Joe Casabona: So if you're a long-time listener, you will know that I am AI hesitant at best. I've had a number of episodes with AI advocates where I've, we've challenged each other, I'll say.
And today's guest, Damien Schreurs, is here to make his case for AI, except where I thought we were starting on different sides or opposite ends of the spectrum, we're actually a lot closer in our usage of AI, than I thought we would be. So, I'm really excited for you to hear this. We do get in the weeds a little bit towards the end, but I think there's a lot of really good stuff, especially if you're looking for ways to improve your processes and workflows with AI. So without further ado, let's get into my interview with Damien Schreurs.
Welcome to the Streamlined Solopreneur, a show for busy solopreneurs to help you improve your systems and processes so you can build a business while spending your time the way you want. I know you're busy, so let's get started.
Joe Casabona: Alright. I am here with Damien Schreurs. Is that right? How did I do?
Damien Schreurs: Yeah. Very good.
Joe Casabona: Perfect. He is the explainer in chief at EasyTech, and he hosts the Macpreneur podcast, which, I had the honor of going on, and I was very excited. Damien, thank you for being here.
Damien Schreurs: Thank you for having me, Joe.
Joe Casabona: My pleasure. So let's start. Let's just answer the big question right now. Is AI a fad?
Damien Schreurs: So for me, it's like going 30 years in the past and asking whether the Internet is a revolution or a fad. Yeah. For me, yeah. So, yeah, AI is definitely a revolution. And for me, I compare that to the invention of electricity.
Joe Casabona: Okay. And, you know, you're, I agree, first of all. Right? I mean, I think long-time listeners will know I'm AI hesitant. And, actually, the reason that you reached out to come on the show is my interview with Alastair McDermott where we kinda had a spirited debate over uses for AI. But, I am reminded of, gosh. I think it was, I think it was the cotton gin, caused, like, a mass pro like, a bunch of protests and, like, a riot, in The United States at least because people thought technology was taking their job. And so we've been seeing this since at least the Industrial Revolution. Right? There's a new technology that helps us do a job better, and people fear what it means for them and their livelihood. Right?
Damien Schreurs: Yeah. And, yeah, until now, cognitive tasks were kind of not too impacted by technology. I was at, actually, an AI and cybersecurity conference this morning in Luxembourg. And there was a guy who was actually kind of half joking about the fact that technology over the past twenty or thirty years hasn't really, really improved our life. Yes, it has. But it has also brought some distractions and stuff like that. We're still submerged by emails. Spam is at a record high. But he said that AI will be or is the technology that for the first time will help us, right, relieve part of our cognitive tasks, and give that to to computers.
Joe Casabona: Yeah. That, you know, that makes me think of, I'm trying to remember. Is it just called The Jobless Future? There was a book that came out in, I don't know, the mid-nineteen hundreds, I would say, that kinda talks about how in the future, nobody will have a job because technology does it all for us. Right? But you're right. Like or the the guy who was half joking at this conference was right.
So far, technology has only made us work more. Right? Because it allows us to work anywhere. And what you say spot on, right, about, cognitive tasks, right, that's spot on. They have largely been, unaffected by you know, we could research more, but the creative thinking process was untouchable. I think you and I probably fall differently on what it is now, but it was untouchable until a few years ago. And now it feels touchable. That's not really the way I wanna word it. But, and I think people are freaking out about that.
But, before we move to that, I wanna level set. Right? Because AI stands for artificial intelligence. There's also AGI, artificial general intelligence. There is something called AI agents, which this is I don't wanna time stamp this episode too much, but Circle, the community website, announced their AI agent, and it's not an AI agent. It's like a glorified chatbot, and that really bothers me. Because, like, the promise of an AI agent is so much more than training me on your content. Right? It's like I'll do things for you. I will do cognitive tasks for you. And what they've released is not that.
But there are a couple of terms that, we were talking about preshow that people conflate with AI as well. Right? So do you wanna talk a little bit about that?
