How AI is Transforming Education

E26 | With Sizzle's Jerome Pesenti
Updated Jan 26, 2024

How AI is Transforming Education

AI For All
December 14, 2023
In this episode of the AI For All Podcast, Jerome Pesenti, CEO and founder of Sizzle and former VP of AI at Meta, joins hosts Ryan Chacon and Neil Sahota to discuss how AI is transforming education. Jerome shares his insights on AI's enormous potential to disrupt education by providing a personalized learning experience. He elaborates on how Sizzle acts as an artificial tutor, adapting to student's interests and providing real-time assistance in problem-solving.
About Jerome Pesenti
Jerome Pesenti is the CEO and founder of Sizzle, the free personalized learning app that leverages AI to help learners work through any problem, step by step. Prior to founding Sizzle, Jerome served as VP of AI at Meta and Vice President of IBM’s Watson Platform. He has been involved with artificial intelligence, natural language processing, and search for the past 25 years.
Interested in connecting with Jerome? Reach out on LinkedIn!
About Sizzle
Sizzle is a free educational app that uses LLMs to make online learning more personalized, fun, and effective by leveraging the latest advances in AI. Sizzle helps learners work through any homework problem, step-by-step.
Key Questions and Topics from This Episode:

- [Ryan] Welcome everybody to another episode of the AI For All Podcast. I'm Ryan Chacon. With me is my co-host, Neil Sahota, the AI Advisor to the UN and founder of AI for Good. Neil, how's it going?
- [Neil] Yeah, I'm doing all right, Ryan. Just another day in paradise out here. How about yourself?
- [Ryan] Yeah, things are well. Things are going well. We have great weather out here in the DC area. Holidays are coming up soon, so we're excited about that too. We also have our producer Nikolai on the show today as well.
- [Nikolai] Hello everyone.
- [Ryan] Today's episode, we have a fantastic one for you. We're going to be talking about how AI is and will transform education and how students learn. To discuss this, we have a very exciting guest, Jerome Pesenti, the founder and CEO of Sizzle. They are a free educational app that uses LLMs to make online learning more fun and effective by giving everyone access to a personalized AI tutor. Prior to founding Sizzle, Jerome was VP of AI at Meta and VP of IBM's Watson platform. Jerome, thanks for being on the podcast.
- [Jerome] Thanks for having me.
- [Ryan] Yeah, it's great to have you. So let's start this off. I think elephant in the room is to hear a little bit more about your background and experience at Meta, being the VP of AI and with IBM's Watson's platform. Tell us a little bit more about kind of your background experience. Tell us, just give our audience a little more context to who they're going to be listening to today.
- [Jerome] I'm Jerome. I'm French. I came to the U.S. 25 years ago to do some research in AI. I came to Carnegie Mellon. And then decided to start a company in AI actually at the time. So in June 2000, which was like the crisis at the time, I decided to bootstrap an AI company before it was the trend. And it actually became a pretty good adventure. It was a B2B company doing, search, text mining, NLP. We call it natural language processing. And that was a pretty good run. I ended up selling it to IBM in 2012. And then at that time I became part of the Watson platform there where I had a significant R&D team to do putting AI in the cloud, which is I would say like ahead of its time. IBM decided to invest in commercializing AI before it was the fashion. The little challenge there is that they didn't have much of a cloud, it was a little hard to create a good offering, but it was a good intent. Decided to leave for another startup doing AI for drug discovery that was based in London, a company called BenevolentAI, it's still around to this day. And then I got recruited to lead AI as you mentioned at Meta, which was a really interesting adventure there for almost five years. It started with a couple hundred people and brought in almost like 10X doing things like PyTorch, Llama. I'm sure you've heard about that. And things like AI for content moderation, AI for ads, which paid everybody's salary, and then AI for recommenders, things like Reels and other systems. And then decided to leave 18 months ago to do my own thing again and do something I thought was really beneficial for humanity, and I thought like learning and education were a pretty good direction.
