Jonathan Green: [00:00:00] Is it safe to let AI put your kids to bed? Find out on today's episode, I.
Today's episode is brought to you by the bestseller Chat, GPT Profits. This book is the Missing Instruction Manual to get you up and running with chat g bt in a matter of minutes as a special gift. You can get it absolutely free@artificialintelligencepod.com slash gift, or at the link right below this episode.
Make sure to grab your copy before it goes back up to full price.
Are you tired of dealing with your boss? Do you feel underpaid and underappreciated? If you wanna make it online, fire your boss and start living your retirement dreams now. Then you can come to the right place. Welcome to the Artificial Intelligence Podcast. You will learn how to use artificial intelligence to open new revenue streams and make money while you sleep.
Presented live from a tropical island in the South Pacific by bestselling author Jonathan Green. Now here's your host.
Today we have a very special guest, the founder of Storytime ai, and I'm very interested in this tool. Very rarely do I start with the name of the tool at the beginning because I've been using it with my kids for the past week.
And it's very simple. You get to write a story with the kids that's somewhere between one and five minutes in length, and they get to choose the characters, the setting, the location, all sorts of things. And I was really curious to see. My kids would like it. 'cause it doesn't matter if I like it because it's not for me.
And that's an important lesson to remember. Every time I watch a movie with my kids, I don't like what they're watching. It's so boring. But they don't like the cartoons from when I was a kid. So I'd love to just get from the beginning, explain the concept and where the vision came from, and then we'll dive into my experience.
Brian Carlson: Thanks for having me today, Jonathan. I appreciate your taking the time to, to have me on your show. Yeah, the concept first came out. I've always been involved in education technology. Former teacher spent a couple years teaching overseas in the island nation of Vanuatu in the South Pacific.
And came back, did a little bit of work in the school system before I switched over in education technology. Worked for a larger education technology corporation, which eventually led me to start my own business, which was online education. And always been involved in that field.
Had my last business, had an exit to a group called Learning Technologies Group. Took some time off, at the right time, had a nice break, and then the AI craze started. So the first question was, what is this going to have in terms of value? Where's the good going to come, the impact going to come.
For education specifically, and landed on a number of different themes that we saw that were going to be really interesting. And one of them was representation and an inclusivity and content, which has always been lacking. We looked at first, what are the AI models doing right now versus where they're going to be maybe a year from now.
And that was why we started with short stories for a younger audience. Aspirations are certainly and roadmap has us going into longer text as we evolve, and we're starting to work on that as well. But for right now, we wanted highly illustrated books. So that's really what I.
Also was part of the genesis for why we started story time was because we wanted to take the best image model options that we had with the best, that we could do at the time, which was, and we started this a year ago, which was a shorter story. Put those two together into an app.
Jonathan Green 2024: So I guess my question is what's the problem that you're trying to solve?
Is it. Kids getting excited to create their own stories. Is it kids getting learning to read? Is it kids spending more time with their parents? I'm trying to understand, so then I'll direct my questions. 'cause this is the first thing I was trying to figure out from the app and I wasn't quite sure because it fills a couple of different niches.
But what's the ideal goal? What's the ideal result?
Brian Carlson: Yeah it's shifted. It's a great question. The original goal was, like you said, was bedtime stories. It was, okay, let's have this as a, as a way for kids to parents to have an easy way to create a bedtime story that would be interesting.
We quickly pivoted and realized there was a lot of other things that it started to solve. I'll start with one of the more unique use cases that we had not thought of that came out right away, which was. For children with special needs and neurodiverse population specifically, we got picked up by a number of different teachers in special education where they were using this for social stories, which is a big part of the IEPs, the individual education plans for children with with with special needs where they're going into a social situation and they're maybe nervous about it.
Or it's new to them. So you build a story with them as the main character in that story, and then you show them what it's gonna be like. We're gonna go on a field trip next week. Here's what you can expect. Here's what you need to be aware of. Here's what you need to do to be ready for it. And that puts the student at ease.
There's other use cases and I think the. That's a one that emerged. We're finding that mul we're finding multilingual parents are a really good use case. You are, you're Indian American and you wanna keep Hindi in the home, but there aren't good materials and you can do this now where you create a personal story about your kid in Hindi at a young age.
