The AI Wrapper Gold Rush: Opportunity or Bubble?

Wiktor Żołnowski: Welcome to the next episode of Pragmatic Talks. Today with me is Dariusz Tarasek, with whom we are going to discuss AI wrappers–not rappers, wrappers. The new big thing in tech? Or maybe this is something that is just another bubble? Stay with us and listen to this episode, so you will learn a couple of things about how to make money in a simple way in a very short time, just by building an app on top of some AI models. You can build an AI wrapper in a week to solve a very tiny problem, but how to make it successful? We’ll discuss today in this podcast. We talk to founders and experts to share real stories and lessons from building and scaling digital products and companies. Pragmatic Talks is for those who want to understand how digital products are really built and grown. No fluff, no buzzwords–just honest conversations. So, let’s start with something that is very, very simple. Let’s try to explain what is an AI wrapper.

What is an AI wrapper?

Dariusz Tarasek: Well, an AI wrapper is an app that gives a user a better experience than just pure AI. It’s built around an existing AI model, like let’s say ChatGPT, but solves a specific problem for a user, right? So, it’s an existing AI technology behind and a completely new layer over it. It’s basically something very, very simple on top of some AI model or models, depends, right? We’re probably going to talk today about multiple examples of AI wrappers. A good question is: do you use an AI wrapper? Yes, I do. I actually didn’t know that there is a name for that, but for a long period of time, I’ve been using, I think, two apps that are some kind of AI wrappers. One is the app that is helping me take care of my plants at home. It’s a very simple app that, first of all, uses AI for recognition, like recognizing the plant on the picture, and then it provides the entire plan on how to take care of this plant. So, this is one thing. And actually, this is one of not so many apps that I’m paying a subscription for, so I’m the customer or the victim of AI wrappers–we’ll decide it later on. And the second app that I’m using pretty often, or I used to use pretty often, was the mushroom recognition app. So, a pretty similar thing, where I was using an app when I was collecting mushrooms just to check if this is something that will kill me or maybe not necessarily.

Wiktor Żołnowski: That’s really cool. Wrappers, I would say. And you didn’t feel like a scam after you downloaded it and you realize today that this is just, you know, a wrapper, not like, you know, brand-new technology?

Dariusz Tarasek: So, with the second app, the mushroom recognizer, I’ve stopped using it at some point since I’ve started simply using Google Photos for that, and I felt that it’s way more accurate than this particular wrapper. So, using Google photo search or search by photo, which works pretty well for mushroom recognition. Still, I don’t recommend that. I think that first of all, you need to learn how to recognize mushrooms. I’m only using it when I’m not 100% sure, just to confirm that I know what I’m collecting or what I’m leaving in the forest.

Wiktor Żołnowski: Yeah, so let’s just sum it up, maybe, right? So, for example, in terms of the apps that you use, you make a photo, and it’s being sent apparently to a different LLM that analyzes it and brings back the answer, right? So you don’t need to use any prompts, you don’t need to know any technology behind it; you just simply make a photo, click send, right?

Dariusz Tarasek: Exactly. So in the second case that I mentioned to you, like the app for taking care of plants at home, it works differently because it’s not only providing me the name of the plant, but it also provides me the entire care plan for the plant. And it also implements it as a schedule in the app, so it shows me the reminders when I should put water there or when I should take care of the plant. My plants are right now pretty awesome after a couple of even years of using this app. But yeah, it’s not just a prompt. Probably I could do the same with ChatGPT or something else, but I wouldn’t have these functionalities of reminders, etc. And now I have all my plants in one place.

Wiktor Żołnowski: Yeah, so convenience, right? Yeah. And basically, your laziness to use it instead of prompting ChatGPT has won, and you paid for it. Yeah.

Why have AI wrappers become so popular?

Wiktor Żołnowski: So, why do you think, you know, like there’s a bunch of other examples that we are going through today, for sure, but why do you think AI wrappers became so popular right now? Since people can use ChatGPT or Gemini or Claude or something else to do exactly the same as they do with the AI wrappers, but still, those wrappers are pretty popular. A bunch of people are creating them, and they are creating them usually pretty fast, in a matter of days or weeks. They’re publishing the app, and some of them are making a shitload of money, yeah, like hundreds of thousands or tens of thousands per month. I think that the biggest are right now like 10 million per year or something like this. Yeah, crazy, right?

