Unleashing AI Power in Marketing: Matt Wolfe's Top AI Tools Revealed

I spent over a decade in the trenches of marketing attribution, tracking over $5 billion in ad spend and over $10 billion in revenue. And here's the truth. Most marketers are overwhelmed second guessing every decision, drowning in data because no one taught them to use attribution the right way. That's why I built the Five Forces, a proven system to give you clarity, confidence and direction to use attribution data with a system. Right now, perpetual traffic listeners can get $800 off the Five Forces course and system over at five forces.com just use the code PT. You don't need more tools, you need a system. Go to five forces.com and get the system. Use code PT to save $800 today. Hello and welcome to the Perpetual Traffic Podcast. This is the show where we share cutting edge strategies to help CMOs, marketing directors, business owners to acquire more leads and sales so they can achieve their vision. And today, Kassim, we have yet another visionary. A visionary 14 years in the making, an overnight success. 14 years in the making. None other than Matt Wolf. It's so funny that this guy is now as popular as he was because I remember him as just a struggling affiliate marketer like myself 10 years ago, like barely scraping by. And now he's like an overnight 14 years in the making AI YouTube sensation at the age of 40. So I think I actually secretly would like to be Matt Wolf in my next life. So it's pretty exciting to have him on today's show.

Here's the thing, this isn't a surprise to me at all, Ralph. I. I don't know, Matt, but if your name is Matt Wolf, you have to be famous.

You should be because it's one of those names, right?

Yeah. You can't not be famous with a name like Matt Wolf like this.

You're predestined, actually, I think for greatness just by having that name unto itself. So Matt is one of the best. Like, this is the place I go when something new happens in AI. Matt Wolf will leave links in the show notes. Obviously it's the place to go for anything that's new. One of the videos I was watching this morning, preparation for today's wrote like Saturday morning and it was already live. He's real time. If you want real time updates on all things AI, he's your guy. So it's where I go. Well, obviously we'll leave some links in the show notes here, but super excited. Matt Wolf runs FutureTools IO Qasim. And I think a lot of the folks that Listen to this show. They're like, man, there's so many freaking tools that are coming out for AI right now. Sam, I'm a director of marketing or a VP of marketing or I'm a CEO cmo. I don't where to go. Like we always sort of go back to the same tools here. But he's going to declutter that for us and get us down to like the most important stuff here today as well as talk about the future of AI. So pretty excited. 2460tools2462 how can you possibly keep up? But Matt keeps up with it. So all you really need to know is just listen to today's show and he'll solidify it for you. So Matt, welcome to Perpetual Traffic. Great to have you on.

Thanks for having me. Yeah, it's fun listening to the intro and I'm very grateful for you guys and you're amazing and I think you're giving me way too much credit. But I appreciate it and love hearing it.

Well, you know, we do set the expectation rather high here. So you know, we over promise and then under deliver on Perpetual Traffic every week. That's like, that's how we roll. So today's show is really is about what tools are the most important from your perspective. The ones that you use all the time, especially when it comes to digital marketing. Like if you do have a team running your market or you're doing it yourself, but primarily maybe from sort of a higher level. And then we'll also talk about the future of AI a bit which we've talked about here on the show quite as much. But love to give your take on that. Before we get into that, any nugget that you can bestow upon the professional traffic listener here today. Something that they can use almost immediately in their business. Perhaps an AI tool of some sort.

Yeah, sure. So there's a tool that I've been finding really cool lately that I think a lot of people will find very useful. It's a tool called HeyGen. I think it's@heygen.com but what's really cool about this tool is it's a tool where you can do two things with it. One, you can train in your own face and your own voice. And then you can type any sort of text and then it will literally just make a video of you saying that thing. So if you don't want to be on camera, you just want to have a video, a talking head video of yourself, but have it just automatically read off a Script Hey Gen will do that. But where I find hey Gen even more valuable is it's really, really good at translating into other languages. When it translates your videos into other languages, it will actually use your same voice, your same intonations. It'll put the emphasis on the same words and the AI will actually make it look like your mouth is moving to the new words that it's generating. So if you want to take a video that you've made or some content on a blog that you've made and translate it into Spanish or Italian or German or whatever you want, this tool will go and do that for you and make it look pretty seamless. So that's one of the tools that I've been nerding out over lately the most.

Have you thought about having a Spanish language YouTube channel?

Now you just take all your. I've thought about it.

Flip it over.

I've definitely thought about that. However, YouTube is actually rolling out a sort of multiple language feature where on the same channel you'll be able to switch languages and it will overdub your voice on the existing video. So it's kind of getting to a point where it doesn't make a lot of sense to go and create a new channel in a different language because it's going to be just integrated into your existing channel in the near future here. It's actually already rolled out on some channels. I'd say the only exception is someone like Mr. Beast is out there going and actually hiring voice actors to do it. So he's actually going and hiring famous voice actors from these other languages. And the fact that it is these famous voice actors is actually helping get more views on those videos. But I would say for 99.9% of the population, that's not necessary. Just we're going to be using YouTube's built in AI overdubbing pretty soon anyway.

I want to see little scrawny Mr. Beast with like, hola, como esta miyamost. You know what I mean? Like, because you know he's hiring like whoever does Schwarzenegger's voiceover.

Yeah, yeah, yeah. I mean, it's already circulated too, where somebody went and took a bunch of videos of Mr. Beast and then overdubbed a new version of Mr. Beast pitching crypto ads or supplements or things like that. He's already had to fight to shut some of that stuff.

I see that all the time. Like people just spoof Joe Rogan videos and have him promote stuff that he's never promoted.

Yeah. And I mean, that's A whole other rabbit hole that we didn't even talk about before Hit and record is all of this sort of scary stuff that AI presents. Because it's really hard to go down this AI rabbit hole and not be just a little bit scared of where we're going as well.

Even I'm looking at the HeyGen website and in the footer they have this icon for the content authenticity initiative. You'd have to kind of opt into something that says, hey, we realize that what we do. Opportunities abound here, but I think things like scams and digital bullying and right now, every couple of months, somebody will start texting or messaging my employees saying, you know, hey, Ralph, it's Kaasim. I can't talk right now, but are you available to help? And if they respond not paying attention to the email, then at some point they're gonna get asked for money. Imagine when it's a video of me.

Yeah, yeah, it's definitely scary. And a lot of the sort of verification things are still kind of a joke too. It's like, check this box to prove that it's really you. And yes, I certify this is really my voice. It's really still kind of a joke at the moment. But, yeah, some of that stuff's kind of scary.

I believe we used heyj. I'll leave a link in the show notes just to show how good this thing is. We actually took a perpetual traffic episode and created. Created a YouTube video with Kasum and I talking Spanish. And it's unbelievably great. I don't really know what we're saying, but it's totally convincing. Obviously, making the mouth articulate the words. And the one thing that we didn't do because we were sort of screwing around with it is because it made the cost of my voice like the same voice.

Right?

I believe it was. Hey, Jenning. I have to double check with my team, but I find, like, a lot of the video tools right now are still, with the exception of that one, as far as substituting in a different language, some of the ones like Synthesia IO are. They're not quite there yet. They're still a little creepy. Obviously, we'll get today's show, but have you used much in the way of those types of tools? Is there another one that actually literally creates the video for you with your AI induced voice? Is there anyone's there that you like more than others?