- [Ryan] That's awesome. I want to get into kind of what you're doing at Sizzle in a second but regarding the work that you did at Meta prior to leaving, and I guess leading into kind of what's going on now, how do you see Meta's current AI initiatives or what do you think about them and just the impact they're having?
- [Jerome] I think there's two aspects. And when I was there, I had, in a way, like almost two teams, you can think of it. One was more research centric and one was more applied. On the research side, the big thing that AI, that Meta is pushing, and that's, it comes from Yann LeCun, who was there before me, and then there was somewhat also Mark's view, which is be really open, right? So develop a bunch of models. I think Meta is still like the number one producer of models on Hugging Face that are used by others. There's a long tradition of creating open things. Not trying to commercialize. Meta is not a B2B company. So we're trying to put this model out there to raise the field and get more performance. So that's one thing. And so we did PyTorch, we did things like Llama, but also a lot of other models like Detectron and others, Faiss. And then there's the other side, which is using AI to really impact the business. And there I would say, the three big things we did were one for content moderation, which while I was there went from being mostly a manual endeavor to something at the end was most of the action and flagging was done by AI, 95 percent of it. And then ads, works amazingly well when you put the most advanced AI behind it to create the most relevant ads that you're the most likely to, showing you the most likely product that you're going to buy. And then recommenders, like things like how do you decide what's the next piece of content you show in your feed? And there again, it's very impactful for the company.
- [Ryan] One question we were thinking about before this was, so obviously there's a lot of stuff going on in Meta with the Metaverse. How do you see AI and the Metaverse kind of working together, being complementing each other, or how do you view that?
- [Jerome] Yeah, at the time, so there was, obviously another big effort, maybe less impactful in the moment from a direct impact because it's still something nascent, but a big effort from a vision perspective. So I ended up, by the time I left Meta, I was actually even the AI team was part of the Metaverse team actually. I would say it's twofold. One is glasses. How do you create very like seamless glasses or a computer that you interact with? And you've seen, there was a new version of the Meta Ray-Ban that just came out. And really I think one of the big part of the pitch was the AI that's part of it. So that's one aspect, we could talk about this. The other is more of the Metaverse itself as a whole with a bunch of agents that are part of that. And especially with the recent enthusiasm for ChatGPT and the characters you can build around it, there's a whole idea around creating these characters if you want that you can interact with either through regular chat or in a Metaverse environment. So I would say this is where the two directions.
- [Ryan] Yeah, that's pretty neat. I saw, I think, Mark Zuckerberg had a video of him using the Ray-Bans to teach himself how to braid his daughter's hair the other day so.
- [Jerome] Yeah, really great use case for Sizzle actually. That's something we'd love to be able to teach you anything like this, yeah. Actually, something we've been thinking about is we'd love for the Meta glasses to get traction because it could make Sizzle a lot more compelling actually. Right now we wouldn't do it because there's not that much usage. But imagine if you could ever get a tutor that sees everything you see and gives you feedback. Even if we don't get the visual in, obviously the holy grail would be to get the visual feedback, but even without visual feedback, just audio, seeing what you do, it's in some ways similar to what we do today. We take a picture of a problem and then we give you feedback as to how you're solving it. So we think this could be really part of an interesting future for learning.
- [Neil] What's holding us back from doing that, Jerome? Is it infrastructure?
- [Jerome] No, I think technically, actually, it's quite possible. I would say it's more adoption. So for a small company like us, we need to develop for platforms that have usage. But if the Meta glasses get some traction, I'm sure, I'll be sure to give Mark a call and say, hey, I think we should do something together there. I'd be excited to do that. I think it could be really one of the killer use cases for these glasses. Or for AR, to be honest, or mixed reality as a whole, actually.
- [Ryan] Yeah, I mean, there's probably applications even outside of thinking about education from a student's perspective but also for professionals. I wonder how this kind of evolves into something with wearables, with kind of Meta, or these kinds of glasses in the Metaverse that allow these augmented reality, mixed reality to come into play, to teach people how to do certain jobs, and have a trainer or tutor that allows when you're onboarding at a new position or what, that kind of approach. I think it'd be really interesting to be able to teach people skills and teach people things for professional sense as well.