So they're starting to see themselves in story for that specific language. But I think the main use case that we're looking at is representation. In books matters. If you see yourself as a doctor, as a young kid, you become a doctor. So seeing someone that looks like you in the story, that's a big part of engaging and the young reader in wanting to not just read more but start to see themselves in specific roles that they might not always have that type of book in front of them.
So it's easy access to that.
Jonathan Green 2024: It's interesting to me. My kids are multiracial. They speak multiple languages, and they did ask, of course, nobody has. There's their language, their native dialect is very obscure. So it's not in there yet, but I maybe someday it's like the 500th most popular language. So it's very far down the list.
But that was the first thing they we're thinking about and it's, I was like you guys speak four languages, use the other three. The other three are in there. So that is very interesting to me. I think that when I was playing with a toy with my kids, the experience was the first thing that they quickly wanted to start putting themselves in the story.
So the first time they did a story. My son did a story about a girl, so it wasn't about him at all. It was a very different story. But then after that, every story was about us and we tried a lot of different ideas. And I guess that's what I was wondering about with the learning to read because they quickly switch to the video element, which seems to be where it's, the direction it's moving into, which is video stories, which is a different solve, where they want to create the story and then.
It's hard to get kids to re my kids and me as a kid to read a book over and over again. You might read the same book two or three times between childhood and 18, whereas they'll watch the same movie over and over and over and over again. So they watch the videos over again. And then of course they wanted to start making stories about each other because that's how they played the game.
And I was like, wait a minute guys, let's. Pull back from that. The, that's the escalation which they wanted to try, which is where they would start writing stories about each other and including the names of their crushes and silly stuff like that. I was like, guys I love the idea, but that's, they were interested enough to wanna go in that direction and they would watch the videos over and over again.
And I did find that for us at least, the shorter story was the sweet spot. So the regular, original story size, which about two and a half minutes, that worked perfectly. And I always think about when I shoot videos of my kids. Doing a video when I'm holding, it feels like it's forever. And then I get home and it's 19 seconds holding a camera for five minutes.
I just can't do that. So my first question is about the guide rails. So we pushed against it. So in the second or third story. We had a boy with his golden retriever solve a murder mystery on a train. I was like, let's do the Orient Express. My kids were like, what's that? I said, classic. I got thy, I'm sure you guys are fans.
So they did solve the murder, which was amazing, but it didn't show the murder. So the images had a rail, which was fine. I was like, I don't need to see it. But they were like, wow, this is very exciting. And that's how far we were able to push it. We weren't able to get past that, but I'm always, I would rather test that with the kids than.
They do it when I'm not there 'cause they're gonna try stuff like that. And it didn't, it was very interesting to see, 'cause I'm like, how graphic is it going to get? And it didn't go beyond, there was a murder in the room. We have to solve it. So I. When you're creating the rails, I guess that's what I really wanna, how do you determine what the line is gonna be?
I know there's strategies of we're gonna ban this word, but there's always a way around that, right? So how do you handle that with one kids creating malicious stories about each other, and two, creating stories that are like more extreme.
Brian Carlson: Yep. Yeah. First of all, fascinating to hear the trajectory of what your kids were doing and what they wanted to see.
The video element was it came outta user feedback because parents also wanted to have the ability to create a reel, create a playlist, and we have playlists in there, I'm sure you saw, where they can create an hour's worth of personalized edutainment that they've vetted and then they kids would wanna watch that over and over again is just what we saw.
So that's interesting. As far as the guardrails, I'm very glad to hear that's what the outcome was. We've put a number of, obviously, like you said, filters when it comes to band words and things like that, and that can only take you so far. So the next layer that we have above that are image classifiers where it will, if something comes back for some reason, that is suspect, it will stop that from coming back.
And it may ask you to recreate the image. You may have gotten a, an image not available, recreate it. And it tries to make sure that nothing comes through. Over time, as we go to older audiences, we will obviously not have the same guardrails, but as of right now there's multiple layers and it's obviously very important for us to do to ensure that's, that's not going to be something that comes across. There's there are ways we can train the image models. I know that's something that that we've talked about. So we're also pretty careful in trying to make sure that the image models that we've used have been trained on image images that we would find safe.
Jonathan Green 2024: Yeah. I'm definitely more worried about an inappropriate image than I am about inappropriate language. Is an image. I still remember the first time I saw something in a movie I shouldn't have. 40 years ago, that six way longer than something like [00:10:00] language. The next question I have, I was thinking about this a lot, is the stories are descriptive rather than action based.