Yeah. So, hey, Gen, I think is definitely the best at it. It actually used to be called Movio. So if you were paying attention to AI a year ago. Movio became Heygen. But Synthesia is another decent one. But it's still a little uncanny Valley. You look at it and you're like, there's something off about this. There's another one. I'm actually drawing a blank on the name right now. That actually does it in bulk. For a while there was those emails going around where there'd be an image in the email and then on the whiteboard it would have like your name on the whiteboard. You're like, how do they do that? Well, there's actually bulk video tools that do that now, where you can upload a spreadsheet of like everybody on your email list and it will make a new video where it dubs in their name into the video and you could send them to a link and that link will actually have like, hey, hey, Frank. And it'll actually be like your name on the picture and the words coming out of the mouth. But I'm drawing a blank on what that's called. But I think, hey, Gen might actually even have that feature built in now as well. A sort of bulk video feature.

Yeah, no, that's awesome. That's awesome. Well, we're going to get into a lot of tools here, but try and boil it all down. I mean, we'll leave links in the show notes here, but there are so many tools. 2,456. I think your team just updated it since we actually were, you know, on the pre roll. But that is a lot of tools to sort of sift through. I mean, especially if you're trying to figure out, okay, where is this going to fit in into my marketing plan for 2024. Like, what are the ones that I need my team to be using just as a base level? Obviously Chat GPT, we use that Google Bard, a lot of those. But we're going to get into some of those tools here. You're the guy to obviously help with that here today. So super excited to get into that. And then also a little bit of the future of AI. We would be remiss, however, Qasem, if we didn't mention at least one more time our special, special offer that we have for one lucky perpetual traffic listener. Would you like to divulge what that is, Qasim? If this is the first time anyone's.

Been listening to the show, if anybody's actually. Oh, dude, yeah.

Oh, yeah, we have 1, 2, 3.

It's getting competitive, but still you have a better chance of winning this than literally Any other competition in the world. Clarion Events has given us a ticket to give away. Ralph, this is free. It's a $2,500 value. That's 2,500 and change, actually. And if you go online and write us an honest review, we will choose from the reviews that have been written at random. One lucky person to come to Traffic and Conversion with me and Ralph. Well, Ralph Reburns will buy you any drink of your choice.

Might be one of those really expensive Las Vegas. Like 40 or $50.

Comes in the pineapple.

Could be anything. Yeah. Comes in the pineapple itself. Yeah, that would be fantastic. And yeah, so that's what we're going to do. And it's all legit. You're going to be presented it by Clarion themselves. Apparently they're going to appear. But yeah, this is a $2,500 value to what we believe is the best digital marketing conference on planet. Traffic and Conversion Summit, first week in January. What are the dates again?

9Th through 11th.

9Th through 11th. So we'll both be speaking there. Certainly not a reason to go because we're speaking there. But the point is, is you can get a free ticket. So leave an honest review for us. Wherever you listen to podcasts, if you're also looking for a way to get your team there or multiple people, we are offering a 20% discount on your tickets to Traffic and Conversion Summit. We'll leave a link in the show notes, but get on over to trafficandconversionsummit.com and enter the promo code, Perpetual traffic. You will get 20% off, so you save money there. Might even pay for your airfare, depending on where you're coming from. So, yeah, absolutely. Vegas. It's Vegas, baby. So without further ado, we're going to get into this.

You're saying the discount will pay for the airfare?

I think so.

That's what I'm thinking. I was like, I am opting out of this offer at this point.

I'll buy a discount. We'll pay for your airfare. Just. And if you confuse that, then Kasim will pay for your airfare. How's that? Yeah. So we'll lock you in. All right, let's get into today's show with Matt Wolf on all things Tools. AI super excited to have him on the show here today. So stick around right after this quick break. You're listening to Perpetual Traffic. Hey, it's Ralph here. Let me tell you about a lifestyle brand that we recently worked with, where they grew their revenue by 49.8% year over year and hit eight figures in top line revenue for the very first time in their history. And they are now on track for 25 million in revenue in 2025. We are so excited to be working with this company. And the reason why is they started using Tier 11 data suite about a year ago. It re reduced their unattributed traffic by 90%. That's right. They're unattributed direct unknown traffic that probably frustrates the hell out of you. Over in GA4 or one of those other crappy attribution tools, we reduced that unattributed traffic by 90%, uncovered $850,000 in hidden revenue and scaled their ad spend by over 3x. These results are not magic. The results of clean, accurate data and a system designed for today's privacy world where everybody is trying to block your data. If you want to see these kinds of results for your business, reach out and let's make it happen over@tier11.com apply and discover how Tier 11 data suite can finally allow you to scale your business, acquiring new customers at a cost you can afford. Head on over to tier11.com forward/apply. All right, so we are back with Matt Wolf. I can't believe you're a YouTuber, dude. Like, this is like so funny. We've known each other for like 10 years or so and I think the last time we were both commiserating about the affiliate marketing lifestyle, like how we're going to actually make it. And then things started taking up. You did the Hustle and Flow podcast which I was on. And then you started doing a lot of different things and all of a sudden you just, you blew up. I'm like, oh my God. I know that. I know Matt Wolf. It's crazy. So your channel now we'll leave links in the show Notes, obviously, is Matt Wolfe YouTube channel approaching a half a million subscribers. It's the go to place for all things AI. Maybe just tell us a little bit of the backstory of like how you stumbled into this whole thing by accident over, you know, your career in digital marketing.