Damien Schreurs: Yeah. That's a pet peeve of mine. Whenever I see, an article about LGBT and then I see a pensive robot, like, the humanoid robot depicted, I like no. No. No.
So, if you look it up in Wikipedia, the term machine because oftentimes I hear machine or robots when people talk about AI. But a machine is something, it's a physical object that exerts force and movement. Right? So, imagine something in a factory. Right? Basically, a car is a machine. Right? Because there is motion, there is movement, there is, there are forces applied on the wheel, and so on. So that's a machine.
And a robot is nothing more than a programmable machine. Right? And when people talk about AI in that context, I think they go maybe two steps further than what it is today. They may be the…it's the unfortunately, I think the wrong image that has been depicted at the cinema. Right? In movies. Right? It's always negative. The future, the robot, the killer robots, and stuff like that. But yeah. So it's not a robot. It's not a machine. It's a computer. It's a smart computer or it's, well, actually, the term that I use is brain simulator. Right? What we have been able to do is to simulate the brain, the way the brain works. And if you look at the brain from this side, so scientists, through FMRI, have been able to locate areas of the brain that can process speech, process vision, and can also, when we talk, we activate some parts of the brain. It's usually, in one place.
And then we have this really unique for humans, the prefrontal cortex, which is where we have all the reasoning part of it. And until, I would say, mid 2024, what the AI companies have been able to do is the back part of the brain. And with O1 Mini, and some people have heard about DeepSeek, the reasoning model, and so on. Whenever they added the reasoning capabilities, they basically simulated the prefrontal cortex of our brain. Yeah.
Joe Casabona: Yeah. And that I mean, that's pretty so we covered a lot of ground right here. I think I agree with you that we should use the correct terms. Right? Because I read in a great book, called The Coming Wave. It's by, I'm gonna, I don't wanna mess up the author's names, and I don't know that I can pronounce them properly. It's by, well, Michael Bhaskar and Mustafa Suleyman, I'm gonna say. And it's really about how humanity fundamentally changes in waves. Right? And so there was, like, the industrial revolution is a good example of that. Right? When fire was invented, when language was invented, and the Internet helped usher the last wave in. It was like the communication wave. And now the wave that we're seeing where it's a convergence of a bunch of technologies is AI or AGI with the advancements that we're doing in robotics. Right? Where we're combining the machine and the brain simulator. Right?
Damien Schreurs: Mhmm.
Joe Casabona: Where we are creating synthetic biology. Right? Where, you know, we're creating, like, smarter prosthetics and using technology to tap into the brain to give people the ability to see. And so really interesting.
The book, I feel, is a little bit alarmist, but I also think it needs to be, because, you know, we can focus on the good all we want, but whenever there is a lot of good, there's also a lot of bad.
And so some of the stuff that you just brought up there really makes me I would really encourage people to read that book. I think that while it is a little alarmist, it does kind of give us what is or could be possible.
And something that made me a little sad was a story he tells about, you know, soon combining robotics with AGI or AI agents. You know, soon you won't really need a waiter or a busboy. Right? They'll just, it'll be a smart robot that comes to your table and takes your order and cleans everything up. And I thought, I always liked you know, I like to make friends with the wait staff. I think there's maybe, like, a cultural difference between The United States and Europe, as far as, like, wait staff goes. Where in America, it feels like you're always getting rushed out of the restaurant because they rely heavily on tips. But, you know, I try to be friendly, and make conversation. And that's, like, a little human interaction that we could lose in a few years.
And I think the immediate negative reaction you, in the preshow properly equated to the five stages of grief. So, do you wanna talk through that a little bit, and tell us why you are seeing the world's reaction to AI as the five stages of grief or through the five stages of grief?
Damien Schreurs: Yeah. Because, so what sparked me to think about that is a documentary. I don't know if you've watched that one, the AlphaGo documentary.
Joe Casabona: No.
