homework. I'd be like, talk about busy work, right? So I think in the future it'll be like, okay, here's an AI platform. You need to basically do a couple of hours of work in it. Uh, at your own pace, and then the system will go, uh, report back to the teacher and says, well, they did a couple of work and this is where they are.
And it must be shared with the, with the assessor and that, and that's it. You know, it's not very complic and the teacher doesn't have to bother with that busy work, you know, and, and worry about, oh, are they cheating or not? That would be much better for everybody. And ultimately what matters is how already they are.
And there will be a test where they will not have access to the ai. So it's in their interest to make sure they do it properly. You know? But even there, you can just do it where it's part of the system, you know? And, and it works for everybody. The teacher and the student and it's a higher quality assessment.
[00:18:33] Alex Sarlin: I agree. I think it's an exciting vision that sidesteps a lot of the angst that people get into when they think about some of the future of AI assessment or you know, all sorts of things. I like the way you just painted it out.
[00:18:45] Jerome Pesenti: I think you really need, another part is you need to free the teacher from like some of these tasks they're doing today.
You know, I think like giving homework and assessing the homework, I think is one of the, one of the satisfy.
[00:19:00] Alex Sarlin: There are certain pieces of this educational journey, both on the student side and the the instructor side that AI really excels at. And you, you were at IBM Watson, you've been following the cutting edge AI for being part of it.
You've been building it for many, many years, and I think we're at this interesting moment, I think in AI history where AI is suddenly. Very visible, obviously, and able to do things that we have not usually connected with, you know, artificial intelligences, including elements of teaching and explaining and, and practice and learning and problem solving and all these things in a very flexible way.
It's just opening up this whole new. Set of ideas of what it could do next. Where do you see sizzle going in the future? So in a world that, in which it becomes a solution for what studying looks like for a huge number of students. And they start to say, I can take any body of material and it will break it down into bite-sized pieces.
Create a study plan for me, make it. Feel fun and engaging in small bits, and then assess my readiness as I go. Where would you wanna go? Would you wanna expand into different types of areas? I know you already do a lot of different areas. Would you wanna go into, uh, subjects that may not already have study guides?
Like if somebody wanted to learn something that is not in a traditional curriculum and could it work that way? I'm curious how you think about the future.
[00:20:14] Jerome Pesenti: Well, today, actually, you can do that. So you don't need to have material. You can type any subject. You can say, Hey, I wanna study like. How to become an astronaut.
Yeah. Or you can ask, you know, like, I wanna hear about French history in the 16th century. It'll give you something pretty reasonable so you, you don't have to bring material. Hmm. But really your vision is to be a learning companion throughout your life. Ah, you know? So we'd love to be the companion where you learn things for school, for work, and for your own interests.
And as you wanna learn, you know, like we, you know, you or something, you know that it's a place that knows you, that knows how to learn, you learn and can assist you and get you there as quickly and efficiently and engagingly as possible.
[00:20:54] Alex Sarlin: That's really interesting. And then could mix subjects. So it could be that you're learning multiple subjects in school and something in your home life and something in your professional life.
It'll know all of that. And as a companion can
[00:21:04] Jerome Pesenti: support you. You recommend what you can do next. You know, we're not yet there, but like, here's so thing I love when vision for me is like, you have a goal in life, right? This is something you're, have an aspiration. What are all the the things you need to do to get there?
You know, I wanna become an astronaut. What do I need to know? Why do I need to know this calculus as is math? And the system could tell you by, well, this is why. And really like, instead of being something that is push on you, it's something that you really want. You understand the context, you understand the motivation.
[00:21:31] Alex Sarlin: Yeah.
[00:21:31] Jerome Pesenti: And if you even put, you know, when you, you wanna be an astronaut and it will make your math exercise in that context. And then you will understand that, or, or your physics exercise or, or your chemistry exercise. And then now you're like, oh, I understand why I need to do these things.
[00:21:45] Alex Sarlin: That is very valuable part of what educators can and should do, and, and Whatis can and should
[00:21:50] Jerome Pesenti: do.
It's hard, right, because 30 kids and wants to be a nurse, another one wants to be an astronaut. The other doesn't what they wanna do. How are you gonna create a teaching environment that address to all of this? This is where AI can be more personalized and then the teacher can look more at classroom activities and like collaboration.
But so I think there's a nice divide and conquer possibility there between the teacher and the ai.
[00:22:12] Alex Sarlin: Yeah,
[00:22:12] Jerome Pesenti: that makes
[00:22:13] Alex Sarlin: a lot of sense. And so when you paint this vision of a learning companion that can support any learner in their academic work and their personal work life, professional life, do you envision the learning companion having a particular personality?
And I guess this also reflects, do you. Instruct sizzle to have a particular personality or to take on a certain role as it supports students. This is another thing that AI does really well. You know, can Im imitate different types of human roles or different types of human communication styles. I'm curious how you think about that aspect of it.
[00:22:43] Jerome Pesenti: Yeah, it's a good question. We've stayed pretty in the neutral right now. We just wanted to stay away from the mascot idea for. I don't know if it's a good idea or not. You know, I suspect at some point people will choose and decide how personalized it is and how much personality it it has. I would say we're still at the beginning there and we haven't figured that one out.