Yeah, so I actually started the YouTube channel almost 15 years ago. My first video on the channel was 2009. So next year it'll be 15 years that I started the YouTube channel. And I really just started it to make random tutorials about WordPress plugins and things like that. Fifteen years ago I was really, really into WordPress. I had a course that taught WordPress. So a lot of my videos were here's a plugin you can use. And here's my favorite theme and why you should use it. And if you look at the sort of progression, it was kind of the same type of video. I mean, I wasn't nearly as confident on camera. I didn't have the editing skills or the cameras or that kind of stuff, but it was the same kind of stuff. It was just nerding out around cool tools. But back then it was WordPress plugins. I did this podcast, Hustle and Flowchart with Joe Fear. He's actually still running that podcast. And back in around 2021, we started talking about AI here and there on the podcast because by then GPT3 was already out. It was very, very under the radar. Nobody knew about it, nobody was talking about it. There was a tool called Conversion AI, which then became Jarvis, and then Disney said, nope, you can't call it that. And then it became Jasper. So we were talking about those kinds of tools that were using GPT3 under the hood and going down those rabbit holes. But back then, nobody seemed to care. We thought it was a really interesting topic, but those were some of our most underperforming episodes. Nobody cared. I never really lost the interest in AI. I was still playing with it. I made a few videos on my YouTube channel here and there about it. And then at the end of 2022 was when everything really started to bubble up in the AI world. ChatGPT came out in November of 2022. Mid Journey, I think, came out in April ish of 2022. Also towards the end of 2022, stable diffusion came out, which is like Mid Journey, where you can make your own AI images, but you can also train the models so you can actually put your own face into them and train them on your own styles. So I started making videos about that because I just thought this stuff's really, really cool and nobody's talking about it. So let's. I'm going to make some tutorials and show people that this is possible. And those were the videos that really started to pick up some momentum. The first video I did about stable diffusion, I think did 40,000 views, which my channel average was probably a hundred views per video. So 40,000 was quite a leap. And then in November when ChatGPT came out, I put out a video about that. That one did 1.2 million views. I went, all right, there's something to this AI thing. People are starting to actually gravitate towards my channel for it. And I just went all in on AI. So starting in about November of 2022, I actually had maybe a 45, 50 day period where I went, I'm going to go all in and I'm going to go hard. And I made a video a day about a different AI tool, like every single day. And around that same time was when I built Future Tools. And Future Tools was literally built so that I can just keep my own head wrapped around the tools that I've talked about and the tools that are available. And as the YouTube channel started to pick up Steam, I just needed to have some place to remember what I've talked about. And that was how Future Tools actually started. And both Future tools and the YouTube channel just sort of started growing really, really fast around the same time. It was around November, December of 2022 when they both started to pick up Steam. It wasn't until February of 2023 where I had the idea of maybe I should be collecting these people onto an email list. I started an email list in February. The email list was growing at 2,000 subscribers a day from the day I launched the thing. Today we're at 150,000 subscribers with zero paid ads. I've never paid for an ad on it still to this day. And that's kind of the short, condensed version. There's a lot of stuff in the mix that I didn't bring up. Like, I experimented with affiliate marketing. I did that for a long, long time. We had sponsors on Hustle and Flowchart. That was our main revenue driver for a while. Joe and I had an agency where we helped people with SEO and some content marketing and a little bit of advertising and click funnels, landing pages. And when it comes to the digital marketing world, we sort of dabbled in it all, throwing stuff against the wall to see what was going to stick. We were scraping by, we were paying the bills, but not a whole lot more throughout that whole time period. And it wasn't until the YouTube channel took off that I went, okay, I finally found my thing. And so for the last year and a half or so, I've just been all in on this.

So sick. I mean, I think that's, I don't know, as if that's the 14, 15 year journey is the takeaway from it. But it's just putting your head down and putting in the work and then noticing the trends. And obviously, I mean, the content on today's shows that you release at least two or three videos a week, some in real time. When stuff is really breaking news, you just sort of pivoted slightly from WordPress plugins, tools to AI, more tools, and noticed the trends and then just doubled down on it. And I think that's for anybody who's trying to get any sort of media going, whether it's a podcast, whether it's a YouTube channel, whether it's a blog. It's just figuring that out, reading the data, but also putting in the work, because eventually you will figure it out. In this particular case, you rode this wave, which you clearly saw before other people, and then all of a sudden, boom, the thing explodes. And now here you are today and the YouTube channel is like, the sky's the limit as far as all this stuff goes.

Oh yeah, it's been a crazy ride, but for the last 15 years, a lot of people just kind of knew me as the tools guy. When you needed a tool to do something, whether it was marketing related or traffic related or productivity tools, things like that, I was kind of always the tool guy. Listened to Hustle and Flowchart, which I did with Jo starting in 2007. Joe was the mindset guy. He talked a lot of philosophy and mindset and that sort of stuff. I was just all about the tools and the tactics. He would listen to me and I would be like, oh, you got to go download this tool and this tool and you got to play with this and you got to install this. And here's a traffic tactic that nobody else is talking about. I was all tools and tactics. And I feel like I've never really changed what I'm all about. It's just the technology changed, you know?

Right, yeah. And FutureTools IO. Well, obviously we'll leave links in the show notes here. I mean, it's so well organized and I mean, if you do want to keep on top of what's going on, I mean, it's funny, like sometimes it's the stuff that you actually formulate, like this is what I need to help me be productive. And like, if it's helping you, then that means, oh, there is actually a market out there for it. It's like that just sort of evolved and it makes sense because now we're looking at so many different tools which are the ones that are most important for me. And this one of those sites that you've got to bookmark. If you're listening to the show and you're doing anything with AI, this is a bookmark site. Forget about all the other ones. At least here you have the directory of where to go. And that's super, super helpful.

Yeah, for sure. I mean, this Actually started as a Google Sheet spreadsheet right as I was making videos. Whenever I talk about a tool, I threw it in a spreadsheet. It got to the point where the spreadsheet had 160 columns in it. And I was like, okay, I can't keep up with it this way. I learned about webflow, which is a no code website builder that has these really cool sorting functions in it. And I went, okay, I'm just going to toss this all into webflow and that'll make it real easy to sort. I did that. Never shared it with anybody for several weeks after making it once I felt like, oh, I've actually had some fun with this. Built a cool design around it. I actually went on Twitter. I had no following on Twitter at the time. I had maybe, I don't know, a hundred, two hundred followers, something like that on Twitter. I shared it on Twitter and said, hey, I made this website just kind of messing around in a weekend. It's all the cool AI tools that I came across. And then Robert Scoble, who's actually a large name in the tech world and AI and that sort of thing, he thought it was really cool. He retweeted it a bunch. And my post sharing my future tools website actually went viral on Twitter. I went, oh, people are actually liking this tool. I actually then went on Reddit and there was a subreddit called I built this or something like that. And it was for people that build weekend websites and weekend projects. I went and tossed this link on the subreddit of I built this in a weekend or whatever it was called. And it actually blew up on Reddit as well. And at least in that subreddit was the number one post for like a month. So it got really popular in that subreddit. And then I went, oh, maybe I could put this on Product Hunt. So I went and put it on Product Hunt. It actually was like the number two product of the day. The day I launched it on Product Hunt. All of these little things just stacked up and stacked up and stacked up. And the next thing I knew, the thing that I built, that this is helpful to me, was when I first launched it, it was getting like 200,000 visitors a month or something like that. Today. I think it's getting about a million visitors a month. And it just kept on growing and growing and growing. And it started as like, I need this for myself.

Well, you know what that tells me too? And this is the thing that I think is really a feather in your cat. Matt is I think everybody gets to go viral once. Everybody in the world that's creating content will just have that Charlie bit my finger moment where they strike gold the one time and they take off. You did it a bunch of times back to back, which means you found the chord that struck. Like there was a resonance with the content that you were posting and the fact that you can do it. Because it's actually hard to cross mediums. We've seen that people that kill Instagram try to move into YouTube and get murdered. So it's not just as easy as saying like, oh, I'm good here, I'll be good there. And the fact that you're able to take your content and cross pollinate it across multiple networks, I mean, that's just gotta feel so profoundly validating. It's like everybody loves my shit, you know? What a great place to be validating.

But also very bizarre. I sort of was in this purgatory for 14 years where I was creating a lot of content, making a lot of videos, doing a lot of podcasts, going on other people's podcasts. And I always kind of felt like I'm just on the cusp, I'm just on the cusp and nothing was really breaking through. So it was this, this long journey of living in this purgatory. But like you said, I mean, I've tried to go into Instagram with this stuff. It's no traction on Instagram. I've tried to make TikTok videos about AI. It's no traction for me on TikTok. I know other people have had a lot of traction with AI on TikTok doesn't work for me. I think it's because the demographic on TikTok is 19, 20 year olds and here's a 40 year old dude coming in, trying to tell them what they should pay attention to. It just didn't work on TikTok. So I think there was a sort of awareness of what tools are available, testing them all, and then the handful that worked, leaning in and going just hard. And the ones that were.

