AI Creation Made Easy - Dustin Stout’s Guide For Content Creators

Jim Fuhs [00:00:01]:
Dustin is is an amazing guy. I've as I as we like to joke, Chris, I've known him since the days of Google Plus. He is just someone that when you see the sort of stuff he puts out there, you listen. And so when I saw that he had started Magi, I was like, okay. I'm I'm interested, and I have not been disappointed, as as I even said, and maybe hey. We'll make a t shirt out of it, Dustin, but, I'm I'm Magi for life. I'm telling you.

Chris Stone [00:00:30]:
Woah. Woah. Woah. Woah. Woah. Woah. Okay. Well, we're just gonna like, I can't do a better intro than that.

Chris Stone [00:00:36]:
Would you please welcome to Dealcasters the one and only Dustin Stout. Dustin, welcome to that match

Dustin Stout [00:00:44]:
Magi for life. Right? Oh. Glad to be here, guys. Thanks for having me.

Chris Stone [00:00:49]:
First time guest of the show. It's it's, it's been a while. We've been hearing a lot about all the great things about Dustin Stout. And now this thing called Magi, which I I I really wanted to set the table for our listeners and our viewers because I think many people think magi.co is just another AI thing. Right? That it's just another generative platform where it does whatever you wanted to do, put in prompts and whatnot, but it's not that. Right? It's it's really a lot more. So I wanted to roll and and set the table, you know, and and roll this over to you. If you're gonna explain this to someone who thinks that it's just a generative AI platform, like Chad GPT or like Runway or something like that, what is Magi?

Dustin Stout [00:01:39]:
Well, I think you did a great job. I I don't know if I couldn't I don't know if I've done a better job than that than than you did in your intro, but really it's just that. So there are a lot of AI tools out there. And right now, we're in an AI arms race. Yeah. You got the Open AIs of the world, the Googles of the world, the Metas, the Anthropics, all these companies, Runway, Midjourney, all these companies fighting to be the best, the best AI in their vertical. Right? You got the text generators, the chatbots, and then you got the image generators, the video generators, the sound generators. It's an arms race.

Dustin Stout [00:02:16]:
And in order to get the best of the best tools, like you said, you have to have like 5, 6, 20 different subscriptions. And first of all, who has the sanity for that? Right. Just happened to learn 20 different apps and manage content in 20 different places. And then on top of that, the financial burden of having to pay each subscription each and every month. Now what Magi does is we take this technology, these exact same apps that you would pay a monthly subscription for. We bring them into one platform, one united interface so that you don't have to relearn a new interface every time you wanna try a new chatbot. You can get ChatGPT, you can get Claude, Gemini, and Meta's AI, and a bunch of others right inside the same chat interface for 1 monthly price. And all likewise, all the image generators.

Dustin Stout [00:03:13]:
I think we have something like 9 or 10 different image generators at this point, maybe 12, if I'm if I remember correctly. It's a lot. We have a lot of them. And there are multiple reasons for this, really. It's not just the cost savings. It's not just the sanity of it all. But to be honest, like, we are all learning this stuff. We are our entire culture of the human race is having to relearn things, how to use this new technology.

Dustin Stout [00:03:46]:
And we all all of us were around during the Google transition. Right? We we knew what life was like before Google and what life was like after Google. The only way you learned how to use Google was using Google. And so the the best part of Magi is, I think, the ability to experiment and explore all the best AI technologies from the comfort of one place. A new one comes out, Magi adds it, you get to try it. And so I think that more than anything gives giving you the freedom to experiment and explore helps that learning curve shorten, big time.

Jim Fuhs [00:04:26]:
Yeah. I I had actually wrote it down, Dustin. I don't know if I'm right, but I was like, at at this moment in time, there were 15 image or video, programs, I guess we'll call them, and 26 large language models, LLMs, on MAGI. That may change by tomorrow because that's one thing that I love. You're constantly updating, you know, and and then, of course, I couldn't even tell you what probably 2 thirds of them do because it's just like you said, it's so much, but I like that it's there. And, you know, as you start educating the community more about how these things work, I I think that also makes it comfortable for people that are just getting into this because I think it's very important to your point. AI is not going away, and I also love your description of this arms race because, right, you're gonna see companies come and go. I mean, it's it's fascinating because I wanna say we used it a while back and maybe it's still around, but in a completely different format, and that was Jasper.

Jim Fuhs [00:05:29]:
Right? Jasper was charging a ton of money to do that and, like, well, now you can do much more than Jasper just by, you know, using MAGI. I mean, what do you what do you say to, like, I'm a business owner and I need to know, you know, how should I get started with something like this? Because I think one of the things that I think and we talked about this a little bit before is it's I don't look I think people have to be careful about thinking like, oh, this is gonna replace my workforce. I think it can enhance your workforce, but I think don't think, hey, this is just gonna write stuff, and we're gonna make tons of money without any human interaction.

Dustin Stout [00:06:11]:
Yeah. I mean, so businesses love Magi. And and the reason businesses love Magi is, again, going back to the the savings. Right? So we have team plans where you can start at as little as $29 a month and have up to 5 team members, use Magi. And we that goes all the way to, I think our largest plan is, a 149, I believe, or 249. You have up to 50 people in in one team. So it doesn't have a per per user pricing. Now we get into the enterprise plans.

Dustin Stout [00:06:46]:
We do kinda do a per user pricing, but that's really just a matter scalability for making sure everybody has the resources they need. But, you know, businesses love Magi for that aspect because, again, they're not having to pay 5 subscriptions per employee Right. To get everything in 1. Billing is easy. IT teams love us because it's like they don't have to manage all the accounts now. Right? So, business businesses love us. Any business who's looking to integrate AI should be using Magi as opposed to, you know, all these other random subscriptions because you get the most powerful AI models. There's there's nothing that ChatGPT can do that Magi can't do.

