Episode Transcript
[00:00:00] Speaker A: We need education. We need good educators out there teaching these kids how to use AI responsibly, wisely. And they're going to be the next innovators in the workforce using AI to just accelerate growth.
[00:00:20] Speaker B: Welcome to the Edtech Connect podcast, your source for exploring the cutting edge world of educational technology. I'm your host, Jeff Dillon, and I'm excited to bring you insights and inspiration from the brightest minds and innovators shaping the future of education. We'll dive into conversations with leading experts, educators and solution providers who are transforming the learning landscape. Be sure to subscribe and leave a review on your favorite podcast platform so you don't miss an episode. So sit back, relax, and let's dive in.
[00:01:03] Speaker C: For today's podcast, I am excited to welcome Dustin Stout, the founder of Magi, an all in one AI platform for so many situations, including many higher Ed use cases like enhancing teaching and learning. Dustin's an expert in blending creativity and technology, bringing the best AI tools into one seamless and beautiful interface through Magi.
With over a decade of experience in web design, social media, strategy and entrepreneurship, Dustin is passionate about helping individuals and businesses optimize their online presence. I am thrilled to have him here today to share his insights on AI for higher education. Welcome, Dustin.
[00:01:47] Speaker A: Thank you very much, Jeff. Glad to be here.
[00:01:49] Speaker C: Well, let's start by. I just love to hear your, your founder's story.
Where did the idea come from? What was your situation? And a little bit about the name. I love the name.
[00:02:01] Speaker A: Yeah. So, as a marketing consultant for the last 15 years, I used my share of AI apps. AI was around long before chat GPT, but the problem was it wasn't very good. Not only was it not very good, but the apps that allowed you to utilize it were very clunky. They were predicated on templates, which essentially means somebody else's idea of how to solve a problem. And the problem you had there was this Netflix effect where you log into the app and all of a sudden you're searching for 20 minutes for the right template, and then finally you find one. And maybe it's worth it, maybe it's not. In most cases, it's not. You had to fit your creativity or your goals into somebody else's boxing. It wasn't a great situation, but when Chat GPT hit the scene, it completely changed the game. I started using it immediately as one of the early beta testers before it hit mass market. And I instantly knew this is going to change the world and started using it for my business, for my clients, and instantly also realized like, oh, this could be better. Like, I wish, I wish it did this, I wish it had folders, I wish I could collaborate with my team, I wish I could invite my assistant in and take stuff over. So all these little quality of life things started to pile up. Thought to myself, well, OpenAI has an API, which means I can build an app that utilizes the same chat GPT technology, but I can make it better. And so I did, I went to work fixing all the little quality of life things that I was bothered by to help my workflows improve. And then as I started doing that, started talking to users and getting their feedback and all these other AI tools started popping up. So I'm like, what if I could integrate all of these so that people don't have to have five different AI subscriptions, they can have them all in one place, the same interface, it's all searchable, organizable collaboration can happen in. So that's really, it kind of snowballed from there. But the name Magi actually came from. I was just sitting around thinking like, what am I going to call this thing? I want it to be representative of the experience that I want people to have. And so to me, this new wave of AI, this conversational AI and even these image generators and video generators, there was something magical about it, right? And so I thought to myself, I want to combine this idea of magic and AI. I don't just want to call it magic AI, because that's just a little bit too on the nose. So what if I just combine the two words into the word magi, which in a sense also has its roots in Latin and in the Bible, we hear about the magi that visited Jesus, and the latin word magus actually means magician, or, or even just wise man, wise person, person of science. So I thought that was a good fit for, for what we're aiming at.
[00:04:48] Speaker C: I love the name, it's brilliant, actually. So let me make sure I understand this, because I think you confirmed something from what I discovered from browsing your site is there's all these lms out there, all these options for people now. Like there was just chap GPT people found, and now there's like, how do you decide? And with Magi, you've kind of have the best of, you have so many different LLMs you're using, so you don't need a subscription to all these, it's just, just magi, right?