- [Jerome] Yeah. This is really definitely like something for us as part of the vision. It's not something today, but today, we have two modalities. We have text, and we have vision, but it's more vision as to like diagrams or pictures, but down the line, it could be like real time vision of what you're doing. And then at some point we want to also give visual feedback that's not just text, but it could be like, hey, this is how you do it, and maybe overlay on what you're seeing, right? So it's this amazing potential of like mixed reality or augmented reality to do learning. Absolutely.
- [Neil] What about haptics? Because I've actually seen a group of grad students develop a haptics AI program, wear these gloves, and it helps you, teach you how to play guitar based on your hand movements and positioning. You can actually help create your kind of your finger motions.
- [Jerome] Yeah, that would be even, I mean I would say it's even more far fetched and all these things again, I think the biggest challenge for us, I think I see from an AI perspective, this will not be as hard as you think because often it's easy to figure out what's the right way to do it and learn from people doing it and learn, you know, the motion and stuff. The challenge is all this platform. So if I'm going to go develop like a specific tutor for this kind of input and output, if I only have hundreds or thousands of potential users, it's not worth it, right? Today the phone is billions of people, so that's what we're focusing on right now.
- [Ryan] When it comes to the mission of what you all are doing, when you told us earlier that you left Meta to do something that you felt was really good for humanity and that was really needed for people, how do you think as a whole, as we look at where we are now, where we're going into the future, how AI is really going to transform education and how students learn versus kind of what we know today.
- [Jerome] One of the reasons I decided to do my own thing is people always maybe don't pay enough attention to what are the objectives of the AI that we put in on the system. I think this is like a really important topic. What objectives you give to the AI? Every AI you put out there is trained to optimize something. And most of the platform we use today are used to either optimize sale or optimize engagement, right? So if you go on TikTok, it will try to optimize engagement, but it's kind of like, it's just engagement for engagement. It doesn't tell you if you're going to learn something out of it, or you're going to get long term value.
Something I had tried to do at Meta is to, and we showed some progress, which we show different objective for the AI. Learning, being passionate about something, being inspired by what you're seeing. So this is something I'm very passionate about, creating AI whose objective is to make you a better person.
And so what we want to do with Sizzle is an AI that will see how much you progress in your learning goals and optimize the content it's showing so that you learn faster, better, and you stay engaged, and you actually succeed in learning. That's really the big aspect. And I'm not an educator. I don't know how to do it, but I will give a lot of choices to the AI, and the AI will figure out what works best from the multitude of content or things it generated itself.
- [Ryan] When you mentioned being able to apply a tutor for each person, what do you mean by that? When you're talking about like an AI tutor, when I obviously think of a tutor, I think somebody that comes over, you go to their place, you learn, they sit down with you, they walk you through a certain subject matter, right? If you have a math tutor, they come, and they teach you math, they sit down with you, walk you through, help you learn in addition to what you're doing in school. How does an AI tutor work? Like when we're, when we hear that kind of phrase, what should we be thinking about in kind of the application of the technology?
- [Jerome] Yeah, so I would say that today, right, the way these AI tutors work is that you can bring to it any kind of problems you have or any kind of question, right? So even the way Sizzle works is you can bring in your problem you're working on right now. Sizzle doesn't know what it is, but you just bring it, and it will help you with it. So it's like this very general solver, and it will give you step by step method to solve that problem. And then if you have any question about any of these steps or any of the answers or, it will also answer and be infinitely patient, right? It can answer an infinite number of questions you may have.
So obviously that's the first thing, which is it just is adapted and responds to your needs as particular and as extensive as it may be, right? You can ask any number of questions. Where we want to go, which becomes a lot more interesting, is that it can also suggest things to you and suggest a curriculum for you that can be infinitely personalized, right?