Dustin Stout [00:07:24]:
And in some cases, Magi can do more. We have little tools and quality of life things built on top of this that was really born out of, like, my frustration with ChatGPT and my limitations. So, you know, really the best way for businesses to dive in. Try Magi, get it for your team, and, get it for all your employees. And it, will be a much more cost effective and more resource less resource intensive in terms of managing the subscription and managing the team than, any of the other apps out there.

Jim Fuhs [00:07:55]:
Yeah. I mean, I I gotta be honest with you. Like, at at first, I thought, like, there's gotta be a catch. I real I I I totally did. I I was like, okay.

Jim Fuhs [00:08:05]:
That's Chris.

Jim Fuhs [00:08:06]:
Where is where where are you gonna hit me? Like, where is this gonna like, it's gonna hit me in the grapes at some point. Like, I know there's gotta be a catch. Like, you know, all these, generative AI, all these all these image play, all these chat, all these things have their own subscription plans. Right? And you get locked into them. And and if and most people go monthly. Right? Because even though it it costs it it's like, you it's more cost effective for you to get a yearly plan. It's like after 3 months, you realize I shouldn't have gotten this yearly plan. Right? And so you you you end up cutting it.

Jim Fuhs [00:08:45]:
But, like, you all these things are, you know, a la carte. You gotta pay all these things. And somehow, you've cobbled them all together, right, into this, you know, collectively smaller price. Like,

Dustin Stout [00:08:59]:
I I just how

Jim Fuhs [00:09:00]:
do you how on earth do you get away with it?

Dustin Stout [00:09:02]:
So it's funny you you say that. I actually have some ads that I am getting ready to test that, that are based on this series of TikTok videos that I've seen recently. And they all kind of start with, I mean, it's pretty much the same, 15 websites that feel illegal to know. So these cats that I have are like, is this a quotation? My leg, is this legal? Right.

Jim Fuhs [00:09:26]:
Right.

Dustin Stout [00:09:27]:
Is this illegal to do? So here's here's the reality. It's actually not that complicated. So each of these companies that Magi integrates with has what's called an API, and that's just a technical term for one app being able to leverage something or use another app on your behalf. And so the way that these, contracts that we have with these other companies, OpenAI and Anthropic and Google, is we are paying Magi is paying for the usage. So we're not paying a flat monthly fee to all these companies. We are paying based on how much we use them. So every time a user in Magi sends a message to to GPT 4, we get charged for. I get a bill at the end of the month for all these AI companies based on how much my users have used it.

Dustin Stout [00:10:17]:
And, that works out great for me because, you know, I can kind of manage my costs. It works out great for the users because they get everything all in one place. And so the way that we sort of manage and and don't go broke, right, is we put limitations on the usage. That's that's really what our plans are based around. The the, you know, multiple different subscription tiers that we have are based on how much usage you get and how many team members or workspaces you get. And that's that's how we stay in business. That's how we make it a win win for everybody.

Jim Fuhs [00:10:51]:
Right. So so as I understand it, like, like, Jim and I were were talking a little bit about this this morning because, you know, with the show coming up, I started using it more and more and more. And then I was like, okay. I really do kinda, like, low key love this thing. And, and I I like, I'm starting to look at the usage rate, and I start seeing going up. And, I thought I had hit, I had hit the the level. And then when I when I went to go sort of what what's called top up Mhmm. You know, like, I went over my plan.

Jim Fuhs [00:11:26]:
It was like, all I gotta do is pay $5 and then it unlocks an additional 10,000 words or whatever, on that. And I started to look, you know, the more you use these things, the more you start to get comfortable and you the more you start to understand, okay, you just can't go in here and just go, like, you know, rough shot and start throwing stuff in and start throwing through. Like, you're gonna you're gonna fill up you know, you have to be a little more tactical about about how you use this kind of stuff. Is each one of these platforms, do you is is it the is it the full enchilada of these particular, platforms, or is it like a a light, quote, unquote, version of, of these of Leonardo, of, you know, all of these things?

Dustin Stout [00:12:15]:
It's a great question. So I I really wanted to touch on, you know, this idea of topping off first because it's something a lot of people don't understand. So some people are under the false impression that when they go use ChatJPT or Claude, you know, either the 2 that they can just use it unlimited. They could just use it all day, all day, and they have no limitations. Well, the fact of the matter is they do have limitations. They're just not as upfront about it. So with chat gbt, I believe it's you. If you're a plus subscriber, you get 50 messages every 4 hours.

Dustin Stout [00:12:48]:
So that means if you're really going to town and you spend an hour and you burn through all 50 messages and you're only halfway done with whatever work you're trying to get done, you hit your limit and chat JBT chat JBT will tell you, sorry. You gotta wait another 3 hours before you can use this again.

Jim Fuhs [00:13:05]:
Ugh. Brutal.

Dustin Stout [00:13:07]:
And you have no upgrade path. There's no way to increase that limit. You simply have to wait. And and that's a terrible position to be in if you're up against a deadline or if you really need to get something done. And let's say it's like 3:20 PM, well, like it is here on the the west coast for me. And then you're telling me you have to wait 3 hours. Like, I'm not coming back to work at 6 o'clock. Like, that's just not gonna happen.

Dustin Stout [00:13:29]:
So it's gonna have to wait till tomorrow. Right? Mhmm. So with Magi, I wanted to make sure that users had options to upgrade. If they're in the middle of a piece of work and they hit that limit, $5 is all it takes to keep working. And and for most people, for most business owners, it's like $5 Yeah, of course. This this copy that I'm writing is gonna make me a $100,000. So we'll pay the $5 to top up, but also, you know, having those extra plans. Your your second question, you know, are the models that we provide in Magi, are they a light version? Are are they they the same as do you get the answers? They are the exact same ones you would get in these native apps.