So it will say, oh, she went on adventure and saved her friend and they had an amazing adventure. But it doesn't say what actually happened in as much detail. And I wonder if that's because it's for younger children, if that's intentional or if that's more a limitation of the language model and what your approach is gonna be going forward.
Because I just think about that as far as replayability or longevity. Yep.
Brian Carlson: Yeah, you're right. There's no doubt that it is it needs to be robust for it, it is partly a limitation for where we started with the language models. We're in the backend, we're testing some different language models already that are having significantly better results for adding dialogue, for example.
You're not gonna see a lot of dialogue right now in the initial iteration, but we're changing that here over the next month or two. And we may even get to a spot where we offer you the option to pick between different models or if you pick mystery versus if you pick action, it might in the backend, be routing it to a specific model based on which one writes better for a specific specific objective.
There's, there is opportunity for us to train. Our own models on different data sets for specific instances, we haven't gotten there yet. But it is something that, we do look at and we do have on the roadmap as a strong possibility. Yeah. '
Jonathan Green 2024: cause what I was thinking about is the replayability of.
How many times will my kids re-watch the same video or re-engage with the same story and what they're drawn to? And obviously my calibration of what's interesting is not the same as theirs. So I let them really dive into the app and they were playing each other's videos and go, look, there's a story about me and my youngest daughter who wasn't there, was like, Hey, where's the story about me?
And so we make one about her. And so they each wanna be involved. So they're definitely all feeling that pull. So it was working in that way, so it wasn't. Negative. It was just like, oh, when something works, I wanna build on it. The element that I really wanted, which we talked about last week in the pre-interview, was I would love for it to.
This is me less the video and more their text has a bouncing ball and they can read along with the book. That's really the element that I want because I'm trying to get them engaged with reading. What I don't wanna accidentally do is create another just video app. So I want something where they're heavily invested in the reading element, but of course they prefer the video like they always do.
So it's, that's a challenge I face as a parent as well, which is they need to learn how to read. That's a pretty important skill and it's where we always struggle because it's hard to get them engaged. Yep.
Brian Carlson: Yeah, so we started where we added the read along we call, that's called speech marks, and we added that on the video first.
We're moving that into the book side next. But at the end of the day it is a challenge because we're finding the same thing. It's why we added the video component, and it certainly. It also was what parents asked us for. They wanted the ability to set it and forget it. They wanted to give their phone to their kid while they're at dinner or making dinner and, with friends and they don't wanna be bothered.
So there's an element of how do we get guilt-free screen time? That concept came up quite a bit and they're saying, listen, it's not perfect in the sense that we're giving the phone to the, to our kid and we're having, to mitigate a certain amount of screen time. But at the same time it's significantly better than the alternatives that they were currently having.
Their kid look at videos that might be on YouTube that are obnoxious or maybe promoting the wrong values. So they could at least control what the message, 'cause you could, you saw when you're in there, you can insert themes. So if I want to insert friendship with, with my brother as opposed to, what they're seeing on YouTube it at least is a little bit more healthy. And you do have the reading element at the bottom of the movies. Are they watching and reading as it goes through? I'm not sure that's the case yet, but at least it is enforcing that concept of reading and creating your own your own books and
Jonathan Green 2024: imagination.
Okay. That makes sense. I'm not like that at all with my kids, that's not my approach to the phone. I don't use it to silence them. I'm very involved. Like I'm with them all the time when they're using it. It's the only time I've ever let them use my phone. Yep. So my, I'm a super engaged parent, which I'm fortunate 'cause I can where it from So I understand why parents want something where they can give and not be nervous.
'cause I certainly know there's issues with YouTube security. I know there's issues with, iPhones have no, the parental controls on Apple don't actually work, but they haven't told anyone they've been broken for four years. So you can set screen limits and app limits and they don't work. They just turn off on your iPad after a couple of days and Apple doesn't tell you.
There's so many forms about that. So that's something that really bothered me. When my kids started to break through all the limits I'd put on the iPad, I go, what happened here? 'cause they have educational games and stuff they're not supposed to be looking at. And they broke through and I found out.
That really bothered me. So I understand the desire to put something safe, and it's like you're caught between these two goals, which is, I wanna give you something educational, but I also wanna give you something that's entertaining and that's, yep. A very narrow road to walk. What do you think is the longevity?