Damien Schreurs: It's on YouTube. And so, basically, AlphaGo has been created by DeepMind, which now, is, yeah, is part of Google. But, basically, in it's, it's a fascinating documentary. So Go is a…
Joe Casabona: I just wanna bang in here and say I think this is these are the same the same guy who wrote The Coming Wave was involved in this project. Sorry. Go ahead.
Damien Schreurs: Yeah. Possibly. Yeah. Possibly. Yeah. Because, Suleyman's, it reminds me of something. And, basically, the game of Go is considered to be one of the most challenging games for a computer to learn.
And the the reigning champion Lee Sedol was, like, 100% sure that maybe the computer would win one game out of five, but he would win easily. And he loses the first match. He loses the second match. And you see the, do you see the guy questioning his humanity? Right? And in the third, I don't remember in which game, he's very well, like, he's he’s on par, he's really about to win. Right? He's really good for that game. He, basically had two games again against the computer. He started to pick up some patterns, right, and so he started to realize, okay, I can maybe outsmart that computer. And then the computer makes a move that everybody, including Lee Sedol, is like scratching his head. It's like, what's this move that's, that's right?
And actually it afterwards, people realized it was not only a smart move but an innovative move, right? And that made him even question even more his talent, right, thinking, okay, I'm a human. I will be always better than a computer.
And so the stage, the five stages of grief are denial, right, when he entered into the tournament, right? There's no way it's going to happen. And then the second stage is anger. And you see him take cigarette breaks and be really angry. And then you have, bargaining, then you have depression.
You see, at one point, I don't they don't say it in the documentary, but I'm pretty sure he was about to throw the towel and not do the five games. Right?
Joe Casabona: Mhmm. Right. Yeah.
Damien Schreurs: And then the last stage of grief is acceptance.
Joe Casabona: Yeah. Yeah. This is so, Suleyman, the guy who wrote The Coming Wave, founded DeepMind and tells this story in the book. So it's so funny that you're, we're both bringing up kind of the same thing.
But, yeah, it's incredible because if I recall correctly, it's not necessarily that the program, the AI, it's not like chess-playing computer program, right, where it was just, like, looking at all the possibilities and doing a probability play. It actually invented a new strategy. Is that right?
Damien Schreurs: Yes. Yeah. Yes.
Joe Casabona: Which is very, which is which is intelligent. Right? It's not just probability.
Damien Schreurs: Yeah. We could say the computer displayed creativity.
Joe Casabona: Yes. Right. It didn't just take what it was told to do and do it. It actively learned.
Damien Schreurs: Yes.
Joe Casabona: And right. So when we talk about the five stages of grief, you go through, no. This is impossible. Right?
Well, no. I mean, surely, this is just a fluke, and then eventually, it's alright. Well, I guess we're in a new frontier where computer programs are better at this game than me. And if we, so if we extrapolate this out to the larger community, we're probably seeing the same thing possibly in the creator space. Right? Because, and I think it was probably most prevalent with DALL E, right, with the image generation. Because you had a lot of artists saying, this can't do what I do. And that's I mean, with art, that's true. Right? It's not gonna come up with what you would come up with, but it comes up with stuff
Damien Schreurs: Mhmm.
Joe Casabona: That could be what a human could come up with. Right? It invents new pictures, and is could be creative about it. So that's, I think this is a really interesting thing to think about, and I think we see it with all technology, and this is but the ramifications feel bigger for this, I think.
Damien Schreurs: Yeah. Maybe my message for the listener, for the audience is, please don't get stuck in denial. Pass denial as quickly as possible. Don't get stuck in anger. Right? Because that's not gonna help.
And, I think what, the question, it's the, it, really AI is questioning what makes us human or what makes humans unique or different from computers. And for the very first time, we have to ask ourselves that question. Right? How am I different or unique? What can I do that a computer cannot do? Even as smart as possible.