Dustin Stout [00:14:10]:
And that's something that I'm very passionate about. And why I've built Magi in such a, the the the word that gets thrown around on it's it's very vanilla, in that it's not it's it's straight from the pipe. So the GPT 4 that you get inside Magi is the same exact GPT 4 that comes out in chat GPT. The Leonardo style options and models that you can get are the same exact ones that you would get if you were in the native app. Claude, Gem, all of them are exactly the same. It's a straight pipe, literally piping Magi, a straight line from them from, OpenAI or Clutter would have me straight into Magi for you. There's nothing and that's what a lot of AI apps won't do. They they want to sort of obfuscate what models they're actually using for you so that they can cut costs and use the cheaper ones and hope you don't notice.

Dustin Stout [00:15:06]:
With Magi, we give you full control. You use whatever model you wanna use, and we we are not gonna, you know, cut any corners anywhere. It's it's completely up to you to which model you use.

Jim Fuhs [00:15:20]:
Wow.

Dustin Stout [00:15:20]:
I just don't

Jim Fuhs [00:15:21]:
know why more people don't do this. I mean, I I it's just

Dustin Stout [00:15:23]:
I mean, it's it's it's

Jim Fuhs [00:15:25]:
it's incredibly brilliant. No. I I it's like, for me, I think there is there's incredible value, and I I know I know this isn't a technical, term. Aggregation is probably the the technical term, but it's curation. Right? So, like,

Dustin Stout [00:15:40]:
we all have

Jim Fuhs [00:15:41]:
access to Yeah. Like, I I equate everything to music. We have all we have all this access. We have access to a 150,000,000 different songs. Right? But we what do we do? We go to a playlist that was created for us by someone who's an expert. Right? Mhmm. So we still subscribe to this particular platform, but there is there there is curation, there is there's an expert that's that's there that's kinda says, hey, I have this for you. But, you know, your exam what what you're doing is you're saying, I own all of the I have all of the music services right here, and I can just pull them all there together for you, and you only have to pay one price.

Jim Fuhs [00:16:16]:
It's just, to me, I I, I think there's incredible value for someone, especially us and our audience who are very curious, solopreneurs, entrepreneurs, small business owners. They're like, listen, I I already pay for CHAT gbt. I pay for this. I pay for this. I pay for this. Why why do I why should I pay for this? And for me, I just think like this to me is not just a cost savings, but, the ability for us to be able to to take something like this and then organize ourselves. Right? So the the other thing I wanted to to kinda maybe pivot into, Dustin, was what are the what are Magi has these other things where you can organize your chats, organize, like, your your like, because you have all of these things going on, all of these multiple services that you have going on, you have this ability to keep everything in house and organize them and pin some things and and do some other things. Talk to us about the the advantages there of of being able to organize yourself within Magi.

Dustin Stout [00:17:26]:
Yeah. I mean, I'm an organization freak. I'm that guy who look my my you look at you open up my, Mac finder and you find very meticulous folder systems, a folder with Unifold that's well labeled and they're all organized and they have a system to it. Right? So I like to stay organized. And so when it came to chats, I was I was having dozens of conversations a day, you know, do dozens of different chats, accomplishing different tasks. Right? And they started to pile up really quickly. And the minute that I brought my team in, I wanted to make sure that I can direct them to where they need to go without them having to do a lot of finding and searching. Right? So we gave the ability to put chats into folders and organize them that way.

Dustin Stout [00:18:11]:
And, furthermore, we actually have what's called workspaces. So if you have a, side project and then your day job and then your personal life, you can actually split those up into different workspaces and keep them all separate. Right? And because we have team collaboration, you can invite team members to certain workspaces and leave them off of other workspaces. So making it super easy to keep things organized so you can find it when you need it. We also have very powerful search and filtering, so you can search through all your content and find it quickly. But also, you know, the organizational standpoint, your team can find it, your team could add stuff to folders, and you can kind of collaborate, Your team can jump into a chat right where you left off and continue going. You can go back and forth. So a lot of things with organization in mind and with collaboration in mind and, a small hint of my, OCD, needing to stay organized.

Jim Fuhs [00:19:13]:
Well, I but I think

Jim Fuhs [00:19:14]:
I think that was important because I remember and that's how long ago it's been that I used chat GPT, you know, when I was paying for it was just like, it was really hard to stay organized. Right? Because it's kinda like, alright, where was that message I was working on?

Dustin Stout [00:19:27]:
And so right away years. You couldn't even search through your chats. It took them 2 years to add that feature. We have it from day 1.

Jim Fuhs [00:19:36]:
Yeah. So, I mean, I love that, alright, this folder has got this project going, and it's saving it, so it's not like I have to start over if I'm wanting to, you know, do iterations on something to get a different result. And I just think that was brilliant. I I love the the copy and paste feature, makes it really easy to turn it into a regular document and then, you know, basically edit it in there and you can share it, you can, you know, put it in, you know, Google Docs or Word or PDFs, whatever you need to do. I mean, even the fact that you've got programs on there that will let you read documents and URLs, and I and I know that's one of the things you've you've talked about, Dustin. I mean, if somebody's just starting out, like, I guess with these large language models, where would you tell them they should start? Because I think it's almost overwhelming. Right? Because, you know, Google's got its thing. Facebook even.

Jim Fuhs [00:20:32]:
Right? You can't even go on Facebook without it wanting you to use AI to search for stuff.

Dustin Stout [00:20:38]:
Yeah. I I think the easiest place to start is just to start a new conversation, start a new chat, pick GPT 4. It's, you know, it's the easiest one to start with because it's the household name. Right? Just pick GPT 4 and ask it. Start a conversation saying, here's what I need to accomplish today. List out your tasks. You can you could even take a screenshot of your task list. Say, here's what I need to accomplish today.