In the past, I think people have talked about adaptive learning. Something that's pretty well known, which is, oh, okay, I'm going to give you the next exercise, et cetera, et cetera, that's going to be like at your level. But I think what's even more interesting in adaptive learning is that the content itself could be changed dynamically for you. Oh, you're a soccer fan. Let me give you problems in soccer, right? You're a baseball fan. I'm going to give you like, you are in a certain language, I'm going to give you content exactly in your native language rather than force you to learn in English. Having something that the content itself is completely adapted to you, so that it optimizes your learning, keeps you engaged, keeps you motivated, and is the one piece of content that's the most optimal to get you to make progress.
- [Ryan] How do you personalize an AI system to do that? Because I think that's a very fascinating kind of approach. And if you really think about what is it that we as people often retain the best, it's when things that we have interest in, that we can relate to. So when you're learning a material, if you can take it out and put it into context, or put it, have it relate to something that's of interest to the individual, they're more likely to probably retain it. So how do you personalize an AI system to do that?
- [Jerome] So I think it's going to be an intersection of two things, okay? One part of personalization is what Sizzle does today with LLM, with large language models. The way they work and the way they personalize is you give it any kind of context and it augments that context, right? And it works pretty well at doing this, right? So I give it the context of the problem plus a question, and it knows how to answer that question about that problem, right? So that gives a very personalized experience. The second piece is what has been done in my prior job, what I was doing is all these recommender systems, which really deeply understand your behavior. They understand what makes you tick. They understand who you are. They understand how you stay engaged. The vision is to combine these two things together, right? A deep understanding of the user and a high contextuality on what you are trying to do at the moment, right? And being able to generate content that's highly personalized to you based on who you are and what you understand.
We cannot ask that for a teacher, right? So you have a class that's very heterogeneous, you have 30 different students, they're all stuck on different things. They have all different learning styles. But a teacher cannot do that, right? It would be impossible. We know that one on one tutoring works a lot better thanks to that. And other aspects also because you have someone that pays attention to you, that keeps you committed, et cetera, et cetera. So we want to be able to go from like a classroom, and we're not going to change that, I think classroom teaching has a big role to play. But being able to have this kind of one on one attention, very personalized, really understanding where you're stuck, what are your gaps and adapting the learning, but not just what the next exercise but even the content of the exercise is and how it's presented to you, and how it's, how it's structured, the type, et cetera, et cetera. We could do all these changes.
- [Nikolai] What do you think AI really is bringing to the edtech space that was like missing before? What can AI, you mentioned personalization, but what can AI really do to improve these platforms and just improve edtech in general?
- [Jerome] With recent advances, the one obvious thing that people are very excited about, that's what we learn a lot about, LLM for learning, is that you have this kind of like assistant like, like a learning partner that's there to be able to answer any question, and you can dialogue with them. So I think this is something that I would say is pretty exciting already. I think the bigger opportunity is that we can generate content and experiment with that content. So the one thing actually that brought me to this field is if you look at most of what is done in the field today, it's based on fixed curriculum. People have an idea how to teach, and they think it's the best way. And when students are in this curriculum, they don't have a lot of choices actually. There's very few learning platform where a student has an infinite number of choices. Compare that to TikTok. You go to TikTok, immediately, you have a wealth of different options. And after 30 minutes, the system that you see is entirely different from another person. Completely personalized to you. Nothing in learning matches that, not even like a thousandth of that opportunity, you know. So that's what we want to do with Sizzle, is to have something that not only is responsive to your needs, in a way that can dialogue, etc, but also is really customized that learning content and inspiring ourself from the things that I learned from my prior job, which is hey, if you put a lot of content in competition, you will get a lot more engaging experience for the user and it will work better for them.
- [Nikolai] So with education, obviously information has to be accurate. How do you deal with hallucinations or like it producing inaccurate information? Is that something that you have to navigate?