I guess I was looking at the pricing model, which is you get, I think 10 credits per month for a certain price, and it's if I'm making a video, that means I can do five. Stories plus video, which is about 12 to 15 minutes of content based on the average story length, is around two and a half minutes when we test it.
So if I can only make 15 minutes of content per month, it's gonna take me four months to be able to make an hour's worth of content. And I wonder about Replayability, what's your experience been with that? Of kids coming back, staying with the app longer, actually continue to use it a year later. 'cause those are the hardest things with any app.
Brian Carlson: We've only been out the door, here about five months since we launched the app. And we stayed pretty much in stealth mode. No marketing to date, we're just getting started with that. So we get user feedback. So we don't have a lot of data, but we are seeing that kids come back and watch the same videos over and over again.
It, you do have the ability to buy extra credit bundles, right? So we are, we're in this balance with trying to find a because we do have costs, right? We have the inference of creating the images each time. So we're having to try to get to a spot where we're gonna be able to cover the operating costs with the pricing model.
And I think that the good news there is those prices will come down very much over the next year. We're just trying to get it a little bit above break even for right now for us as we push the pricing down. With the image and the text models over the next 12 months. But what we're finding is a lot of people are buying extra credit bundles, so we're trying to make it as accessible for everyone as possible.
So we're far, for example, we have competitors out there like Epic, and you and I talked about Epic here offline. They have a library of books and no videos or I don't believe they have videos yet. Books only. And they're limited in languages, and it's $70 a year for access to those, to that library.
We're gonna have our library fully free with. Hundreds of thousands of books and videos, and currently we have 22 languages we'll be building to 74 this month, and that's free there. So the only cost that we have is if you want to build a personalized story and that comes out to about, depending on what model you look at, about 30 cents a book.
So if somebody wants to go above and beyond the 10 credits, then there is obviously the ability to buy bundles at prices that are, we think very affordable compared to traditional options.
Jonathan Green 2024: I certainly didn't think that the pricing was unreasonable. Considering what similar tools that I use to create videos for adults charge that make a short form video that's 30 seconds, and they charge, they all do credit systems, so you can't figure out what they're charging, but I looked into the numbers.
They're all significantly more expensive, so I didn't think it was an unreasonable price point, and that is a hard needle to thread because a lot of people will, if it's at a certain price, they'll just go, oh, I can just figure out how to do this on my own with an AI tool, which is. Real hassle. I use AI a lot.
I'm like, I don't wanna do this myself. It'd be an hour for me to create, to replicate the process and create each image. 'cause you have to render images one by one and then put them together. The story, I'm like, it's not an easy process. So I use an automation tool for that reason. So I definitely found it interesting.
I think that you are entering an interesting market and it's trying to. Go in a positive direction for ai, which is why I wanted to shine a light on this tool. So my kids did like playing with this. We'll continue over the following weeks to see what's possible, to see if they stick with it and keep wanting to play with it, because that's really, I can get 'em to play with anything once or twice.
So that's really give you the test of time. So I'll definitely follow up with this interview in the future. But thank you so much for being here. Where's the best place for people to check out your app, find out what you're doing, and connect with you if they have more questions and wanna see what's going to happen next for story time.
Thanks
Brian Carlson: Jonathan. I appreciate you having me on today. The best place to find me and find Storytime is find us@storytimeaiapp.com and there's a download link there and that will get you in and get you a couple free credits. And feel free to ping us on socials and anyone who's out there who wants to test in your audience, we're happy to add some free months of credits in there for them.
So just ping us on socials and we'll set
Jonathan Green 2024: you up with that. We'll put all the links below that, below the video version and in the show notes you're listen to audio version. Thank you everyone for being here and listening to another amazing episodes of the Artificial Intelligence podcast.
Thanks again, Jonathan.
Thanks for listening to today's episode starting with ai. It can be Scary. ChatGPT Profits is not only a bestseller, but also the Missing Instruction Manual to make Mastering Chat, GBTA Breeze bypass the hard stuff and get straight to success with chat g profits. As always, I would love for you to support the show by paying full price on Amazon, but you can get it absolutely free for a limited time@artificialintelligencepod.com slash gift.
Thank you for listening to this week's episode of the Artificial Intelligence Podcast. Make sure to subscribe so you never miss another episode. We'll be back next Monday with more tips and tactics on how to leverage AI to escape that rat race. Head over to artificial intelligence pod.com now to see past episodes.
Leave, review and check out all of our socials.