Dustin Stout [00:21:04]:
Which of these things can you help me with this? And start there. I actually did it as an exercise, I think last week. I took a screenshot of my to do list and I was like, okay, which of these things can we work on right now? And it it told me, it did read the the screenshot. It knew every task in that screenshot and was able to say, okay, we can accomplish this, this, and this. Which one do you wanna start with? And I said, well, let's start with that one right there. And, it did. And the the the thing about this, again, going back to this idea of, like, we learn through doing, we learned how to Google by using Google. The more you use this, the more you realize you can use it for.

Dustin Stout [00:21:43]:
And I think that's, again, exploration and and just trying things out is the best way to learn.

Jim Fuhs [00:21:50]:
Yeah. I love how you've got the personas that are built into it as well. So, you know, I I I, you know, I created a a Chris Stone persona. Very scary for me. And, I I also love, like, if your intention is to get found, there is an actual SEO content writer that is a persona that's that's, that's built into into into this as well. So it's, you know, for me it was a it was a little, it was a little crazy to, you know, start with a blank page. Right? Like Mhmm. Start feeding the the information, you know, act as a research writer, here's all of my information, here's here's this, that, and the other, you know, like you said, start with with g t p, GPT 4.

Jim Fuhs [00:22:36]:
And then it's got this these selected personas that are here as well, educational content, create academic researcher, assistant, bible scholar, brand voice generation, business adviser. There's there's all kinds of different things. And then when you select these personas, they're giving you, different, sort of responses based on, that particular, persona. You know, when you get those things, I think the temptation is to go, wow, that took me 30 seconds. And holy crap, I can just take that, and now I have a blog post. Or I can take that and now that's gonna be my Instagram, copy for this photo that I that I gave it. Or this is gonna be that, and I'm just gonna copy paste. Let's talk a little bit about the human loop in this.

Jim Fuhs [00:23:30]:
And what is what is give me your thoughts on, you know, obviously, you know, AI is is drinking water out of a fire hose right now. There's all kinds of people using it. There's there's things are swirling around. Everybody has this fear of missing out on all of these things. And so they wanna they wanna just keep, you know, doing all of the things. What is your take, Dustin, on, the human involvement in, a lot of this generative space? And I'm not just talking about the text based stuff. I'm also talking about the image and video.

Dustin Stout [00:24:05]:
Yeah. So I think a lot of the models that are leading the way are extremely good at sounding human like and responding in highly intelligent, thoughtful ways. But if you are if you have any degree of domain expertise, if you are a expert copywriter and you have GPT 4 write you some marketing copy, you'll find out really quickly that it's not quite as good as you. And and and this, you know, from being a marketer for the last 15 years, I noticed that a lot. And so I think it's still at a point where you need a human in a loop. And you need to just make sure that you're checking everything it says because we also know that AI tends to lie. It can constantly lie to you like nobody else. It will make some talk.

Dustin Stout [00:24:50]:
So, you know, always wanna you always wanna fact check the things that it says because it does have that tendency. So, you know, it's up to you. Your reputation is on the line. So in my marketing agency days, when somebody would hire me as a ghostwriter and I would write these 2,000 word articles for them, their reputation was on the line. And so they didn't like, let me just write it and publish it. They had an editor or somebody who had to approve it. Right. They had to work look through that entire blog post, make sure it was in their brand voice, it sounded like them.

Dustin Stout [00:25:26]:
Right? And it represented all their values and all of the objectives, all the KPIs that they wanted, and then they published it. So I think in the same way, if we look at AI, these assistants or these bots, whatever you wanna call them, if we look at them as an intern, you've just hired the world's most creative, smartest intern in the world, fastest intern in the world, who never gets tired. Even if that was a human intern, you would want to audit or or approve their content before it goes out. Right? You would view everything that they do for you with a bit of a critical eye. And I think that's what we need to keep in mind. Like, we should look at everything it produces critically as if it were produced by a, you know, somebody just out of college who is looking for their first gig.

Jim Fuhs [00:26:19]:
Yeah. Love it. And, you know, because I think the temptation is like, oh my look how look how much time this saved me. And if you do go back and fact check, you still are saving the time. You still are hitting that creative thing. It's it's like, don't don't just take it and rubber stamp the content and and and just slide it out because your credibility is at risk when you do that. So so so take this stuff. I love what you said.

Jim Fuhs [00:26:47]:
It it just flat out lies.

Dustin Stout [00:26:50]:
You know, it's like Honestly, like

Jim Fuhs [00:26:52]:
I I, you know, I also I I look at especially the text based stuff, I look at it as this This was already created. So like when something gets spit out to me and I look at it, I'm like, oh, this looks really good. This looks almost like, you know, like, like something I would have seen before. Well, that's because it probably has been put out there by somebody else. Right? Right. So even if it's just, like, read through it, nah, I would not have said Delph. Nah. I probably wouldn't have put a rocket emoji in there.

Jim Fuhs [00:27:25]:
Yeah. I probably wouldn't have done this. I, you know, let me I I probably would compare, Dustin to, like, the AI Jesus or or, you know, you know, I don't know. Whatever. Like, add add yourself into it. It's okay that, like, even most of it is kind of already a framework for you. We do it already in in, like, you know, utilizing things like Canva. Right? Or a lot of the Adobe stuff.

Jim Fuhs [00:27:52]:
It's like, look, here's a template, here's a thumbnail. But, like, if you just take the thumbnail and you just change the words on it, it's somebody else's stuff, and it's not going to resonate, with people because it's been used before. And the algorithms on these platforms are probably not going to tap your stuff to get it seen more because it's already been used before as well. Right? Right.

Dustin Stout [00:28:17]:
Yeah. And here's one thing I'll say to that end is there are some some advanced techniques you can use to really get it to output content that sounds exactly like you would've written it. But it's not easy. Right? It takes and, you know, it takes you you watch someone do it once and you can probably recreate it. I have one persona that I created recently with the help of Jeff Hunter. And if you haven't heard of Jeff Hunter, Google him. He's a phenomenal copywriter, really smart dude, brilliant marketer. With his help, I crafted a viral post persona, that all I have to do is really describe what kind of social media posts I wanna write about, give it the topic, give it some experience from my personal life.