- [Jerome] Yeah, absolutely. That's a big concern. The systems today makes mistakes. Parents and tutors and teachers as well make mistakes, but it makes maybe too many mistakes. Not enough to be, it's still actually quite useful. So our accuracy right now that we can track is, I would say, 75 percent of the time it's fully accurate and 90 percent of the time the steps themselves are accurate, but the answer may not be fully accurate. So it's not perfect. I wish it will be much better, but it's enough to be really useful, and the students that use it feel like it really gets them unstuck, okay. But we are, it's still a research topic, but there are a lot of opportunities to improve this. For example, LLMs tend to make calculation mistakes. Actually we have calculators, so you need to get the LLM to use a calculator rather than trying to do the calculation themselves. That's a big opportunity to do better. You can also check itself. So actually LLMs are pretty decent at figuring out that they are themselves wrong. It takes them longer though. So you can create like more accurate sets of well-solved problems and train the LLM with that, et cetera, et cetera. So we have some, it's not perfect today. It's still good enough to be very useful and this is why we have pretty good traction with Sizzle today, but we need to make that better, absolutely. Ultimately though, it will be a system that still make mistakes, and I think people need to accept that. This is technology. As every AI technology out there which can make mistakes. There's no way to get to 100%.
- [Ryan] How have conversations that you've had with educators been when it comes to adopting new technologies or technological advances for classrooms or for educators themselves? How does that kind of currently stand? Do you, what do you see as the biggest hurdles you have to overcome for this to be something that more educators and more classrooms adopt and bring in as opposed to potentially avoiding adoption because they're unsure of maybe the accuracy of the information or they're unsure how to use it. How do you bridge that gap? Especially when a lot of educators, or I guess educators in general, the range of age and experience with technology is so vast. How do you navigate that to have this something that can be more widespread and more easily adopted?
- [Jerome] So first, I would say that in some way, I don't have to navigate that. They are, and there's this, there is a current state. And I'm not saying it's easy, but it's a current state. The current state is you have students who are using these tools, okay? And they will use them. It makes their life much more effective, much better. There are companies as well who use this tool. My company, today, I will not recruit a developer that doesn't know how to use GPT-4 or Copilot to code. It makes you literally 30, 50 percent more effective. So you need to use this. So if you're a school, you need to train people to use these tools because if you don't, you're going to produce people who are subpar in their jobs, right? So it's not like they have an option. Why I think it's tough is that it's coming very quickly, right? Much faster than people assume. And it's hard because it's hard to adapt to that. But they will have to adapt. What we at Sizzle are trying to do is we want, I've talked about objective, we want to create a system that's optimized for learning. Not cheating, not showing you the solution, learning. We're going to incentivize it. Doesn't mean you cannot see the solution for a system because if we don't show it to you, you'll go somewhere else. But we're going to nudge you to answer it. We're going to make it easier for you to answer it. We're going to tell you when you're not learning. And we're going to optimize for that learning. So I think as a company, that's our responsibility to optimize for that. But the school system will have to adapt and, and maybe hopefully promote the tools that are more responsible rather than the ones that are less, right?
- [Ryan] Those copilot tools you talked about, obviously, we've seen a lot of, especially developers, find a lot of value in using these copilot tools to code, to develop software. How do you, when it comes to an earlier education in implementing these tools, these AI solutions to help people learn more effectively, how do you make sure that they still are developing the problem solving skills and the understanding of the fundamental lessons that are being taught or the fundamental topics that are being taught and not being super 100 percent reliant on the AI, on the software, on the technology because obviously that is a big part of the educational process too, is that they fundamentally understand what they're doing and they don't rely solely on the tool to give them an answer every time.
- [Jerome] Yeah, so for us, as the user, the learner interact with our tool, you can see how much effort they are making. And our goal will be to get them to go up in that effort scale, if you want. There is this little acronym called ICAP. I don't know if you're familiar with it. It means Interactive Constructive Active Passive, which is basically like a decreasing order of effectiveness of learning, but also it's a decreasing order of involvement of the user, right?
So interactive is having a discussion. Constructive is problem solving, active is taking some notes, and passive is just watching. And so we, as we interact with the user, will try to get them to go up that scale and do a lot more interactive, constructive things than just passive interaction with the content. And that's the key. We want to optimize for that.