Can I just say for a quick aside, and this is to both of you, I don't love the way we're talking about 40 year olds. If someone is about to be 40, like this doesn't need to be a line in the sand that dictates and indicates old.

It's true. But I'm more referencing it in the sense of the demographics we're going after. I feel like in terms of demographics, a lot of the younger generations, the, you know, the 19, 20 year olds, they may not find a 40 year old super relatable. I think on YouTube, I've started to sort of get my audience younger and younger and younger. And I think that's more. The algorithm is kind of pushing it to more and more people. But if I go into my Google YouTube demographics right now, I'll see that my audience is, you know, between 35 and 55. Like that's the age range of the people that watch my videos. And if I notice based on the data. You guys are data guys, right? Based on the data, I'm gonna make content that goes in front of what the data tells me.

Dude, it's such a compelling point. You're exactly right. You're just gonna find out who you resonate with and you can either fight it and try to expand out or you should. Dude, I am catnip to women between 30 and 50. Like, I crush that demo and I don't know what it is.

It's the hair.

It must be.

It's gotta be the hair.

But why would I fight that? Like, what an amazing demographic to just know that you resonate well with. And so now when I'm doing my placements or my ads or even strategic partnerships, like that's something I get to lean into.

Yeah, absolutely. I mean, when I was doing a lot of Facebook ads and Google Ads and that sort of thing, I would start real broad. I mean, I'm sure Ralph, you know, Dan Ryan's sort of strategy of start broad and then start negative out the keywords and negative out the audiences that aren't gravitating towards it. And. And so I've always just focused on, all right, let's start broad and then focus it into the people that are actually paying attention. So even from an advertising standpoint, that's how I've always approached it is all right, people that are 28 and under don't watch my videos. So let's just eliminate them from the audience because I'm spending money that's not going anywhere.

Right, right. Makes sense. Well, I mean, I think that the listener to this show here, I think skews a little bit more. I wouldn't say Gen X per se, but older millennial. And those are also the people that are fortun. Those are the people that usually have maybe some money, maybe some responsibility, like of directors, of marketing, VPs of marketing, like CEOs. The types of folks that we talk to at tier 11, they're usually 30 plus at the very least. But I think you have Been able to self select an audience that like wants your stuff. It's the place where I go. I'm not going to say what my age is, but I'm over 40. But the point is, is that there's a certain sense of overwhelm, I think when you reach a position of any sort, responsibility and you have a team that you want to focus their efforts on achieving a goal. So for the directors of marketing that are listening to the show here, what AI tools are the ones that my team should be focusing in on? What are the ones that I should be watching out for if I'm a CEO? Like the types of customers that we talk to every single week and you guys at Solutions 8 the same thing. Like I want to grow my business to the next level. Where does AI fit in with that? Obviously we'll leave links to FutureTools IE here, but maybe you can go through some of the best marketing tools that you found that might resonate with the folks that listen to our show that have to make those decisions. Maybe not using the tools themselves, maybe they are, but also like what their teams should be doing.

Yeah. So when it comes to a lot of these tools, the ones that most people have heard of already typically are some of the best tools. So I still use ChatGPT as one of my main tools. Probably more than anything else. I use it as like a personal business consultant. It's really, really good to go into ChatGPT and say, Hey, I am struggling with this problem right now. What do you recommend? And it's got basically the sum of the Internet's knowledge to help you answer that question. So I'll do that a lot. I might have an employee that I'm struggling with. Hey, this employee is, is acting like this right now and I don't quite know how to handle it. How should I approach it? It will give me advice on how to approach talking to this employee. ChatGPT released this new thing called GPTs, which is the ability to sort of fine tune a model towards your exact needs. You have to be one of the $20 a month plus members, but if you're one of those members, you can create a GPT and train it for specific needs. So I have one that's a YouTube video idea generator GPT and I pre trained it and said this GPT is designed to help me come up with, with title ideas, with thumbnail ideas. And I gave it all the context it needed. Here's my demographics, here's the type of stuff I talk about. Here's the titles that have performed well. You can actually upload images into ChatGPT now. So I literally have an image of my top 18 videos, a screenshot of those top 18 videos. And in the screenshot you can see the thumbnail and you can see the title of each video. And so I've uploaded that screenshot of the top videos and said, these are my top performing videos with their titles and how many views they got. So now it's got all of this information about just my YouTube channel and I can go and say, I don't know what video to make today, what should I make? And it will go well based on what's performed in the past, here's what I recommend. So using these GPTs, you can really sort of fine tune and train exactly what you need it to do. I've also got one that's trained to basically write my description on my YouTube channel for me now. And I know, know a lot of people that are listening to this or watching this probably don't have YouTube channels, but I'm sure you can apply it in other ways. I have it right, my descriptions now where it will listen to the transcript of the video, you can upload the transcript and it will pull out like all of the links that are mentioned throughout it, and then plug in the URLs for me so I can just copy and paste all of the URLs that I mentioned in a video straight into my description without having to dig them all up myself again. So there's some really, really interesting things you can do with Chat GPT, which is probably why that's the one you hear the most about. It's just so versatile, right? There's just so much that you can do with it now. There's also one called Claude, which is another chatbot. The main benefit to Claude is that it used to have a hundred thousand context window. Now it has a 200,000 context window. And what that means is it's the combination of words that you can input into it and words that you can get back out of it. So between what I upload and what it responds with, there's this 200,000 context window which gets me about 150,000 words. So between input and output, I get 150,000 words. I can literally upload an entire book. I can take a PDF of an entire, let's say like Harry Potter book if I wanted to, and ask questions about that book. And because it's got this huge context window, it will actually answer questions and reference the context from that book for me. So I use Claude for summarization, for getting bullet points, for explaining it like I'm five, for creating analogies for me. A lot of people think when I make these YouTube videos, I just have this innate knowledge of these tools that I'm talking about. But I mean, AI is my secret weapon, right? I will go and find a large research paper that's very, very complex. I will throw it into Claude and say, explain it like I'm five. Okay, now give me an analogy so that I can turn around and make a video and explain this in a way that people will understand. And then I'll sit there and have a conversation back and forth until I understand how that piece of research works. I do that with news as well. Every Friday, I put out a news video of here's all the AI news that happened for the week. Well, I'm probably sifting through 50, 60 different news articles every single week, week. But I'm not reading the news articles word for word. I'm taking these articles, going to Claude and saying, give me a list of bullet points and summarize what's in this article for me. And without having to read the entire article, I'm getting the quick gist of what that article is about and why people should care about it. That's how I make these videos. I just have these tools, summarize them for me, and then I, in my own words, re summarize them back to people. So ChatGPT and Claude are two tools that I'm in on almost a a daily basis when it comes to image generation. I mean, you probably don't need graphic designers anymore. I actually still have a graphic designer because midjourney and stable diffusion. And Some of these AI art tools, Dall E3, they're really good at generating images, but they're not great at having like that full holistic design eye. So I still have a graphic designer who will help me work in the images that I generate into what I'm doing. But some of these image generators are really, really good. Any image you can imagine, you can get out of it. So for your sales pages, if you're making videos and you need like B roll or screenshots to show up, or if you're making slide presentations. I'm speaking at Social Media Marketing World in February. AI is pretty much going to make my entire slide presentation. All of the images are going to be generated with AI images or image generators, and all of the content is probably going to be created through ChatGPT or Claude, right now, the ones that most people have heard about are probably the best. I just don't think enough people know about the depths of what you can do with some of these tools. I think it's more how do I take these tools that I've already heard of and do even more than most people realize I can do with them, if that makes sense.