[00:05:13] Speaker A: Correct. Yeah. So the challenge is there's this arms race that's happening, right? And it's not just a chatbot arms race. You've got the open AI, the anthropics, the geminis of the world, all fighting to be the best AI model. But then you have the image generators, you have the stable diffusions, you have the Leonardo AI, you have flux, and all these image models ideogram coming out, and then you have the video generators, you've got Runway, you've got luma, you've got Sora on the way from OpenAI. So everybody's sort of competing, and especially in the LLM space, I feel like it happens more frequently. You'll have one come out with new features and leapfrog the one before it, and then four weeks later, another one leapfrogs that, and then two weeks later, another one leapfrogs that. So, I mean, it's really absolute chaos. If you want to get the best of the best AI, you have to subscribe to five different tools. But with Magi, we solve that. You only have to subscribe to one tool, and it's all in one place. And every time an AI releases an update we implemented into Magi for all of our users, instantaneously, as quickly as we can, sometimes it falls in the middle of a development cycle. It might take four days, but in many cases, like in the case of GPT 4.0, their last big update for OpenAI, we had it within 2 hours of them announcing it publicly, all of our users had access.
[00:06:30] Speaker C: Wow.
[00:06:31] Speaker A: So it's really cool to be able to just deliver the best of the best to people without them having to go get a new subscription, or I cancel one and go get another.
[00:06:40] Speaker C: So with all of these options and possibilities for people to be using Magi, talk about some of the common use cases for higher education that you're seeing.
[00:06:51] Speaker A: Yeah, I mean, there's the obvious ones of, you know, teachers in higher ed using it to either help them come up with ideas or lesson plans or refinements to lesson plans, or, you know, come up with a different aspects of maybe analogies or things like that. Those are the more obvious ones. But some of the really cool use cases that I didn't think about this, I only thought of the productivity side of it, helping me with my creativity or helping me with my work. But the good people at the University of Notre Dame, John Behrens, who heads their program over there, he's actually giving students access. So they bought a large team plan, and they're inviting their students into their team, giving them access, and they actually have lessons on how to use generative AI. And so I never thought of that application actually using it to teach, because obviously we have a whole generation of students who are coming up in this new AI age, and they're going to need to learn how to use this because it's going to be a big part of the workforce when they go out into it. So I see universities like Notre Dame paving the way and actually bringing this into the classroom as a tool to learn about these LLMs, how they work, and even the image generators. And so in John Barron's own words, he said, he looked around at the landscape and he said, there's no better tool than Magi to allow us to teach our kids how to use all the different models in one way. It just makes their classroom teaching so much easier. I thought that was just a really cool something that I never thought about. But now we're really working hard to make even better for those schools because I truly believe the teachers in the field are crucial in this stage of our culture, the shift that's happening, and we need education. We need good educators out there teaching these kids how to use AI responsibly, wisely, and they're going to be the next innovators in the workforce. Using AI to just accelerate growth.
[00:08:51] Speaker C: Yeah, that's a really great lighthouse to have the University of Notre Dame, they're one of the leaders. So to have them using Magi is really kind of a testament to your tool. I think it makes me think of the governance.
From what I've seen out there, most schools haven't quite got around too much.
They're working on their readiness and their governance now, and it seems to be pushed down to the faculty level at this point at a lot of schools, even k twelve. So are you seeing that as an opportunity to really have a faculty level adoption, or do you, do you try to sell it as an enterprise model or kind of both?
[00:09:35] Speaker A: Yeah. I mean, for the large schools and businesses, we do have enterprise plans that kind of cater to how they need to access it, because our regular plans, they're typical software subscription base, where you enter a credit cardinal, you pay monthly. But with schools in particular, you know, they have different requirements, the different ways they need to pay for software. So we've accommodated that for them. But we've seen some, some really good adoption. I have a friend who works at the local school district here in Kern county. He's the IT guy down at Kern, Kern school District, and he's implementing it for their admin staff and, you know, doing all kinds of things and he's leading the training for this because obviously there's a stiff learning curve for people, you know, been doing things for a long time. And, you know, in education, you've got no shortage of new things to learn and keep you on your feet, right? And so AI is just another one of those things. And so it's great to have really smart it people, you know, adopt this in, play with it, explore it, and then start deploying it into the admin staff to help them with, you know, tasks that are, that AI is actually really, really good at and can help them get things done faster and feel less burdened by just the mountain of work that they have to do.