So, yes, so these AI are faster than humans to process information. You cannot change that. Right? Computers will always be faster than us to process information. But, until now, the AI systems that I've seen have not been able to extract meaning from the processing that they have done. Right? They process information. They display the result of the processing. But we humans are still uniquely qualified at making some judgments, right, and deciding whether it's the right direction or the wrong direction or maybe sparking another idea. Right? At the moment, computers, AI computers, do not have ideas of their own.
Joe Casabona: Right.
Damien Schreurs: They seem it, the way they communicate with us seems that way, but don't be fooled by the fact that they speak our language.
Joe Casabona: Yeah. I think this is a really crucial difference, and this is something that I try to communicate to people all the time. AI is not coming up with new stuff. It's basically taking the average of everything it knows and spitting out the most likely information. Right? Like, if I ask it questions about my audience, it's not going off and surveying my audience, or it's not inventing new things that it thought of with my audience. It's taking everything it's learned about the audience I've told it about and then spitting out stuff it already knows kind of reconfigured.
Damien Schreurs: Yes. Yeah.
Joe Casabona: And so It's like buying a bunch of LEGO sets, dumping out all the pieces, and then building something new with it.
Damien Schreurs: Yeah. And maybe, maybe you could come up with an arrangement of LEGO blocks that we have never thought about or that we have never done ourselves, but it will still be the Lego blocks. Right?
Joe Casabona: Right.
Damien Schreurs: It cannot at one point think like maybe when you were a child, you say, okay, let's take these, toothpicks and make some arms and add them to the to our Lego figure. Right? And picking up stuff from left and right what what we can see in the environment, outside of the scope of the the Legos, to make something truly, truly unique.
And there's there's a big difference between knowledge, experience, and wisdom. Right? They have knowledge for sure, but they don't have experience. And it's only when you combine knowledge with experience that you can then go to the next step, which is having wisdom.
Joe Casabona: Yeah. What an incredibly succinct way to say that. Right? Because I think, you know, I'll link it in the show notes, but, the impetus for you reaching out was an episode, a friend, I'll say a friendly debate, I had with my friend, Alastair, about how we use AI. Right? And, I think we came down on different sides. Oh, man. I just called it great things require different, great things require time. Is that what I really called it? Anyway, I'll link to it in the show notes.
But, I think the main thing that stuck out for me in that debate was when he said, what if I train AI on everything I've ever written, on everything I think about a topic, and then I tell it to write a book on that topic, then will I have sufficiently written that book? And I said, I could give AI everything I've written today, and it would not know what I experienced tomorrow. It won't know what I experienced ten years ago that I wrote down that I didn't write down.
AI can't replace the human experience. And I think the way you put it, right, knowledge plus experience is wisdom. I think that's such a great way to drive home the point that. Sure. With quantum computing, we can we'll be able to tell the computer to write a book on AI, and it'll do it in a or, you know, or whatever on the, you know, the history of the Internet. And it'll be able to do it in a second. But it'll just be like a dry regurgitation of facts. Maybe not dry, but it'll be a regurgitation of facts.
Damien Schreurs: Mhmm. Yeah. I think well, can I go a little bit about the healthy debate you had last time?
Joe Casabona: Yes. Absolutely.
Damien Schreurs: Because, I don't know, maybe the listeners are experiencing that right now, but I'm when I'm listening to podcasts, especially interviews, I'm like, no, no, no. I want to jump in. I want to…
And I think there was one thing, one, maybe, one angle that was missing in that discussion, and it was the angle of what's the purpose of using AI. Right? Is it to convey, our thoughts and our thought process into the world? And in that case, I'm against using AI. Right? It makes no sense to use AI when we want to express our opinions or our way of seeing the world. But when I have recorded an episode of my planner and I want quickly a first draft for a summary and takeaways. Right? I'm okay to give it to an AI to speed things up.
And, yes, I will read. I will make some small edits left and there, left and right. But I have, I am not ashamed of using AI for that because the, first of all, the source content is my voice and my guest's voice. Right? It's content. It's coming from me. It's totally unique. And I just want a small cognitive task to be performed quicker.