Dustin Stout [00:28:59]:
And in a single button, it produces it. And every single time, I I almost don't have to edit anything. It might be a word or 2. It might be a sentence that I remove, but, you know, that'll save me. And even if it's just a short post, you know, social media post might take me 5 minutes to write. Right? But with this, it took me 30 seconds, and then I had to edit out one sentence. So, you know, may not be huge time savings, but any time savings can help. And, you know, there are instances where if you create a persona that is, you know, really well tuned and and really, well crafted, it can spit out things like a blog post that would need very minimal, editing.

Dustin Stout [00:29:42]:
For someone like me, writing a blog post is typically an 8 to 12 hour process. But with the help of AI, because I researched the heck out of it. I know way too much about SEO. But now that I have these personas, I can kind of step by step process through a blog post with the help of AI and get it done in under an hour.

Jim Fuhs [00:30:03]:
It does cost Yeah. And I think another thing too that I think this is a difference with what Magi offers that maybe others don't think about is, 1, we're talking about. Right? The custom personas that you can save, but also even the prompts. Right? We can it's like prompt engineering is, you know, becoming a big thing where people like, well, what's the prompt I use? And we're able to save that as well. Is that correct?

Dustin Stout [00:30:28]:
Yeah. You can save your prompts inside of Magi. And then once you're in a chat, you can open up your little sidebar and have access to all the prompts you've saved. You can search through them, find them, and in a single click, add it to the, the input box and send it off.

Jim Fuhs [00:30:43]:
What I found when I was creating some stuff in Magi, this is turning out to be quite the Magi commercial for you. I I feel

Dustin Stout [00:30:50]:
guilty, and I haven't paid to be here. No. But I no. This is

Jim Fuhs [00:30:56]:
serious. Is is there was this little thing and I just I looked next to my prompt and it had a little magic little thing and and I hovered over it and it said enhance your prompt. And I clicked on it. And it and it, like, turned my prompt into some, like, as if an English lord had said, now you will now it will create something even more magically delicious for your for your fan. And see, that's really, like, the heart

Dustin Stout [00:31:23]:
of how I create products. That that one feature alone really shows you what I think about. And I I've had a lot of developers, a lot of really smart people start using Magi and they're like, well, why don't you do this? And why don't you do that? And the reason why I don't I don't add a bunch of crazy technical features that are great for like engineers and developers is because I build products for the person who isn't a developer, who isn't a prompt engineer, and who doesn't have time to go learn all these highly advanced techniques. I wanna build a product that, Joe, somebody off the street who's never used AI in his life can come in and start using it right away. And that one prompt enhanced button, it's built for exactly that, for the image generator. And it allows you to put in a very simple prompt and hit the button, and the AI will use all of its trained knowledge on what a good image prompt is and add tons of beautiful detail for you. And then you can then, you know, modify it as needed. That's really the heart of Magi, to democratize AI if I can steal Canva's killer marketing campaign from Magi.

Jim Fuhs [00:32:32]:
She was thinking about me, Chris, because I I wouldn't have been able to do all that stuff.

Jim Fuhs [00:32:36]:
Can you can you enhance your prompt to have it, Ed, like eating crayons like a marine? Is that no. Is that, like, part of the enhancement process? Dustin, what are you talking about?

Jim Fuhs [00:32:47]:
Well, you know what? Now I might have to go into the image thing and try to create something with that in mind, Chris. That's that's

Dustin Stout [00:32:52]:
eating crayons.

Jim Fuhs [00:32:53]:
No. I think that's that's like the most amazing part is it's not just it it's like so much of it is about prompting. Right? And it's not it's like, give me an image that's cool about podcasters that has this vibe, and here's my colors. Right? It's like the sort of, like, stock or whatever. And that might not be the right way to go about it. Like, certainly, the colors, you know, and whatnot. But, like, understanding how to prompt these models to to understanding how to, you know and each one of them is different. Right? Isn't isn't am I to understand that proper? Like, runway is different than Leonardo is Yeah.

Jim Fuhs [00:33:32]:
And chat gpt is different from than than all of these other things. And so understanding how to how you know, what it likes, what it does to actually animate the things you want to animate as opposed to it it it not doing that is is key to this whole thing is is is Yeah. Is prompting. And so that feature alone for me is, like, a huge time saver.

Dustin Stout [00:33:56]:
Mhmm. Yeah. Again, going back to this idea of exploration, like, you would never know which model is better than the other or which strengths and weaknesses a given model has unless you tried it or unless you read the blog post that I wrote about. Right? But so but again, experimentation is key. So having these different AI models, and it's most apparent with the image models. Right? With the chat models, it's it's a little more difficult to see which one's better at x y z and which one's better at a b c. But with the image models, it's very clear. You ask it for something photo realistic realistic.

Dustin Stout [00:34:29]:
Dolly doesn't do a good job. Another one doesn't do a good job, but then you use Flux and you're like, oh my gosh, this looks real. And then you tell Flux to do something that's, impressionistic or or, abstract. And you're like, I don't know. And then you give it to DALL E and you're like, oh, wow, DALL E nailed it. So you get to figure out what these different models are really good at so you can use the right tool for the right job.

Jim Fuhs [00:34:54]:
Is there a is there a resource for things like that? Like, so so if I'm a Magi subscriber, I I know that you like, there's a there's a Facebook group. Like, are there places where I can say, okay, I want to work on persona modeling, or I want to work I wanna understand the differences between, like you talked about DALLE versus, you know, all of these others. You know, is are there are there resources that you, you make available to to Magi subscribers or maybe to the public?