- [Ryan] That fundamental understanding is not lost when you're building tools and making and say, hey, look, this is going to make teaching more effective. This is going to make learning easier. But at the end of the day, we're still not trying to have you be completely reliant on the technology. We need you still to understand.
- [Jerome] There is a huge responsibility for people developing in this space. And it's not just learning. Like with all these LLMs, right, you can create infinite amount of content. Are you going to really spam the world, create a web that degrades with all this content. There's a lot of risks attached with this. There are today. I'm not worried about doomer risks, like from the AI taking over, but I do think there are risks with the AI systems we use today that we need to be very conscious about.
- [Neil] What about accessibility, right? There's gonna be a lot of power, good or bad here. How do we ensure that there's fair access to most people to be able to take advantage of the opportunity here?
- [Jerome] Well, right now, Sizzle is free, so it's accessible to everybody, and we have users from everywhere in the world, from Somalia to South America to Canada to the US, and anywhere in the US, right? And it's gonna be part of our DNA is to make a base offering that everybody can use. Actually, we are inspiring ourselves from some other companies like Duolingo, for example. When they came out, we're free for language learning for everybody. And then they managed to find a good business model by creating a premium offering that didn't really degrade the experience of the free learners. So this is what we are inspiring ourselves. This is the big passion for us, is access. Rich families in New York, I'm based in New York, they can have tutors. They do, actually. Everybody else, they struggle more. And then if you are in the in a country where you have little access to education, it's even way worse. So that's something I would love to help, help with, I'm not going to solve that, but at least make a dent in improving the situation.
- [Ryan] So if we think about kind of everything you've talked about here, what you have going on is super fascinating. I love the fact that you're able to build learning modules and a learning experience for people and connect it to things that interest them. That's something I know personally works really well for like myself, just being able to, I digest knowledge and information better when it's around something that I care about, that I have interest in. So being able to do that I think is going to pay huge dividends for people out there that just maybe struggle with learning and a lot of times they struggle because I just think they lack interest in the topics but if you relate it more to what they have interest in, I think you're going to see them be able to retain the information a lot better, which is awesome.
When you look at what's going on in the edtech space, what you all are doing, just your background experience at Meta, at IBM, what do you think of the current landscape right now and how real the hype is for a lot of the AI kind of talk? What should people that are listening to this really understand about where we are as society with AI and also maybe what they should be careful about because we've talked about some of the things that you need to be aware of and understanding when you're using a tool like Sizzle with the information and how accurate it is and how it's going to evolve and grow but what's important for people out there to really understand about what's going on right now in the AI just like industry as a whole?
- [Jerome] So if we talk about learning and education, right? I, my feeling is the industry has been been stuck for a long time, right? Online learning doesn't look that different from regular learning, and it hasn't been, the promise of the MOOCs didn't come through. People were not more engaged, were not more successful at completing them. So that's been like the state for a long time. I think the current advances in AI, that are more recent, are just throwing a wrench in the system. It's going to disrupt the system. It's going to disrupt classrooms. It's going to disrupt the notion of homework. It's going to disrupt essay writing. I believe, I'm optimistic, that overall it will be a really good impact. It will force, really, this whole field to rethink this aspect and think how we can really improve. For me, and for Sizzle, it's an opportunity, right? When things are disrupted, things will change, and it's an opportunity really to make a difference. Five years ago, it was hard because it was cultural, it was culturally difficult for things to change. But now, it's pretty clear, the system will have to change.