Joe Casabona: So, I like that, and I think we're about to move into how we each use AI because I agree. I do the same thing. Right? I love the distinction you made. But there was something that I think you started to say that I really want the listener to hear. Right? We talked to you and said, get through, don't stay on anger. Don't stay in, depression. Right? Essentially, get to acceptance. Right?
And so I'd love you to elaborate on that a little bit more before we get into how I'm accepting AI in my everyday life.
Damien Schreurs: Yeah. And I want to preface what I'm going to say by the fact that using AI does not mean accepting. So you can be using AI and still be in a bargaining phase or a depression phase. Yes.
But then, for me, the acceptance is just realizing, right, that the fact that AI will be embedded at every stage of our technological life is inevitable. Right? It's the same way that when we had the Industrial Revolution. Right, you could jump up and down or strike and say, no. No machine will ever replace me in the factory. Do not use steam engines. Right? It's bad for the environment, blah blah.
The thing is that there are economic forces and demographic forces. There are forces completely outside of our control that push society, our human world to go in that direction. And I said earlier, I equate that to the invention of electricity. For me, it's imagine when electricity was invented, right? it was dangerous, right? We didn't have all the safety standards that we have right now.
Joe Casabona: Yep.
Damien Schreurs: A lot of jobs were eliminated, right? People who would, light lamps in the streets.
Joe Casabona: Mhmm. The lamplighter. Yeah.
Damien Schreurs: The lamplighter disappeared. But at the same time, there are so many jobs that were created. Thanks to electricity.
And nowadays, you could not imagine living without electricity, right? We would be paralyzed as a society, completely paralyzed. And the genie is out of the bottle, and it will be the same with AI. AI will be embedded. And we won't, I don't even think that we will talk about AI in the future as we talk about it today. We'll just talk about features of computer systems or even machines and robots and stuff like that, appliances that will have AI embedded because it makes our life easier or it makes the product better to use, actually.
Joe Casabona: Yeah. And this is I think, this is the the important thing to think about. Right? I came, you know, I came of age, I was born, I don't, maybe not before the cell phone was invented, but certainly before people, you know, consumers had them. Right?
I got my first cell phone in my freshman year of high school. I was 13, and all I could do was make calls. I couldn't even text on it. Right? So, but I remember my friend in eleventh grade, twelfth grade saying, I don't need a cell phone. Right? Like, so like, so again, why do you need a cell phone? Why are you so important that someone has to call you at any time during the day?
About a year later, he got a cell phone and had, like, a full-blown addiction to it. Like, wouldn't put it down. Him getting a girlfriend probably had something to do with it. But, like, I think about the early adopter versus the Luddite, right, on that technological adoption curve.
Damien Schreurs: Mhmm.
Joe Casabona: And, I think the, it's up to the early adopters to find the bugs and the kinks and educate the wider audience on the benefits and pitfalls of any new technology. Because you're right. There are economic forces that will ensure this. Right? Ensure, I think, is one of those AI trigger words now. But, we'll ensure that AI is everywhere. Right?
Damien Schreurs: Mhmm.
Joe Casabona: Just like you need the I mean, like, there are so many cashless venues in The United States now. Like, you can't go anywhere without a credit card or a debit card or your cell phone to pay for stuff.
And so I think understanding the technology early will help you evolve with it and safeguard against it, and this is probably why Suleyman wrote the book that I referenced earlier. But, I wanna get into how we're using it, how you and I are using it today because I still follow your convention. I don't, if I'm trying to convey my thoughts and feelings on something, I don't use AI to write it for me, because how could it know? I could tell it. Here's how I feel about something. Write an article. Like, that's that's gonna be so obviously fake or so obviously AI-generated. But I have been using it a lot more. So, I'm curious. I've been talking for a long time. I'll let you go first. What are some of the ways that you're using AI today?