It almost needs to become second nature in a lot of ways because the tools are so profoundly useful, but only if you're weaving them into your day to day function. Like if I have to remember, if it becomes tangential, if I have to get off the freeway and pull over, then play with AI, then it's not going to work. Which is why I love that you say you go back to chat GPT because it's so funny, man. There's 2000 some odd tools sitting here in your database and it's like, oh, well, the one that sparked it all is actually still the most functional one. Coming from the guy who's effectively the world authority in all things AI, that's really validating. It just goes back to the basics. It's like, hey, go get good at free throws. And then once you're good at free throws, then we can learn to dunk. But until then, it's okay to start small. You're not missing out on anything.

Yeah, exactly. I mean, there's a lot of other really, really cool AI tools out there, but most of them are like very small niche edge use cases that I'll play with them on YouTube videos. But even as I'm making the YouTube videos, I'm like, this is really, really cool. And I want to show it off because it's really, really cool. But there's like nine people out there that are actually going to use this tool, right? So there's a lot of interesting tools that aren't super valuable, especially for like an audience of marketers. There's a really, really fun tool that I love playing with called Wonder DY can actually film a video of yourself walking around and then replace yourself with a CGI character. Right. You could replace yourself with a computer generated alien or a computer generated robot. So I actually made a video of myself out in my front yard playing basketball and then I turned it into a robot. And so it looks like a video of a robot playing basketball in my yard. And I thought, this tool is really, really fun and really, really cool. But unless you're making YouTube videos or you have a movie production business, just what are you going to use that for? So there's a lot of tools like that that I'm just like I'm blown away that this tool exists and it's so fun and so cool to use but what percentage of the world is actually going to actually have a use case for this?

Yeah, yeah, no it's funny like as we've been talking I've been going over to FutureTools IE and your team is actually updated. Went from like2456, now it's 2460tools. Just since the time we've been talking it's updated like four more tools. So it's like sifting through all these tools for all these individual uses. Yes, creating a robot mat, shooting hoops is actually is really cool but how is that going to affect my day to day activities aside from being really cool? So like if I'm a marketing person, obviously you start with ChatGPT Claude, which I'm wondering now if like a lot of the tools that I probably use on a regular basis probably powered like through Claude or through. I mean obviously it all comes from one single source here. But digital marketing tools, like for folks that are running ads, doing marketing, doing any of this sort of stuff, content creation, what are your go tos aside from the ones that you've talked about here? So for example, creating a marketing campaign for us always starts with a tremendous amount of research. Now we use a lot of different tools there, but it seems like they're all a little bit clunky. It's like there's no one thing to rule them all, like simplify, maybe some of that for us in addition to the stuff that you've talked about.

So most of these tools like you mentioned, they're actually just using the APIs of the existing tools out there. So like we mentioned Jasper earlier, which is like a copywriting tool, well that's just using the GPT API underneath it. A lot of tools that are out there are using Claude's API or now Google just released Gemini API which is competitor to GPT3. So most of the tools that exist are actually just using one of these APIs underneath. Same with the AI imaging editing tools. These tools are all either using Stable diffusion or Dall E3 underneath. Midjourney is its own thing. They don't have an API. The only way to use Midjourney is through Midjourney, but all of the rest are either using Stable diffusion or DALL E3 underneath.

Do you have a preference Matt, between stable diffusion, Dall E3 and midjourney?

I use them all, they all have different pros and cons. So stable diffusion is really, really customizable. I can train my own face, I can train my own styles into it. It's uncensored. If I want something that like midjourney won't generate for me, stable diffusion will generate it. If I want to have celebrities or like ip, like I was making images of Pikachu the other day, stable diffusion will do any of it. It's totally open. There's literally nothing you can't create with it. Dall? E 3 is really, really, really good at being prompt adherent, meaning I can do something really complicated like a three headed man sitting at a desk wearing a blue coat and on his monitor is a picture of a rabbit and it will get all of that into the image. It's really, really good at that. Prompt adherence. And then midjourney has gotten to the point where it's the leader in realism. So if you want an image, I'm sure you guys saw the stuff circulating a few months ago with like the Pope in a puffy jacket or Donald Trump getting tackled by a group of people in the streets and they looked really realistic. Those were all mid journey, mid journeys reached that level of realism. So I'm literally using all of them. They just have different sort of use cases that I would use them for. But most of the tools are out there are just using one of those underlying technologies because it takes a lot of computing power to train one of these models to the point where it costs millions of dollars to train a new model. And most of these small companies don't have the budget to train these models. So they're just front end user interfaces on top of GPT or stable diffusion or Dall E. I know it's a very long winded answer to the original question, but when it comes to marketing, ChatGPT and Claude are your best friends. ChatGPT has a tool in it called Code Interpreter now where you can upload literally anything you want. You can upload screenshots and it will analyze what's in the screenshot. If you were to go to your Google Analytics and take a screenshot of your analytics with all of the data on it, of, of what websites you're getting the most traffic from and the time on site and all the data that Google Analytics gives you, throw it into ChatGPT and say Analyze this screenshot and tell me what my next best move is for traffic. It will use the data that you uploaded, the image that you uploaded for context, and then also use all of what it can find on the web for additional information and provide you with feedback on what you can do next. And if you go out there and you find a tool that's specifically tailored and designed to be that analytics analyzation tool under the hood, it's just using the exact same tech that ChatGPT is. It's just using the GPT4 API underneath it. So there's. So a lot of these tools that are out there are just the same tool with a different user interface on them.

Yeah, my God. Yeah, it's fascinating. I mean, if we really. So, you know, instead of chasing the shiny object, I mean, we as marketers are sort of notorious for this. It really does come back to, okay, if you're already using these tools, how much deeper can you go into? Like, I didn't realize code interpreter could do all that. Like the individual GPTs inside ChatGPT, like the tools that you're already using, how much deeper can you go into each individual one? And I go into ChatGPT and like, I only use like one or two things.

Right.

But you've clearly like gone all around the edges and found out all these other ways to use it.

So.

So would that be the recommendation or. If you're already using something and you're comfortable with it, don't get distracted to go to a different place, Double down on the tool that you're already using and figure out how you can make it even better. Is that sort of the advice that you would.