Dustin Stout [00:35:27]:
Yeah. So we I try to continually update our blog with new research and and, trainings. It's a little hard, because we're still a small team and I'm wearing a lot of the hats still, but doing my best to create that educational content as best as I can. The other place is our help center. So we're going through a process of going meticulously through every feature in the app and creating new training videos on how to use them because the app has changed so much in the last, 6 to 12 months. We've added new stuff and it's a lot of the old stuff is obsolete. But on top of that, inside your dashboard, if you're a Magi user, you'll see that you now have access to something called the AI Academy. And the AI Academy is a hub of of great AI training content, not exclusively about Magi, but we partnered with the AI Academy.

Dustin Stout [00:36:16]:
The creator of that is his name is Steve Dotto. And Steve invited me in to partner with him. So all Magi users get free access to his community, his, his hub of training resources. And every single week on Wednesday, we go live with a new training, topic and we'll dive real deep into whatever the topic of that day is and allow the people who are there live, get to ask questions and will answer those questions in real time. And so that that library of content is just built up and it continues to build up every single week.

Jim Fuhs [00:36:48]:
Does that also Dustin, does that apply to your team members? If you're, like, the owner of the account, do they get access to that as well?

Dustin Stout [00:36:55]:
Yeah. Okay. Yeah. If you were a active Magi user, you get free access to the academy.

Jim Fuhs [00:37:01]:
So I wanted to talk a little bit about this, and and we talked earlier about this is because, you know, a lot of this AI stuff is so fun. It honestly, it's it's been a it it's a blast to be able to create your alternate persona, to animate my thumbnails. Like, there's a lot of stuff.

Dustin Stout [00:37:19]:
Mhmm.

Jim Fuhs [00:37:20]:
And I I start to see this stuff, and then the more I the more I get into it, the more I'm I'm starting to discern what are practical uses versus what is just am I am I just playing here? Right? And so you, Dustin, have this, you know, overarching visibility to everybody that's using platforms, not only just through Magi, but probably in general, you're seeing what what everybody's, what everybody's doing. A lot of our listeners, a lot of our viewers are podcasters, livestreamers, solopreneurs, entrepreneurs, and they look at this kind of stuff and they're like, okay, this is cool, but am I just jumping into a playground here to do this stuff? Or are there really practical uses for my livestream, for my for my podcast? And so are there any maybe case studies, examples that you have of MAGI members or or examples that you can point to where they're utilizing some of these platforms more than just like, here's my transcript, spit out some show notes. You know, that that kind of stuff is is is cool. And, yes, of course, you know, we we use that. In fact, ladies and gentlemen, like, everything you see on this show right here was all created with Magi. I did that purposely. So if you're seeing any of the show notes, the thumbnails, all of that stuff, we did that because, like, we wanna show it to you. So if you're watching the show, you can kinda see, okay, these guys used, this.

Jim Fuhs [00:38:46]:
Look how we cobbled in a little other other stuff. But back to the question, Dustin, what are maybe some specific sort of practical ways that a podcaster or a live streamer can utilize, Magi properly? I think there's

Dustin Stout [00:39:02]:
there's so many ways in so little time.

Jim Fuhs [00:39:05]:
But I'll I'll try and keep it

Dustin Stout [00:39:07]:
specific to, you know, what I I think your audience is familiar with, you know, podcasting, you mentioned. We have one podcasting company. We have many podcasting companies, podcast production companies that use Magi regularly, and they do exactly what you said. They'll take the transcript from, a podcast recording and they'll turn it into show notes, but then they'll take that also and then turn it into a full on blog post based on that, which is as a marketer, I told I have told every podcaster to do in the last 12 years, turn those into actual blog posts for SEO. But on top of that, then turning that into social content that you can then drip out throughout the week or throughout the month. You know, so AI is really good at taking content and repurposing it in many different avenues. And so that is that's a huge use case. It can't be understated.

Dustin Stout [00:39:58]:
A lot of people use it for that. And bloggers, podcasters, YouTubers. So we have a, a lot of YouTubers use us because we have this feature where if you paste a YouTube video link into the chat, Magi will recognize that it's a YouTube link. And before it sends it off to the LLM, Magi's gonna go fetch the transcript from that YouTube video and attach it to your prompt so that the AI that you pass it off to knows everything that's said in that YouTube video. So then you can then, you know, easily start doing stuff with your YouTube videos.

Jim Fuhs [00:40:30]:
Wait. You just made every like, the we we here we are, 46 minutes into the show, Dustin. You just gave me, like, like, are you serious right now? So in a new chat, if I paste a YouTube link, it pulls in the transcript from the YouTube video. Automagically. Wow. Oh my god. Alright.

Dustin Stout [00:40:50]:
Yep. Yeah. And in the same way, if you post a web link to any public web page, the same thing happens. Magi recognizes this is a web page, probably want all the content from that page to be included in the context, and so it will attach it to your prompt. And this makes it super easy to say, Hey, give me feedback on this landing page. How can I make this landing page copy that? And all you have to do is paste a link. You don't have to go over to the web page, drag and copy along with text and paste it over and hope it works. No.

Dustin Stout [00:41:25]:
Can you just drop a link in your magic?

Jim Fuhs [00:41:26]:
Or for me, like, I'm pulling it into another program, like Descript, and I'm exporting the transcript and I'm attaching it to to chat gbt. And, like, you

Dustin Stout [00:41:35]:
know,

Jim Fuhs [00:41:36]:
you know, from the attachment, you know, utilize here's the tran you know, like, oh, jeez.

Dustin Stout [00:41:41]:
Okay. Yeah. That one feature is huge, the the YouTube one. Do you have a good and here here's another use case. So, you know, I'm I love following business practitioners and I love following the the latest advice from, you know, marketing strategists and stuff. And so, like, I'll watch a YouTube video where they outline this brilliant process or this brilliant campaign strategy. And I'll be like, oh my gosh, I wanna do that. And so instead of me having to let go take notes throughout the video, you know, because I'm terrible at taking notes anyways, I just grabbed a YouTube link, pop it into Magna and say, hey, write out a strategy based on this video, and and then list out all the steps I need to take and who I need to send it to on my team.