Damien Schreurs: So, I use it for ideation. I use it as a, I've created, so I paid $20 a month for the ChatGPT Plus account.
Joe Casabona: Mhmm.
Damien Schreurs: I have created a few custom GPTs, and one of them is called my co-CEO. So, it's a, can I say a strategy, a thought partner. Okay. Kind of. Kind of. Yeah. Just to give credit to people. It's, I discovered this co-CEO concept from the host of the podcast, the AI marketing podcast, or it was used to be AI marketing.
Damien Schreurs: I'm gonna double-check now.
Joe Casabona: AI-driven marketer, artificial intelligence show. Oh, man. There's so many.
Damien Schreurs: The artificial intelligence show with Mike Kaput and the main host. And he's the main host.
Joe Casabona: Oh, Paul Ritzer.
Damien Schreurs: Paul Ritzer. Yeah.
Joe Casabona: Paul Ritzer. Ritzer. Okay.
Damien Schreurs: So Paul Ritzer has come up with this idea of the co-CEO. He even made organized, a webinar, which I attended, and he has published on SmarterX, his website dedicated to education for about AI.
He actually published and explained exactly how to create that co-CEO. And so, basically, I told that in custom GPT, I fed it a prompt that an instruction that was based on what I've seen from the webinar. But I adapted to, obviously, my company is EasyTech. I'm providing IT training and coaching in Luxembourg and abroad. And so I fed it a bunch of information about my company, about myself, my dreams, my aspirations, but also, the podcast, the Mark Barner podcast, who's the ideal audience, and stuff like that.
And as well as some instructions, how can it help me. And then you can attach documents. So I fed it also a bunch of documents. So for the financial side of easy tech, no absolute values, but percentages. Right? Percentages by, customer segments and stuff like that.
Joe Casabona: Nice.
Damien Schreurs: But it's pretty good at helping me, right, think about either issues or strategies. Right? And, so that's one way that I use it. It's more like helping me come up with ideas or strategies.
And then we have the typical thing for post-processing text. Right? You give it, so it's because it's called the Large Language Model, it's perfect for using text and processing text. So, I feed it the transcripts, and then it spits me a summary, and takeaways, something that I've been doing as well because, the Macpreneur Podcast, the guests, they share the tools that they are using. So it's full of tools and software and devices. And so now I give it that as a, please give me a list of all the tools that have been mentioned.
Joe Casabona: Yeah.
Damien Schreurs: And that's pretty good. I really like that feature. And, and so it's more text processing kind of things. I've doubled with image generation, but at this stage, it's more, it's more for giving me concepts for images or thumbnails or stuff like that, but I will do them afterward by myself.
And then something that I started doing last week, so it's very, very fresh, it's embarking into the AI agent bandwagon. And when I say AI agent, I really mean multi-step processes.
Joe Casabona: Yeah.
Damien Schreurs: And, the tool that I've been experimented with is called Mind Studio. I don't know if you've heard about that or..
Joe Casabona: No. I haven't.
Damien Schreurs: MindStudio. And, you need to be a little bit comfortable with coding or programming concepts. But it allows me at at the moment, it helps me with two sub-workflows of my podcast editing process. So, the one thing that took me 10-15 minutes was to proofread the subtitles and the caption. So I used Descript to edit a podcast. It spits out an SRT file.
My episodes are a bit like yours, like thirty minutes to forty minutes. So a lot of captions. If you feed that to Chargegpt, it will choke. It's too much, right? It's bigger than what is called a context window. How many tokens can it output? The token window for the input is huge, but for the output, it's still limited. And I realized that to proofread them manually, I needed to do roughly 100 or 120 maximum at a time.
Joe Casabona: Okay.
Damien Schreurs: Copy them. So select, copy, go to ChatGPT, and proofread, I have my text expander snippet. With the prompt, it takes, it was taking one mini a few minutes to do it, then copy the result, paste it back into the transcript, and go to the next one.