Now, what we need to hope for is that what we get at the end of it is much better from what we have today, right? And it's not guaranteed, right? You could have a system, as you say, where people don't learn anymore. I don't think about it because I think when we had the calculator, it didn't make people dumber. It made them, they could focus on something a lot more interesting than like tables and multiplication. Math is a lot more interesting than calculation, okay? Essay writing is a lot more interesting than just, what drives me crazy when I see my kids, I have four kids, it's like when they struggle to fill the page. I'm like, who cares about filling the page? In business communication, you try to say in as few words. It doesn't matter how many words you put. It should be the opposite. You should be rewarded for saying the same thing with less words. The education system has always been thinking, oh, you need to fill the pages and use these big words. It's not the right, the good way of writing. But leveraging like an LLM, for example, writing is really interesting. I'm sure you have done it. We're all doing it. It's non trivial. It's not about, oh, let me get GPT to write my essay. And that's a terrible outcome when you do that. It doesn't have your voice. It's very cookie cutter. But getting GPT to tell you, oh, on this sentence, you could do better there, or you could do it in a simpler way, or you could do it in a nice word, iterating like this. That's going to make communication a lot better if we can use these tools.
- [Ryan] Yeah, I think technology is also going to help people that are learning focus on the things that really matter that you should be learning, like you're saying, is it's less, I heard this said, this is a number of years back, I re quoted it to somebody that was a teacher, and they did not like this feedback at all, but I, the statement was something along the lines of given the time that we're in now, it's probably the least valuable to be book smart than it's ever been before because anybody asks you a question, you have access to it on your phone, right? You have access to it through social media, you have access to things. But I still think the ability to use technology to fundamentally understand certain topics and improve problem solving skills, have more active learning, spark curiosity, those kinds of things is what is going to probably pay off the most when people move into the next stages of their life and learn and become professionals and have jobs and just grow with society, and I think oftentimes that's lost on professors and teachers that are stuck in their way of doing things. I feel like there's, you go to every university, there's teachers that teach from the same textbook for 20 years because it's just easy for them. So this is going to hopefully help us stay more, more forward when it, with learning and relevant with learning, which I think is super fascinating.
- [Jerome] I agree. And but the challenge though, is that, so what we learn has changed a lot, how we learn, why we still have lectures as like maybe 70 percent of the time people spend, it's still baffling me. So this is what it is to change. We don't need different, we are learning different things, a lot more interactive, a lot more problem solving, we should learn differently.
- [Ryan] I really like the idea of the ability to use, and you talked about this when we first started about, the Ray-Ban Metaverse glasses to be able to teach you to do like hard skills. And I think there's a large generation of younger people who have just never had taken the time to learn hard skills, like how to fix certain things around their house, how to do kind of these manual labor type things that I'm growing up and you're taught that as a kid by your parents or especially older generations, they know how to fix everything. That's why most kids now call their parents on how to do certain things.
- [Jerome] My kids don't call me though. I'll be honest. They don't call me. I'm terrified. If Sizzle gets there, it'll be really exciting, and I think learning motor skills and with visual input and visual output could be a really exciting way to do things.
- [Ryan] Totally agree. Neil, any last words from your side?
- [Neil] It's been a fantastic discussion and as a guy that teaches a few university classes during the year, I get it. We still teach using a lot of 19th century techniques and there's better cognitive science on how to wire our brains and actually learn knowledge and skills. So, I'm really glad for the work you're doing Jerome with Sizzle. If people want to learn more about you, your work, check out the app, what's the best way to stay in touch?
- [Jerome] Yeah, at this point, we are heads down in trying to make this happen. To be honest, the app is one person done, right? It's a very small thing right now. I think it's very useful. But yes, you can download the app from the Apple store, from the Android store, and you'll just go on, Sizzle AI, and check it out there.
- [Ryan] Thank you so much. This has been a true pleasure to have you on. Love what you're doing over at Sizzle. Just to, your background's super fascinating. I appreciate you shedding some light on the things that you've contributed to, you've done, you've learned with our audience. I think they're gonna get a lot of value out of this. So excited to hopefully one day have you back maybe later this year or maybe I guess going to next year, talk more about kind of the evolution of the education system as AI gets further and further adopted. So, it'd be great to stay in touch and talk again soon.
- [Jerome] Awesome. Thanks for having me.
Special Guest

Hosted By
AI For All
Special Guest
Jerome Pesenti
- Founder and CEO, Sizzle
Hosted By
AI For All
Subscribe to Our Podcast
Apple Podcasts
Google Podcasts
Amazon Music