Jim Fuhs [00:42:25]:
Which model do you use specifically for that? Is there one that you would say does it better than others?

Dustin Stout [00:42:30]:
So for something like that, I use GPT 4. And the reason I use GPT 4 for that is because GPT 4 is the best at structured output. So like list and and styling things and and formatting, it's the best at that. But when it comes to copywriting, I'm a big fan of Claude 3.5 Sonnet. So if it were I were actually writing marketing copy, I would switch over from GPT 4 and say, okay, let's start writing the the copy for this. Wow.

Peg Fitzpatrick [00:43:01]:
See, Chris is taking notes over there.

Jim Fuhs [00:43:03]:
Oh, yeah. I'm trying to take I don't know why I'm taking notes.

Chris Stone [00:43:05]:
All I gotta do is feed it into Magi, and

Jim Fuhs [00:43:07]:
it's gonna tell me what that is.

Dustin Stout [00:43:08]:
The Steve will not be blank because that's that.

Jim Fuhs [00:43:14]:
Oh, man. That's, that's killer. So let's let's talk let's let's pivot a little bit off of this because I'm I'm already like, I can't wait for the show to end so I can go and start, start utilizing no. I'm, start utilizing this stuff. So you're someone that obviously has to use these things. Like, you're not and you so so that you can recommend things. Like, I've seen videos where you're like, hey. Because some of them you have to attenuate, oh, 1 through 5, like, how how it should how it should model.

Jim Fuhs [00:43:48]:
And you'll say, I found the sweet spot as 2, a scale of 1 to 5 because if you give it too much, it's going to be, you know, too built in, and it's not gonna give you the animation that you want. And, like Yeah. So you're you're really going in, and you're you're you're testing these things all of the time. But I know you're you're also a family man. You also, you know, have a lot going on, and and you've you prioritize that over what you're doing. So I I appreciate that as as a human being because I know, you know, this is a this is a great program, but I also love to follow, people like yourself that that have their head on straight, you know, as far as, like, hey, I dropped my kids off, you know, I brought my kids back from school, so I'm cool, I can I can do this show, you know, and and, you know, prioritizing your life, your faith, everything you've got going on, I think is it's super important, that you're doing? But I I really wanted to tap into because you're you're an expert in a lot of these different things. I really wanted to kind of find out from you, like, where do you see this going in the future? Like, maybe some some, like, emerging trends that you're seeing that are actually not, you know, look at look at the we just talked about, you know, the the new launch of of Sora, like, all of these things that are that are happening. Where do you see maybe overarching, maybe specific get as get as specific or overarching as you want.

Jim Fuhs [00:45:19]:
Like, what do we what do we do in a year from now?

Dustin Stout [00:45:22]:
So I I think the, literally 2 weeks ago, I was I was sitting in my office. I think it was a Saturday morning. It might have been a Saturday morning. I try not to work on the weekends, but Saturday mornings are like when I feel I have like some free space to think because it's not the grind of the week. And, you know, like, usually we made a big breakfast for my kids. We do pancakes and hash brown. I'm like, you know, you go about to go to food coma from big directions. I think so.

Dustin Stout [00:45:50]:
I've got some space to think, and I don't feel like, you know, I need to get a bunch of tasks done. But I was I was thinking about this question, like, where is it going in the next year? We have all this talk about AGI, and I think, you know, between you, me and the folks watching and that, you know, the CIA listening in, I do not think there's a whole lot of reality to the the AGI talks happening right now. And, and if you listen to a lot of the talk that specifically Sam Altman and his idea of AGI, the target keeps moving. It keeps kind of lowering the bar. And I think that's for good reason because I think it's a it's a much further away than we realize. But what I think is very close in the near future is these AI models are finally good enough to start doing tasks autonomously. So right now they need a lot of human in the loop. You need a lot of checks and balances to see, you know, make sure they're doing their job and make sure you're not blatantly lying to you.

Dustin Stout [00:46:51]:
Right? But I think they're they're just on the brink with this new model called o one from OpenAI. They're thinking more, more thoughtfully about the things they do. And this kind of marks the the age of where we can kinda trust them with a little bit more autonomy. And this is why on that Saturday morning, I, you know, just kind of thinking forward, like, what do we need to do in this next season of Magi to bring our users the best of the best and, and really help them go into the future. And that that is agents. So right now people throw around the word AI agents kind of loosey goosey. And I have a bit of a pet peeve about it because I don't think they're talking about actual agents. To me, an AI agent is a AI model that can take action, based on its own decision making and and actually take action outside of the container that it's in.

Dustin Stout [00:47:48]:
So right now in Magi, our personas, they can't really do anything outside yet. But very soon, probably within the next 2 or 3 weeks, we're gonna start adding these skills or actions to the AI models where you can tie them to things like Zapier, or you can tie them to things like Gmail, or tie them to, you know, whenever your YouTube account, whatever app, we're gonna unlock this ability for them to make connections outside of the app and take control of whatever tasks that you give them the ability to take control of. And that will then lead into what I think the next phase of AI is, is autonomous, workflows. So giving these AI agents the ability to not just do a task and take action, but pass the next step in the process off to the next persona or the next agent. So that then we start to have these autonomous workflows where we can give, you know, a set of AI agents an entire standard operating procedure and just let it go do it. And that's where I think this whole I think 2025 is gonna be that. I don't think we're gonna get anywhere close to AGI, but the the level of intelligence where they can these AI agents can accomplish things in a sequence without human supervision every step of the way, I think that's what we're gonna see in 2025.