Dariusz Tarasek: Well, I believe people who developed this, they just wanted to create something on their own. And since AI became a thing and somebody realized that you don’t need to build the whole technology from scratch, you just need to solve a specific problem, it became much easier these days to basically build a product around it and start selling it. Yeah. So, I believe this growth of popularity of AI wrappers is at its beginning, right? Because we will see probably more and more tools that will enable you to ship wrappers even faster. Right? Today we’re going to talk about a few of them, but in general, I believe some of the people who create this, I believe, may fall into this get-rich-quick scam, as maybe drop shipping used to be. I’ve seen so many ads at the time of courses of people who were trying to sell me the idea that I don’t need to own the products, I don’t need to store them, I just need to organize the shipment from the vendor and basically build the front-end layer, the marketing around it, and sell it. An AI wrapper, in theory, can work similarly. You don’t own the technology, you just utilize it. And basically, you need to be a good salesman to provide it to the users who need it.

Wiktor Żołnowski: You’re a product manager with a lot of experience, years of experience. Don’t you feel or think that this sounds kind of like cheating?

Dariusz Tarasek: Yeah, it sounds like cheating, but if people like to use it and you are one of them, why not, right? The big difference between, I would say, a good AI wrapper and a bad AI wrapper is the question of if it really solves a problem. If there is a friction–let’s say I hate prompting ChatGPT for a certain purpose, so I would really like to click one button and have the same effect, right? This is a good example. The bad example is when you basically mitigate an app or copy it and are selling a scam, right? And you don’t even mention that you are using the AI. You can rebrand the tool and start selling it with a payment wall to the same people, trying to kind of base it on the fact that they may be naive, or maybe they are not aware that there are existing apps or just ChatGPT that can do the same, right? So probably, this is a thin line between those two, but in general, like why not? If you bring value to the table, why not? I think it’s not cheating then.

The future of the AI wrapper market

Wiktor Żołnowski: Okay, so it also fits into my theory that building products is becoming easier than ever right now with the tools that we have. Okay, so what will be the future of it? Yeah, like don’t you think that simply OpenAI or Google or Meta or Anthropic, whoever is owning the technology that is behind it, will sooner or later start building their own, maybe not a wrapper, but native solutions that are solving these small problems that users have?

Dariusz Tarasek: Yeah, that’s a good question because from that perspective, you can assume that the market for each of those solutions is getting smaller and smaller because all big LLM models basically start providing more features. The good example is that ChatGPT at the beginning didn’t have a functionality to upload your PDF file and ask questions and get the information from it. So, there were created multiple companies that basically provided you a solution to upload your PDF and ask questions, right? The good example is pdfchat.com, and I’ve been using it as well at the beginning. And apparently, they made 10 million dollars a year, yeah, on just providing this functionality. And the funny thing is, people still use it because probably it’s convenient for them. And I think they’ve built on top of it more features, and they basically became a platform instead of just being a wrapper. So to answer your question, what’s the future of them? I think it’s a good way to maybe build your proof of concept, the MVP, right? Ship fast something that solves a very specific problem. But if you want to start building something more advanced, something that will become a platform with a wide range of use, then you probably will disappear.

Wiktor Żołnowski: So, it’s pretty similar to the model that, let’s say, AWS has. Yeah, like they are providing a whole different infrastructure for building websites, web apps, etc., but they are not building so many apps on their own. So they are not trying to replace the internet, but they are trying to provide the solutions for other builders, provide the infrastructure. And they’re making a lot of money on that. And the same is with the AI models, say, like OpenAI or Google or someone else who’s also providing the infrastructure for other people to use it and to, of course, pay for it to the proprietary owner of this technology. And in terms of the future and in terms of the usage or the market for this kind of app, I’ve read somewhere that for now, it’s like the end of 2023–like October, end of October right now–there’s less than 10% of people who are using AI from time to time. So, still more than 90% of people, among whom there is like, let’s say, 70% who are using the internet, they are not using AI at all. So for those people, these kind of solutions might be some kind of the best way to start using AI, even if they are not aware that the AI is somewhere there behind it.

Dariusz Tarasek: Yeah, yeah. Like on the top of my head, I see many features that my mom would like to use because she’s not very fluent with prompting ChatGPT, right? So I can build wrappers for her for, I don’t know, doing shopping, maybe other things as well. So yeah, at the beginning when the internet became a thing, less than 1% of the population in the world were using it, right? Right now, it’s totally different. Today, only about 11% of people worldwide are using AI and are aware of its capabilities, right? That gives us around 89% of the population of the world that still don’t know about AI as much as we do, for example. So from our perspective, it may look silly, right? Why not use ChatGPT? But for many certain cases, people would love to have something simple that does literally one job.