And with Mind Studio, I've basically automated that process. And the way Mind Studio works is you enter a prompt of what you would like the agent to do. This process, when it needs to work on its own, it also takes ten, sometimes fifteen minutes. But I can do something else. Right? I can work on another part of my process, and it costs me 20¢ because you pay by the tokens that you use.
And so it was ridiculous, the when I say, I want to make one day a newsletter, with the title, an outrageous title, like I build my first AI worker and its wage is ridiculous, it's like less than a dollar per hour wage. And it's I yeah. It's great.
Joe Casabona: Yeah. And so this is something I wanna say. Right? Because, like, I thought I never I didn't get into the ChatGPT API for a while because I thought, oh, the things I wanna use it for, it's just gonna be, like, more than I wanna pay. It is obnoxiously cheap, like, obnoxiously cheap.
Damien Schreurs: Mhmm.
Joe Casabona: So, yeah, I'm like you know, I'm building I'm working on, like, some Zaps and Zapier to use it and and stuff like that. I think that's a really powerful thing to use it for. Right? And one of the things that you said is probably the main ways I'm using it.
And I know we're coming up on time here, but I wanna kinda talk through this a little bit. Right? Because you mentioned custom GPTs. I haven't gone and built one yet, but I have found that the projects feature in ChatGPT has been pretty good for me because a project is a folder if you don't know, and then you can give it instructions and custom prompts. You can't as far as I can tell, the main difference between, like, a project and a custom GPT is you can't attach files to it, and you can't make a project publicly available. Or am I incorrect in thinking that?
Damien Schreurs: You can attach documents. So that's or maybe it has changed compared to the first time. But, yeah, it's true. I never looked at I didn't look at projects until recently. And I realized that nowadays, the only difference is you can't share a project. You can share an individual discussion that you might have, but not the whole project. And, what I realized, is I've not not done it yet because I've always been on, let's do a custom GPT. Let's do a custom GPT, even if it's a fully private one. And then I realized, oh, but I could do that in projects. It's easier and so on.
But then I realized, yeah, but if one day I want to if I do a great project and I realized maybe I could share it with my audience, right, and have it maybe as a lead magnet, I can't. I haven't found a way yet to convert a project into a custom GPT, but it shouldn't be too difficult to do.
Joe Casabona: Yeah. Yeah. So that's a really great point. I didn't even realize that I must have had, like, blinders on or I just haven't looked. Sometimes there are discrepancies between the desktop app.
Damien Schreurs: Yes.
Joe Casabona: Yeah. So okay. , if I'm looking in the desktop app right now, I'm not seeing a way to attach files. So, yeah. But if I go to the web app, I'm seeing it.
So. yeah. Perfect. So I use mostly projects. I haven't looked into custom GPTs yet, but, like, say, exactly what you said. If I wanna have a lead magnet or something really interesting, that I wanna share, like, it would have to be a custom GPT.
But, the main thing I'm using it for are proofreading my work. Right? So I have a project called my writing. I'm gonna upload some samples now of writing I thought was particularly good, but, you know, the instructions are like, you know, these are the things I wanna know. These are the things I say, like, when I ask for feedback, this is what I expect.
Damien Schreurs: Mhmm.
Joe Casabona: And so now I just say, like, give me feedback on this. And, like, it uses the instructions. I think that's great.
In the custom project for or the project for this, it's very long because I tell it everything about the show and who the target audience is, and how I format the show notes. And then the prompt is, hey. Here's the latest transcript of the latest episode. Me and Damien talked about AI. Give me an episode description and the top takeaways. And then, like, what links did I promise? Right? Or, like, what should I put in the show notes?
And it's been, like, a game changer. Like, that the fact that ChatGPT can, like, take a transcript and kind of understand, oh, yeah. You probably wanna link to this in the show notes. It has made my life so much easier.
Damien Schreurs: Mhmm.