[00:10:48] Speaker C: I go to a lot of edtech conferences, and what always comes up, everyone gets pretty excited about it, but there's always a big contingent that are very hesitant and nervous about the ethical implications of AI. So what's your take on that? What do you see in higher ED and how do you answer those questions?
[00:11:09] Speaker A: I've thought a lot about this because my wife is actually a teacher, she's a middle school teacher. But, you know, the, the ethical implications of this, I think are the most important things to think through before starting to use it. And as I've thought more and more about it, I mean, really it comes down to the authenticity and the responsibility of the user to say it is. It is my responsibility to make sure anything that I do with AI is checked by my eyes and is approved by me. So it's not like I'm just hitting the YouTube generate button and sending off whatever the AI spits out, right? In business.
As a consultant over the years, I worked for many large companies, many small companies, and one of the things that we did for them all the time was copywriting. And most importantly, we did organic content marketing. So we wrote articles for them. And so the people who owned and ran the business or worked there, they weren't writing the articles. They hired us to be their ghost writers. And so I see AI in that very same way. AI is just a ghostwriter. And the only difference is it's not a human doing the ghost writing. But at the end of the day, it's still up to you to check the work, right? You still have to have a human in the loop. And so for me, there's really no difference there whether an AI generated or a ghostwriter generated or a personal assistant wrote it. It's all the same. Because at the end of the day, you as the responsible human being whose reputation is on the line, you have to make sure that the content is up to your level, up to your standards, and represents you in the best possible way. And so for me, the ethical concerns kind of dissipate at that. As long as you're open, transparent, and you yourself are doing all the check work, you're being the editor in chief, so to speak.
[00:13:03] Speaker C: I've heard that echoed different ways from so many AI leaders out there. And I think the way I like to sum it up is, and the way I feel to the way I use it is it has to be a starting point on any point, and it may end up being less of a starting point, more of like, you're always going to have to check something in the near future before it's consumed by your, your audience. But yeah, that's, that's kind of a thing we all need to understand, I think. So can you offer some tips if college faculty want to start using AI in their classroom and they're just not sure how to start, I think you've touched on this a little bit, but what are some great ways that faculty can start using this in their classroom?
[00:13:48] Speaker A: Well, I think, again, for me, the thing that I tend to use it most for is analogies. So I speak at a lot of conferences and I speak to different groups of people. I do a lot of these podcasts and teaching myself. And so I'm constantly trying to evolve my analogies to the audience that I'm speaking to and create connections that they can relate to. And so I often use AI, like saying, here's a concept I'm trying to explain to x, y and z audience. What's a good analogy I can use and have it brainstorm ideas with me. That's to me, I feel like I do that several times a day, every day, just give me ideas that I can come up with different analogies. And another great use case is help me outline this lesson. I want to hit x, Y and Z points. What are some things that I'm missing in this? Or what are some alternative angles I can come at this outline with to make sure that my audience gets the fullest possible experience? So really just using it as a creative partner in every aspect of, you know, building learning experiences or building lesson plans and so forth. And of course, the image generation side, that's pretty obvious. So my wife, one of her jobs, she's an Ela teacher, so she teaches English, reading, language arts, but she's also in charge of the yearbook and the school newspaper. So she actually, last year she used Magi a great deal. The image generative capabilities to help with the yearbook, you know, backgrounds, textures, you know, things that are more, you know, out of focus, because really, the focus is the photography, but you still have background elements and so forth. And so that was a pretty cool use case for her for newspaper and such. But, yeah, I mean, she's also used it to help her find missing elements in her lesson plans and to sort of critique her lesson plans. And so that's another thing that AI is good for. If you want it to help you critique something, then it'll give you some great feedback when you can't get that from, like, a human who doesn't have the time to read your 15 page lesson plan for the year.
So, yeah, lots of cool ideas out there.