[00:23:04] Alex Sarlin: You know, Ian, you just expanded. When you say, uh, we've thought about the mascot idea, I think I know what you mean by that, but I'd love to hear how you think. You mean like how Duolingo has a duo or some of.
[00:23:14] Jerome Pesenti: The question also is like, how do you, what do you project right in that ai, you know, what does student project and ultimately, it's not a maco, it's not a person, it's a system right now, so does it need to have a personality?
Does that personality needs to be adapted to you? You know, I don't have answers to that to be honest.
[00:23:31] Alex Sarlin: I don't know if anybody does. I think everybody's just trying different approaches and yes, we're all gonna learn as a, as a, you know, as a society, as a civilization, what works in this space and what doesn't.
And it probably, frankly, will be different for different people.
[00:23:43] Jerome Pesenti: And this is what we're trying to set up, actually. We're trying to set up a framework where we can learn these things. So we will do experiments where, hey, if there's more of a personality in the system, is that gonna work better or not? We do experiments every week.
Right now, we launch new things. And we see do people stay longer on, you know, learn longer and more efficiently with this or not. So that's what we haven't done yet. But you know, it would not be very complicated to do that in our platform. And the beauty of digital system is you get the result in numbers very quickly.
You know, like if you had to do that in the classroom, you know, for example, does the personality of a teacher help or hinder learning? It will be. A major undertaking is trying to do that. But actually in the digital form, I could do that in a few weeks, you know, and give you an answer.
[00:24:30] Alex Sarlin: Yeah. It is one of the great powers of educational technology, the scale and ability to test that way.
And you know, when you say that, it reminds me again of Duolingo, they were, that's a company very famous for instilling that constant, continuous testing framework into product. It sounds like you've taken a similar approach where you're constantly testing new ideas. The idea of being able to test the value of a personality is very.
Interest to me and that it's not like there's one right answer. You know, the value of a personality and what personality and for what type of learner. All of those create a really interesting space for us all to learn about what learning. Yeah. And
[00:25:03] Jerome Pesenti: one, our goal at some point is to be able to even like have scientific evidence on some of these things, you know?
How personalized does a curriculum need to be? How personalized the practice, you know, how much is it important to anchor it into reality versus not? Like, yeah, if we could test this thing at scale, we can start getting some real answers and learn from that, you know? And maybe people are different or people are all the same, you know.
Let's test these things.
[00:25:25] Alex Sarlin: Yeah. It feels like the beginning of a whole new era of learning science where we can test things in a, in a much more just field way. Right. Not with 20 undergraduates with a, you know, a very particular type of thing.
[00:25:38] Jerome Pesenti: Learning sounds very hard in the lab. It's very hard in the classroom, but on a digital platform, especially when it's bite-size, you know, like when you have a lot of, you know, basically like optionality at every moment to present different things.
Then you can learn very fast. And that's what we're trying to create. That's the platform and the vision of ol
[00:25:57] Alex Sarlin: Yeah. Researchers listening to this and others who are following the education and generative AI research, we've gotta make a plug here for the Stanford Scale Accelerator for learning. They just launched the ed tech research repository hub, so they have over 400 papers there about what we do know about education, generative ai, but frankly.
It's still not that much. It sounds like a lot, but we're still not that much. We're very, very, we research actually,
[00:26:19] Jerome Pesenti: we inspire ourself and we use it to I'll platform.
[00:26:26] Alex Sarlin: Yeah, it seems like sizzle is moving very quickly to sort of figure out what works for who in this really interesting context of studying for upcoming exams or studying for any particular topic.
It's really, really interesting. So last question for you. You bring an unusual depth of background on ai. To the founder role in EdTech. We have, you know, I've talked to hundreds of founders on here. They all have very different stories. Some came from other fields, some came from the classroom, some came from administration.
But you come from like a AI background. You founded multiple AI companies over 25 years. What do you think that does in terms of your leadership style or your sort of approach to the technology compared to somebody who may just be learning about AI in the last two years?
[00:27:09] Jerome Pesenti: It's about being extremely, I would say, data driven and data centric.
You know, AI is one aspect of the automation and. And so what we're trying to do is be rigorous, develop, understand how people improve and keep iterating on the products. I think it's very like rigorous data centric approach is key, you know, and try to be state of the art in everything we do.
[00:27:32] Alex Sarlin: Yeah, and it sounds like you're being scientific rather than ideological.
You're not saying, oh, this type of learning is the best. We're just gonna do that. I don't know. You know, it's not me. I'm not gonna tell you that. I think we want to figure it out. I. Admirable approach. Really interesting works. Everybody should check out Sizzle ai. It's, it's at S z.ai, right?
[00:27:51] Jerome Pesenti: Yeah. And you can just go in the app store or the play store.
It's easier. Just search for sizzle there. Yeah, and you'll get it.
[00:27:56] Alex Sarlin: That's probably a better idea. I'd love to get a study plan for some of the things that I'm trying to learn. I'm always trying to learn more things in this crazy fast changing time.
Work with Sizzle for quite a while. It's really nice to actually put a face to the name. This is Jerome. He is the CEO and founder of Sizzle AI and has long experience at VP of AI and Meta and IBM Watson and Co CEO of benevolent ai. Thanks so much for being here with us on EdTech Insiders. Thanks. 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.