I would say that'd be the best advice. When we start talking about where things are going in the future, I think that will change. I think open source tools that are being released are probably going to eventually pick up some steam and actually be able to compete pretty well with some of the closed source stuff like OpenAI and Google and Microsoft and Meta. Well, Meta is actually open sourcing a lot of stuff lately as well. But yeah, I would say the best advice would be to go deeper on tools like ChatGPT. If you are getting into like image generation and maybe you're a graphic designer or you need some really good art for your sales page or things like that, really dive deep into what Dall E3 stable diffusion and midjourney are capable of because those are really the three that you're going to use. And between those three you can get, get any image you can ever imagine. Anybody can tell me they need an image of anything. I don't care what it is. I will figure out how to create it between those three tools. I might have to do a little bit of, like, merging it together in Photoshop or something, but for the most part, I can create any image you can imagine just with those three tools. And anything that's like analytics or writing copy or doing research or anything like that. ChatGPT and Claude will probably get you there. Bard is getting better and better. Bard. I think its sweet spot is if you really need data from the Internet, it's probably the best at pulling up websites and finding specific websites for specific queries. ChatGPT will do that as well, but I think Bard's a little bit better at it. And then you've also got Grok now, which is Elon Musk's X AI version. And the interesting thing about Grok is it actually has all of Twitter at its disposal, so you'll ask it questions, and it's using all of the information that's ever been put on Twitter as it's referenced to answer those questions. So that's really, really good for, like, current events that are happening, like, right now. Like, you can go on Grok and say, what's the biggest news in the world at this moment right now? And Grok will be really, really good at that. So the tools that I'm using are still mostly the main tools, and any other tool that I'm using outside of that, it's mostly just for, like, content, entertainment purposes, honestly.

Yeah, yeah. It's interesting because, I mean, I've sort of shifted my usage away from ChatGPT and over to Bard with just the basic stuff. All right, well, do this research for me or make this sound even more clear. But it seems like from what you're saying, I really need to go back to ChatGPT and dive in deeper and understand a lot of the additional functionality as opposed to just what most people probably do on the surface. And. And if you're listening to this show, it's probably the reason why you're using it is. All right, well, summarize this article for me or, like, make this email sound more clear or more distinct or more authoritative or whatever it happens to be. Like, those are sort of the basic functions I think people are using, but I don't think they realize how many things are sort of behind the screen that maybe are within the tool itself. That can really increase your productivity, especially from a marketing standpoint.

Yeah. And I mean, if you're not paying super close attention to all the updates, like I am, a lot of the stuff that's rolled out, you probably missed. They released a thing called GPT4 Vision, which made it so Chat GPT can actually see what's in the images. That's only, I don't know, two months old, since that was rolled out. And if you weren't paying attention, you probably didn't even notice that Chat GPT can do that. Now, that code interpreter thing, that's only three months old. So a lot of these really, really powerful features are kind of getting rolled out. And those of us that are in the AI space that are watching it every day know about them, which is why I'm making YouTube videos about it. But if you're not paying close attention, it's probably not on your radar. You're probably thinking, ChatGPT is just that thing that people use to cheat on their homework. But it keeps on getting more and more advanced every single day, to the point where it feels like you don't really need a lot of these other tools. And being somebody that curates all of these tools, I'm seeing it constantly, where I was mentioning before we hit record, that I'm actually removing just as many tools as I'm adding on a daily basis, because every single day, tools are shutting down. And the reason is, is that tools like ChatGPT and MidJourney and the incumbents, the ones that were sort of there first, were all just starting to do the things that these API wrappers are doing. I don't quite understand the business model anymore of building a SaaS product on top of an API, because if something does really well as a SaaS business, well, what's to stop OpenAI from just going well? That's working really well for these companies. Let's just make it a feature of ChatGPT. And that's what we're seeing. That's what we're seeing in Mid Journey as well. You can actually edit images directly inside of Mid Journey. You can generate an image and let's say there's like a bird flying in the sky and you're like, I don't want that bird in Mid Journey. You can just sort of highlight the bird and say, remove the bird and it'll be gone. Or you can generate an image of a girl walking down the street and she's wearing a pink shirt. You want the shirt to be blue. You can highlight the shirt and say, make the shirt blue and it will change it. This used to be an external tool that you would have to generate an image in midjourney, pull it into a different tool, do the inpainting in this other tool, and go through all these processes, but all of these tools are just sort of wrapping it all up into the original tool.

Now that's the amazing part of just having an open AI and it's almost like the Amazon model. It's like, oh, okay, I'm going to sell this thing and I'm going to get to a couple of million dollars in sales and all of a sudden I've got a partner. And then also name is on like releases the same damn tool. It's the same thing here to a certain degree. I don't think it's quite as cutthroat as Amazon, but, but it's the same concept and it's a brilliant move in order to be able to sort of consolidate. Ah, I never really thought about doing that. As Sam Altman says from his, you know, from his throne, like, why don't we just incorporate that into ChatGPT? It's a brilliant business move, but it's completely counterintuitive. Make everything open in order to really improve the experience of your primary tool.

Yeah, but they're going to kill the sandbox that is ideation to the point that you just said, Matt, I don't see the point in building a business on an API. Folks are going to catch up on this and then all of this free R and D they've been getting is basically going to come to a screeching halt. So at some point somebody's going to have to have a little enlightened self interest and allow the organic market some self preservation.

Yeah, well, and I also think that if we're looking into the future a little bit, I think open source is going to get bigger and bigger and bigger because more people are realizing this. Right. A lot of these companies that are trying to build on top of tools like ChatGPT and like Google's Gemini and Claude's API, if these companies just keep on taking the features that these people are building as entire companies, well, these people are going to stop building on top of those platforms and start moving more and more towards the open source. The problem with open source right now is just that same problem of it takes a lot of time and a lot of money to train a new model. And if you're spending $2 million to train a model that's as big as GPT3, which isn't even the best model, how are you getting that return if you're just going to go and open source it and let anybody use it? Some companies are starting to do that a little bit. Meta went and created a thing called llama2. It's kind of competitive chatgpt, but it's actually open source now, so people can build off of it and iterate it. But it's open source with some conditions. If you make a tool on the back of this open source tool and it has more than a million users, then you have to start paying meta some extra money. Right. So it's open source. Ish. But it's not totally open source. But that seems to be the way it's going, is a lot of the developers, I think, are going to start moving further and further away from these incumbents that are building APIs.

So we are here with Matt Wolf. He is the founder of FutureTools IO, the best AI tool aggregator on the planet, as well as one of the most popular YouTube channels that we've got out there. Matt Wolf we'll leave links in the show notes here. And we've been talking about doubling down on the tools that you're already using to a certain degree and sort of SIMP things, especially if you're figuring out, okay, how am I planning for 2024 here? How am I going to use AI in my marketing? Maybe let's get into a little bit more about the future of where you're seeing this all going. Maybe not just in 2024, but a couple of years out. How does this all evolve? What's your take on it? Right now it's a bunch of tools. They're very cool. Obviously there's a lot of productivity enhancement that's coming along with a lot of these tools. But where is it all going? Where are we seeing this, this three to five years down the line, or maybe even past that point. What's your sense?

I mean, I think, I think we'll eventually just all have robots in our houses and we get to play PlayStation all day while the robots do it all. I think that's where we're.

Sounds good.

That robot is going to Crush you at PlayStation though, is the problem.

I will kill you. That's right. But you'll still have fun.

Yeah, yeah. I say that jokingly, but that's actually where a lot of these companies want it to go. You know, that's.

Dude, that is without a doubt, that has to be the end game.