[00:16:00] Speaker C: I have a question for you, Dustin. So we've seen the evolution of the text, the LLMs, in the last few years. Everyone's seen this now, like, how much better they're getting. We've seen even the image generation. I saw someone present, you know, 2021 through 2024. Like an image, a prompt. What would that, what would the output be for this image prompt? And it's now, it's almost like you can't tell from a photograph. But now what seems fairly new relative to these other use cases is voice. Where are you at with kind of the realistic voice generative AI, if you're even going that way? Is that on your plan?
[00:16:40] Speaker A: I mean, it's definitely on my radar. It's just not a modality that I am drawn to. So some people, that's their natural modality. They would much rather hold up their phone and hit record and speak to an AI and then have it speak back. One great use case I heard was, when you're in the car, you don't want to have to look at the screen or touch the screen and OpenAI's new voice mode where it's like a continual conversation where you don't have to really manage it, you can just kind of set it and talk back and forth. I think that's a really great modality. It's just not one that I'm drawn to. Some people are drawn to that. So I think there's definitely a lot of great use cases in it for people who are verbal processors. I'm a finger processor. Like, I type to process my thoughts, but that's just me, so I think it's very helpful. I haven't found many practical applications for voice output in so much as in production. So podcasts is like the one great use case. If you don't want to use your own voice. I don't know why you wouldn't, but there are some people who want to have sort of like that automated podcast type thing, I guess that can be very helpful. And for the hearing impaired, you know, it's a really great upgrade to the tools that have been available for the hearing impaired for a long time where something can read out text to them. It's very helpful. So there are definitely some, some great use cases out there. You know, it kind of just really depends on what your mode of operation is. And for me, the work that I do, it's often text input, output based. And so we spend a lot of our time focusing on that.
[00:18:15] Speaker C: I'm similar to you, I think, in that I prefer text to soak in information. And I kind of wonder if it's a little bit of a generational thing, like, could be. I think I'm older than you, but like the young, the Gen Z's and younger, I think, are much more apt to like intake audio learning. I don't have any data on that, but I'm sure it's probably out there. And I'm thinking about this podcast thing because now there's a few tools out there, they're doing incredible work. But I'm like, why would I ever use this? And I think if I was in a classroom, I might like to listen to a summary of a podcast of a giant document. If I was riding my bike, you know, that might be kind of something. I would do nothing where it's entertaining, though. I wouldn't want an AI, I think, to entertain me in any way. So that's kind of how I'm thinking about that.
[00:18:59] Speaker A: Yeah. And I can totally see it. For summarizing, one of the things that I recently used it for was I was looking for some. There's a lot of grant programs out there for AI research, and we pretty much all we do is AI research, research and then implementation. So, like reading these grant programs, like, how do I, what are the requirements? And sometimes they're very long and dense documents. And so I ran one of those through Magi and got a text summary of it, and then I can place that quickly into an audio processor and just like have it read it to me while I'm designing or doing other things. And, you know, that's one way that is very helpful in that aspect. If you don't have time to read or energy or you want to multitask, maybe that's a great use case for. But one of the rare times that I actually use voice output.
[00:19:47] Speaker C: Right. So in your journey over the last few years, what are some of the biggest challenges you've faced in your magi to higher ed specifically?
[00:19:58] Speaker A: Yeah, you know, the. I think the biggest challenge was we weren't ready for, for higher ed at first. I was mainly, you know, building this for businesses, marketers, for, you know, small business owners, you know, my background. And the biggest challenge was realizing, oh, we do have these schools who are using it and they have some great ideas about how to use it. How do we shift the infrastructure that we've built to be more accommodating to them? And that was a real. The team plans were not nearly as comprehensive when we first got our first higher ed customers. And so we had to really think hard about, you know, how do these team structures work and how do we allow for having multiple teams and allowing instructors or the IT department to sort of segment off the different teams so they're not intermingling and, you know, all that privacy related stuff. And so, like, I think the biggest challenge for us was just that initial shift to, oh, we have higher eds interested. Let's. How can we make it easier for them? How can we build a system that is just more conducive to what they're doing? And that was really the biggest challenge. And then after that, really it's just a matter of staying on top of technology because, as you know, AI is moving quite fast and there's always something new coming out, always, you know, improvements and updates, especially when you're supporting so many different AI protocols and models and modalities. So, you know, really the challenge is just staying on top of it, making sure that our product is working at its peak, but also pushing the envelope a little bit, staying on that bleeding edge of technology so that our users have not just access to it, but the best access to it and the most accessible way to get access to it.