Jim Fuhs [00:49:09]:
Yeah. I think I I I do agree with you. I I think, it it would be interesting to see how much people trust. The that right. I think, you know, when when someone hears that, it's like, you know, it it wasn't that many years ago when we thought, like, I don't wanna put my credit card on the internet, You know? Now it's like, you know Okay. And we're we're we're giving it away, like, you know, we're walking through scanners that are, you know, pulling down our our our CCV files, you know, and, like, whatever. And and so I think, you know, so that those walls are breaking down, and then there's security issues and all of that. But I think age the the the the agents that you're talking about, it's like, well, if if if it works, and and you're like, okay, you know, I've I've created this thing that that I that I or that I don't have to do myself, and the agents can do it for me.

Jim Fuhs [00:50:05]:
Yes. Is it replacing jobs? Absolutely. I mean, there's just no doubt about it. Like, when that when that happens, like, you know, all of a sudden, you're like, I I wanna book a flight. You know? It it's a it's a prompt, and all of a sudden, it's done. And you have a flight, and there there wasn't a human involved in it. Yeah. And that's crazy.

Dustin Stout [00:50:26]:
Yeah. But It is. Yeah. I and you mentioned, you know, the loss of Josh Singh, which I think is is totally legitimate. You know, there are a lot of people This fear of like, am I gonna get replaced by AI? And, you know, my answer to that is always not yet. You AI is not gonna replace you yet. Somebody using AI is gonna replace you or somebody using Magi is, to be honest, is gonna replace you. There we go.

Jim Fuhs [00:50:49]:
There we go.

Dustin Stout [00:50:51]:
But I think on top of that, when calculators were invented, we didn't lose the mathematicians. Right? We didn't stop learning math. When calculators invent were invented, it freed us up to solve more complicated problem. And I think that's what AI is doing for us. It's going to allow us to offload some of the the stuff that is just not necessary for us to spend so much time on and free us up to do more important work, more difficult, more complex problems that haven't been solved yet. So that's why I'm really excited about the future of AI.

Jim Fuhs [00:51:27]:
And and I also think too, you know, one of the things you know, because people always like, oh, my job's gonna get replaced by this. But I think it kinda goes back to, but what are you doing to improve yourself? Because there's so many other things that AI can't do. I mean, like, we're especially right now. Right? AI is not gonna come fix my toilet. It's not gonna fix my electricity. You know? Right. Those were high paying jobs, but people like, oh, but that's hard work. Yes.

Jim Fuhs [00:51:56]:
And I think that that's part of the problem, you know, not to go too far off top, but I think that's the whole thing is, like, okay, if AI is coming, how do you embrace it to make yourself more, you know, adaptable, flexible? Because you can use it for a lot of good I mean, I love the fact that, you know, like we're talking about, if I am going to give a class or something, I say, okay, give me a outline for this topic and, like, oh, yeah, I forgot about that. I should add that into my outline. It doesn't mean it's gonna write the whole, you know, word for word what I'm gonna say, but it gives me those things that, you know, our brain's not as sharp as a computer in theory. So I I think that's where people should look at it more that way as opposed to, oh, it's gonna replace my job, sort of a thing.

Dustin Stout [00:52:42]:
Absolutely. Supplemental is is the word. It's a supplement to your current skill set.

Jim Fuhs [00:52:50]:
Ladies and gentlemen, you have got to check out magi.co. If not, because Dustin Stout is just awesome as a human. But no, seriously. I mean, I don't know honestly how, creator doesn't have this if I'm being completely honest. Because I've tried I I went through this whole thing and I'm like, cancel the subscription, cancel the subscription. I, you know, it's all of this stuff is in there. They're adding constant, programs to this. So you can go in, and there's tons of ones that I didn't even know it even existed.

Jim Fuhs [00:53:29]:
And for me, and I I shared this earlier, as well. You know, I started playing around, with some of the stuff just to, you know and but I'm using this for client work. So I'm doing an award show with 1 of the podcast clients that I have, and and I created an award. I animated the award. I'm adding it to, their the video, for for that. There's an example there. I started animating, some of the, thumbnails for some of the shows. Dustin, I shared, I shared that with you.

Jim Fuhs [00:54:04]:
These are dolly images, that we created, some vertical images. You seem to be nodding or or just maybe a little bit nervous in this image here, Dustin, which is like, I don't know what's going on, right there. But, you know, and and so I'm learning as I'm going through this. It's like, okay, This doesn't quite look like me, so and it's probably because I was looking to the side. So I'm starting to learn I'm starting to learn how to manipulate. Okay. When I feed in an image that I wanna animate, I need the whole face. Right? And I'm starting to to to understand how all this stuff works.

Jim Fuhs [00:54:39]:
But the only way, folks, to understand how all this stuff works is to get in and and and start, you know, getting curious, start utilizing that stuff, and there's no better way than to, to do this with, magi.co. Jim, I think you have a a a link. It might be an affiliate link. It

Dustin Stout [00:54:59]:
it is.

Jim Fuhs [00:54:59]:
It is. And we're gonna we're gonna put that in the show notes. Dustin, we're gonna roll it over to you. How should people connect with you other than going to magi dot co? It is there a place where is it is it LinkedIn? Is it Facebook? Is it Instagram? Where do you like to hang out?

Dustin Stout [00:55:17]:
Well, as a former social media marketer, I'm on all the things. So you can find me on all the things. I'm at dustin w stout on all of them.

Jim Fuhs [00:55:27]:
Beautiful. At dustinw stout on all of the things. Ladies and gentlemen, please, go check out magi.co. And as always, don't fear the deer.

Jim Fuhs [00:55:40]:
Thanks for listening to Dealcasters. Congratulations. You've taken another step forward in your content creation journey. Please don't forget to hit the subscribe or follow button here in your favorite podcast player so you can be reminded every time we drop an episode.

Jim Fuhs [00:55:57]:
We love hearing from our listeners and viewers. And if you're wanting to watch our shows live on Amazon, feel free to follow Dealcasters Live as well at deal casters dot live. Follow us on Twitter or subscribe to our YouTube channel where we also

Jim Fuhs [00:56:22]:
also email us at dealcasters@dealcasters.live. Thanks again for listening, and you know the deal. Don't fear the gear.