Joe Casabona: Because I was going through and telling my VA to find the links. She's in the Philippines, and that's, like, a very unclear thing, especially if it's, like, a jargony tool. And so then I'd have to go check the links anyway. Now, I don't really need to do that, which is really nice.
Damien Schreurs: Mhmm. Yeah. And, yeah, with what I like with Mind Studio, and I've been positively, actually, impressed by that, is that if you or maybe the listeners, if you are familiar with make.com or Zapier, it's very similar. So you have different steps, different boxes. Every box is an action. It could be an AI action. And whenever it's an AI action, you have to give it a prompt. The prompt can use parameters or, yeah, parameters that come from earlier steps. So, the same way that maybe come work or Zapier work.
And what's even really cool is for every individual AI step, you can choose which model to use. And you can pick from Claude, you can pick from OpenAI, you can pick from DeepSeek, you can pick from, Google. So, the Gemini, Jemma, those from Meta as well, the Llama models, and so on.
And, something that I realized, so the process, I've made a second, AI worker, which I would call for my podcast, and that's what I was talking about. So I feed it now the transcript, not the captions, but the transcript. It proofreads it and then creates a summary, and takeaways, and then show notes for the podcast, YouTube description, Instagram and Facebook post, and a LinkedIn post, all that one after the other. And I fine-tune which model to use for which step, and, Claude's 3.7 is Mhmm. Definitely the best for the summary and takeaways.
Okay. So much so that I copy pasted the output of Claude or from that AI, worker. Right? I copied the output, and I put it in an AI detector, and it told me 1% probability of AI. It thought it was human. I almost do not touch those summaries, but I also refine the prompt. I have a very specific prompt that makes the output very human like with a lot of bullets and numbered lists and stuff like that. But, yeah. Really impressive.
Joe Casabona: Really impressive. That's cool. And so Mind Studio, I'll link that in the show notes. I'm looking at the pricing, and it's $0 per month plus usage costs. So, like…
Damien Schreurs: Exactly. And you get $5 credits. So you can start right away, you can start experimenting for free. And I still have $2 left, I think, or $1 left. Now, If you don't put a credit card, you have limited access to the different models. So if you want to be able to experiment with all the models, and all the LLMs that are available, then you just put your credit card. You will not be charged unless you go below the credit limit that you have put in.
Joe Casabona: Gotcha. Yeah. This looks real I'm gonna have to check this out. This looks really interesting.
Well, this has been great. Thank you so much for coming on and sharing and talking. I'm glad we were, like, a little philosophical in the beginning, and then we got to the practical, which is great.
Damien Schreurs: You mean in the weeds?
Joe Casabona: In the weeds.
Damien Schreurs Yeah. Exactly.
Joe Casabona: That's alright. Damien, if people want to know more about you, where can they find you?
Damien Schreurs: Yeah. They can go to macroner.com, and they will find everything about the podcast, and how I help solopreneurs who cannot imagine running their business on anything else than a Mac.
Joe Casabona: Yes.
Damien Schreurs: Solopreneur. I call it Mac.
Joe Casabona: Yeah. Love that. And, again, I'll link to our discussion of that episode too because that was a lot of fun. I don't usually get to talk about my Mac setup, so I was really excited when you invited me on. This was great. Can you, can you say your name properly in your accent for us so that people know I tried really hard to say it the right way.
Damien Schreurs: So I'm Damien Schreurs.
Joe Casabona: Damien Schreurs. Oh, I can't get, I, you know, I can roll my r's because of the Italian accent, but I gotta get the throat, you know, a little bit. ‘
But, Damien, thank you so much for joining us today. I really appreciate it.
Damien Schreurs: Yeah. My pleasure. Thank you for being here.
Joe Casabona: Thanks to everybody listening. You can find this is gonna be rich show notes, so check the description wherever you're listening. You can find everything over at [streamlined.fm] as well. If you like this episode, let me know. Send over some feedback. Same link, [streamlined.fm]. Link in the description.
And until next time, I hope you find some space in your week.