[00:21:44] Speaker C: Yeah. Yeah. So how do you ensure, like, if students, in fact, they are using this for a lot of their information, how do you ensure the privacy and security of that data?
[00:21:56] Speaker A: So I'm a privacy first guy, always have been. We eliminated Google Analytics from our websites.
Everything that we do is privacy focused and our server stack is just the hardest encryption possible. We have a policy that we do not use any AI model that utilizes user training. So with tools like chat, GPT or cloud or all these AI tools, you have to opt out. You have to manually, manually go find the setting that says, do not use my chats or conversations or my input for training models with Magi, it's by default, you're opted out. We will not ever sell your information, use your information. I'm not building an AI model. I have no interest in building my own AI model. So we won't use it for training your info. Your data is always private secured and the highest priority to keep it that way. And so, you know, if a model updates itself and changes its terms and says no, you can't opt out. Sorry, we're not going to support it. So that's just our commitment to, you know, keeping things private. Everything is, you know, 100% yours and will never be used to make anybody else wealthier or, or make their models smarter.
[00:23:10] Speaker C: Seems like one of the huge benefits to Magi, I would think, right there.
[00:23:14] Speaker A: Yeah, I mean, it's, it's nice that these other apps offer the ability to opt out, but you still have to be aware of it and find it. And there's been certain situations in the past month where OpenAI, some trigger got happened, and then everybody who had opted out was suddenly opted in again. And so you kind of have to stay on top of that with Magi, you don't because it's always by default.
[00:23:37] Speaker C: How would you recommend higher Ed get ready for AI?
[00:23:41] Speaker A: The only way to get ready in all reality is just to, to dive in.
I use this analogy a lot when, when Google came on the scene, we're old enough to remember that right before the Internet, before Google, the Internet after Google was very different, but we didn't really know how to use Google.
What are we going to use this for? Like, I remember my first search was something like, charlie bit my finger, right? That funny YouTube video. That's all we were using for funny stuff on the Internet. But then we realized, oh, I can use this for research. And we kind of learned how to change our language and how we search for things. And as we started to use this thing called Google search more, we realized there was more we could use it for. So the more you use a thing, the more you realize you can use it for. And really the only way to speed up that curve with AI is to really just dive in and start using it, start exploring it, put something in, put in a problem. Here's my lesson plan. What else can I, what, what am I missing? You know, just ask it a question. You know, here's a, here's a long email that I just got from my team. Can you summarize it into 15 bullet points for me or five bullet points for me? You know, just kind of put stuff in and ask it to do things. Think of it like your personal assistant, a human being who knows everything and all it needs from you is to tell it what to do. And then you'll start coming up with ideas.
[00:25:12] Speaker C: Well, thank you, Dustin. We're gonna close out this episode and just I'd like to ask you, how can people find you?
[00:25:20] Speaker A: Yeah, just head straight to Magi co Magaico and you can explore all the fun features and options we have there. I am on all the social things. Happy to answer questions and I give you ideas. I'm an extrovert, so I love talking to people. I love sharing. I love teaching. So happy to connect with, especially higher Ed because again, I think you're training up the next generation of AI superstars. So it's up to us to do our best to steward this new information and to hand it off to the next generation to do great things with.
[00:25:56] Speaker C: Perfect. Thanks, Dustin. You can also find Magi on edtechconnect.com. and that's the show. Thanks, Dustin.
[00:26:04] Speaker A: Thank you.
[00:26:14] Speaker B: As we wrap up this episode, remember, edtech Connect is your trusted companion on your journey to enhance education through technology. Whether you're looking to spark student engagement, refine edtech implementation strategies, or stay ahead of the curve in emerging technologies, Edtech Connect brings you the insights you need. Be sure to subscribe on your favorite podcast platform so you never miss an inspiring and informative episode. And while you're there, please leave us a review. Your feedback fuels us to keep bringing you valuable content. For even more resources and connections, head over to edtechconnect.com comma, your hub for edtech reviews trends and solutions. Until next time, thanks for tuning in.