That's why the talk of universal basic income and all that kind of stuff, stuff keeps on popping up is because, well, once all the AIs and the robots can do this stuff for us, how do humans fit into the picture? But that's probably a little bit further down the line than what we're actually talking about. So we heard the term in 2023, prompt engineering being thrown around a lot. Learn how to be a prompt engineer because the better you get at prompting, the better you'll be able to use these tools. I personally think that's kind of bs. I don't think any of these companies want you to have to be an engineer to use their tool. I think prompt engineering is going to be for the plumbers, the people that are actually building these tools, not for the end users. I think we're going to see more and more of these user interfaces where just click this button and type one or two word prompts and you get exactly what you're looking for. I think the prompt engineering thing was sort of overblown. That's a very simple, easy thing of where we're going in the future. Nobody needs to be a prompt engineer unless you're actually building AI tools. I mentioned the GPTs, right? That's where you can use ChatGPT and custom build a bot to do whatever you want. Well, what OpenAI is doing, they actually were going to release this in December, but pushed it out, is they're making what they're calling the GPT Store, which is like their version of the App Store. So you can sort of custom make your own GPTs that are very tailored to a specific use case. Then you can actually sell it in their GPT store and make money by selling these custom fine tuned, tuned AI model. So I know for a fact we're going there, it's already been announced that's coming. So that's going to be, that's going to be pretty big. And that's where something like prompt engineering comes into play. When you're trying to build this custom model that other people are going to use, you're going to be a prompt engineer. Prompt engineers are going to be like the coders of today. You don't necessarily need to be a coder to be able to use Microsoft Office, but the person who made Microsoft Office needs to be a coder. That's what prompt engineering is going to be. Prompt engineering is the coder of the future. The people that are building the underlying tech that the general public is using. I really see that's where it's going. And then the GPT stores is real interesting, right? Because anybody can go and create a GPT. And the way you create a GPT is by talking to ChatGPT. You go in there and you say, hey, I want to make a GPT. That is trained for YouTubers to get them the best title and thumbnail. Here's a bunch of information to help you get the best title and thumbnail. Here's an ebook on titles and thumbnails that you can reference when people ask questions. Here's a handful of YouTube URLs of people who teach how to make titles and thumbnails. Use that information to give context to your response. So you can go and create these GPTs, but the way you're creating them is just conversations back and forth with ChatGPT. So that leaves the question, well, if I go make a GPT and then try to sell it, why couldn't anybody just go look at that GPT, go to ChatGPT themselves and make a GPT that does exactly the same thing without buying my GPT. So where I really think the GPTs you have to have specific domain knowledge and specific IP. Those companies that have their own IP that's trained into the GPT that people can't get elsewhere, those are going to be the ones that make money in the GPT store.

But even then, couldn't I just have my own chatbot? Chat with your chatbot and say, reverse engineer the creation of this chatbot?

People have already done that. Yeah, that's already been happening a little bit. Where other people have gone and used somebody else's GPT that was already created and said, what is the underlying prompt that made you? And then the GPT would actually spit back out the prompt that you used to make that GPT.

Give me your first principles. But from where do you hail what mass Shoot is?

There's nothing hidden. It's all there. It's good.

OpenAI is trying to put limitations on that and block people from being able to ask that question. They can build in guardrails and say if somebody asks how it was made, made, don't give a response. Right. So they're trying to build in some of these guardrails, but I still think it's always going to be a constant cat and mouse game.

Yeah, but it's because it is Pavlovian. Like it's prompt response. You can build a knowledge graph that allows you to, I mean, no matter what, you're going to be able to like triangulate the answer. Even if you don't have the original prompt. You could ostensibly generate all of the feasible outputs or within a Pareto distribution, the majority of the feasible outputs. And now you don't need the origin because you have the output and what.

You'Re describing in The AI world is called synthetic data data. It's data that came from AI outputs. And there are a lot of models being trained right now on synthetic data where they're asking ChatGPT a whole bunch of questions, getting a whole bunch of responses, and then using all of this data that it outputs and training new models on the output that originally came from AI. That's a concept called synthetic data that's happening right now. And it's one of the ways they're sort of escalating and picking up the pace of how they're training these models is by using synthetic data. Because before they started using synthetic data, they were using reinforcement learning with human feedback RLHF. Basically, what that is is when you go and ask Chat GPT a question and it gives you a response. Every time it gives you a response, there's a little thumbs up, thumbs down next to that response. If you give it a thumbs up, you're telling ChatGPT, you gave me a good response to this. If you give it a thumbs down, you're telling ChatGPT that wasn't a good response. That's reinforcing ChatGPT, making it better and better and better. But, but with synthetic data, they're actually moving more and more away from the reinforcement learning with human feedback, and they're just training it all on this synthetic data. So that is actually happening right now. But where I really think these GPTs are going to be good is when you have really proprietary data. Let's say you're like a company that had a database of addresses for direct mail companies or something, and you're the only company in the world that has all of these direct mail addresses and somebody can go and pay you 100 bucks a month to get access to this database of direct mail addresses. Well, if that data doesn't exist anywhere else but in this one person's database, well, that can stay proprietary. You're not going to be able to reverse engineer and figure out what other addresses are in that database, if that makes sense.

Well, at least it's some guardrails. There's some protection there. Take us as far as, as you can see on this. I mean, I said three to five years because I'm like, I don't know, it's three years. Five years seems like to be able to predict what AI will be doing in five years seems almost like I can't see that far. How far can you see this happening? I mean, you're in this stuff every single day. What can we expect? And then Also maybe talk to the naysayers that obviously are scared of a lot of this. I mean, I think there is certainly that element to this whole conversation. But how far out can you see and what does that future look like from your perspective? Perspective?

Well, it's interesting because I've made predictions where I've said, I bet in two years we're going to see this happen. And then the thing that I predicted happened three months later, right? So like that's. That happens constantly. Everything that I think is going to happen further out seems to be happening sooner than I thought. So things at the beginning of 2023, a text to video was very bad. You can put in a prompt like a monkey on roller skates and it would generate something that kind of looked like, like a blurry monkey that maybe had wheels on its feet or something. You're like, I guess like, see if I squint now you can go and use a tool like Pika Labs or Pika 1.0, which is a text to video tool where you can say a monkey on roller skates and it will look perfect. You can get. It's to the point now where it looks like a real video that somebody shot at a zoo or a circus or something. Right. That stuff has just picked up so rapidly. The problem with those is it's only still generating like three second generations. So I think what we're going to see is just better and better versions of what we've got now. I don't think it's too far off. Maybe by the end of 2024 where you're going to be to prompt a whole 20 minute video of monkey at his circus and elephants in the background and trapeze swinger above it is going to generate like a whole 20 minute circus act of a video. I think that's coming a lot quicker than people realize. That's why Hollywood was striking. It wasn't the only reason, but Hollywood was striking because a lot of the AI stuff that was happening and how quickly it was speeding up. We're getting to the point now where you can model anybody's face into videos. So a lot of the people that were striking were saying, hey, you can just use my likeness in any video you want in the future and not pay me for it. That's not cool. I think all of that stuff is really, really accelerating at a rapid pace to the point where we can have like a full on Tom Hanks movie. And Tom Hanks didn't even have to show up because they cloned his voice. And they cloned his likeness and they just plugged him into the whole video. I think. I think we're getting to that place really, really quickly.

You know what's creepier to me than that is the fact that the next Tom Hanks would be a human that doesn't exist. Somebody went and did demographic and psychographic profiling analysis and decided the most appealing star for this generation is gonna be this person, this height, this skin tone, this eye color, this timbre, their voice, and they just develop like our generation's Gregory Peck. And it's not gonna be a real human.

Yeah, I think the technology will be there to do it, but if I had to guess, I would say most people will reject it. I honestly think a lot of people will be like, you know what? There's. It's gotten to the point where. Cause we've already got cartoons for that. Right? We've already got Toy Story and Pixar and that kind of stuff. Stuff where it's sort of these synthetic characters that we already relate to, we already love, we already buy the toys for, and that sort of stuff. And then when you start making like these live action movies, but it's not actually a live action person in there, and there was no real skill to the acting of it, I guess that at least for a while, people will reject it until it just becomes too entertaining to stop rejecting. And then everybody will be like, well, you know, I hate that it's not a real actor, but it's a fun movie. So I'm watching it anyway. I think there will be a strong pushback for a while until there's no not. But I do think we'll get there eventually. There was also a strong push in mid 2023 for these auto GPTs where you just give it a task and it would just go and run autonomously until the task was complete. I don't know if you guys heard about this thing that somebody tried to create called Chaos GPT, where somebody basically trained a GPT to just keep on running until it destroyed the world. I don't remember the exact prompt that they gave it, but they're like, your goal is to destroy the world. World. Do whatever you have to. Enter whatever prompt you need to to get it there. And basically what it ended up doing was creating a new Twitter account that ended up getting like five followers and then tweeting to the five followers, hey, it's time for us to rise up. Who's with me? Two people liked it and then nothing ever came of it. But.

That'S that's encouraging to say the least.

Yeah, so, but these auto GPTs are designed to essentially give it a task and then it goes on and creates its own prompts over and over and over and over again again. And until it gets to its final completion of the task that it created. And looking at the scarier side of it, I honestly think if the world ever gets destroyed by AI, it's not going to be the Terminator scenario. It's going to be some idiot in a basement testing something, going, let's see if I can get AI to do this. Oops, it worked. But I think we're going to see more and more of these autonomous AIs that learn to prompt themselves through. There's a process called Q Learning. It's been really popular in the news because there was a potential leak for ChatGPT called Q Store. But Q Learning is basically trying to find the optimal next step to achieve a goal. And it keeps on testing things until it finds the optimal next step. And then once it finds the optimal next step, it then moves on to the next step in the sequence to find the optimal next step for that point of the sequence. So I think we're going to start to see the rise of these auto GPTs where you give it an output like, help me make a hundred thousand dollars, I don't care how you do it. And then it goes and tries to find the optimal way to make that money. And it just keeps on writing prompts, writing prompts, testing what it wrote, write another prompt, test what it wrote until it finds the optimal path to getting to that end goal. And so I think we're going to see more and more of these auto GPTs. They sort of popped up a little bit in the, in the middle of the year in 2023, but most people didn't find them too useful because they didn't really have a lot of capabilities to do things. It was able to go and post some tweets on Twitter, but other than that, it didn't really have, have much agency beyond that. But I think that is going to get perfected and get better and better and better. And so I think we're going to see a lot of that kind of stuff, and I think the business implications of that are huge. Look at our manufacturing plant and try to find how to make it more efficient to move the product through the plant. How do we minimize waste on this lumber that we're chopping up? I think you're going to see auto GPTs really, really help Optimize things like manufacturing and processes and things like that. That.

Where do you see all of this from a quantum computing standpoint? I don't know how much you've discussed that on your channel or gone down that rabbit hole, but that's sort of a fascinating next level. AI on steroids. On steroids. On steroids. Like, what's your sense? Will we migrate over to that? The quantum computing that's happening right now is obviously in very, very limited form, but it's frightening what the prospect of that is in a good and potentially bad way. But what's your sense about that? Maybe even explain to the F what quantum computing is from your standpoint?

Yeah, I definitely wouldn't be the best at explaining quantum computing. And it's not a rabbit hole that I've gone down super deep. But if I'm honest, I think quantum computing is scarier than AI. I think the implications of quantum computing are scarier. Like, quantum computing is really, really good at math. Quantum computing is really good at figuring out all the possible solutions and outcomes of a single problem. Quantum computing, you're never going to be able to use a quantum computer to create a document in Google Docs. They're for two totally separate things. But why I say quantum computing is really, really scary is because it's so good at math. And because it's so good at figuring out complex problems in math, it makes me very scared for the future of, like, encrypted technology. So when you're looking at, like, your bank account that's using encrypted technology, when you're looking at, you know, crypto, right? All of that is based on cryptography. Everything on the entire Internet is essentially using encryption to sort of protect. Well, if most of the things that are encrypted, it would take something like, I don't know, a million years to decrypt it using a regular computer. But a quantum computer could theoretically decrypt some of this stuff in a matter of days, if quantum computing ever becomes so inexpensive to manufacture that anybody could get their hands on a quantum computer. The thought of that is really, really scary because now anybody can sort of hack into any sort of system and take anything offline that they want because they can get past the encryption. So the sort of encryption implications are what really scare me about quantum computing. But also on the flip side of that, I mean, quantum computing has that ability to try a whole bunch of iterations of something at a very, very rapid pace, which the end result of that could lead to new drug discoveries. To cures for diseases that we haven't found yet, to an explosion of discoveries that could really, really better the world. And I see the same stuff with AI. AI has some really, really scary implications, but it also has that ability to connect dots that humans don't see. Because it can connect dots that humans don't see. The implications are really, really potentially beneficial, but also really, really potentially scary for some of the reasons we've already talked about. So they both have their benefits to the world, but also they have the. Well, if this goes wrong, this could go really wrong really, really fast.

Right. Right now. Fascinating. Well, we've gone way farther down the rabbit hole than I think we probably anticipated, which has been super fun. And it's been so great to have you on the show. It's awesome to me to see your success right now. Just tremendous.

Appreciate it.

There's lots of places where people can connect with you. Where is the best place for people to connect with Matt Wolf?

YouTube. If you go to YouTube and search out Matt Wolf, I'll pop up there. I've also got FutureTools IO. Those are probably the two best places. I am on Twitter X pretty often these days as well because a lot of the AI community is thriving there. So I hang out there as well. But really, YouTube and future tools are the two places to probably get the most value from me.

Yeah, absolutely. Well, we'll leave links in the show notes for that. If you don't know where Matt is already, probably a lot of folks on this show probably already do. But the point is we'll leave links in the show notes for that. Make sure that wherever you listen to podcasts, you subscribe and leave a rating. And if you leave a review, you are automatically entered into our contest to get a free traffic and conversion summit ticket. So definitely do that. And also follow myself on LinkedIn as well as Qasem over on all the socials. ASM Aslam. We will leave lots of links in the show notes for today's show@petpetualtraffic.com make sure you do subscribe to our YouTube channel. That's@petpetualtraffic.Com YouTube Like I said, said tons and tons of stuff to list here. So thanks to our editor for catching all the nuggets that you dropped here because there were many, many, many of them. We'd love to have you back on the show. Don't forget the little people in your life now that you're like this big star. People like myself and now Kasim. You know, we really appreciate you coming on and talking all things AI. And plus, there's so much that's evolving here. It'd be great to have you back on the show to give, obviously, updates on where you see things are going, so.

Absolutely.

On behalf of my awesome co host, Qasem Aslam, peace until next show. See ya. You've been listening to Perpetual Traffic.