Wiktor Żołnowski: Yeah, and it’s also about the prompting that you mentioned. Yeah, that prompting is a skill; it’s still a skill. Yeah. Like if you want to have great results with AI, you need to be good at prompting or prompt engineering. And people are lazy and they don’t want to learn new things. So whenever we provide them solutions that are a shortcut to achieve the same goal, they will choose the shortcut. They would even pay a few dollars for that shortcut instead of just making this effort and learning new stuff.

Examples of successful AI wrappers

Wiktor Żołnowski: And talking about paying a few dollars, let’s see some examples of successful AI wrappers. So, there is a company that made $6 million only by wrapping an existing AI, and their niche is student essays, right? So you have support to write a beautiful essay. You can do literally the same with ChatGPT, but I guess they have a more sophisticated AI agent behind it that is trained and provides better results, right?

Dariusz Tarasek: Let’s take another example. Tinder replies. You know, there are so many people who use dating apps, they don’t know what to respond to or how to be successful when responding. Yes, exactly. You know, take a look at this: $270,000 a year made on the AI that literally is just giving you an answer to the conversation. I didn’t try it by myself, but I don’t need it. But so, you know, those are just a few examples, right? Take a look at this: Cal AI, a calorie tracker that is literally built on existing GPT, makes 12 million a year. Re:GPT, a dating coach built on top of OpenAI GPT, makes 1 million a year, right? You know, these are just a few of the wave of GPT wrappers. And if you are struggling with building a new, innovative technology, let’s just ask ourselves a question if it’s still worth it to try to build something innovative, right? If you can see such companies, they literally make millions by wrapping AI. What do you think?

Wiktor Żołnowski: Yeah, like this is something that is a huge opportunity right now, but I am not sure if this is something that will stay with us for long. Still, I think that there will be a bunch of products that will be popping up like this from time to time, but the number of problems that could be solved that way, I think it’s… and I believe that at some point, people will realize that those are just the wrappers and they are not worth being paid for. Or maybe there will be a lot of competitors, and the prices will be going down and down to the point where it will not be profitable anymore. And at some point, the entire bubble will pop up or collapse.

Dariusz Tarasek: Yeah, that’s the thing. When I search on forums like Reddit and stuff, you can see so many new ideas, so many new AI wrappers. So it looks like there are actually more people who are building wrappers than people who are literally using them and paying for them, right? So if you can build something literally in 24 hours, the competition will do exactly the same, right?

The business of building and selling wrappers

Wiktor Żołnowski: Yeah, so the biggest challenge is not to build it, it’s to sell it, to ship it to people, right? And to have a distribution, to ship it to as many people as possible, to convince as many people as possible to actually pay you for that. And I think that a good sign of the bubble, that this is probably some kind of bubble, or maybe not for now, it’s a huge opportunity, but I believe that at some point it will become a sort of bubble. That’s my opinion; I can be wrong, of course. I believe that the good sign for that is that there are a bunch of online courses on how to build your own AI wrapper. There are some tools that are allowing you to build your own wrapper in a few hours or something like this, and to ship it and to start selling it, which is actually showing that those products could be pretty easily copied by others, at least the basic one, the most simple one. Some of the advanced products, as you mentioned, the products that offer something that is way beyond just the AI wrapping, just adding some default prompts to some AI model in the background, like platforms that are built on the top of these wrappers–that might be something that will stay with us for longer, and that will be, some later, a pretty nice product.

Dariusz Tarasek: Yeah, but again, API cost is zero, the distribution cost is everything, right? Stop coding, start selling. That’s probably the thing we should be doing. And if you’ve built something in a weekend, now how do you get paid for it? And how to check actually if people want it, right? Any good ideas on that? How to check, or maybe how do you check what people want before you start building? Do you have any ideas on that?

Wiktor Żołnowski: Yeah, you know, it’s probably the matter of building something that is even simpler than an AI wrapper and starting to sell it before we build it, right? It’s like a landing page. Yeah, you can create a landing page in one of those tools that helps you build–also those are sometimes AI wrappers who help you build a website–and start advertising. You know, spend $20 on Meta ads, try different formats and see if at least a few people will register, right? If yes, that might mean that, yeah, you should go ahead. If not, probably you should kill it as soon as possible and take care of something more important in life. I’ve seen some other methods that someone was showing, that you can actually check the keywords that people are looking for. And you can try with something like “AI for recognizing mushrooms,” as I mentioned. Yeah, if this is something that has a lot of traffic, a lot of searches in Google, for example, and the keyword difficulty is pretty low, that means that there are not so many competitors and probably this is the niche that you can address. From the other hand, I think that no one who is building a wrapper is counting on building a product that will become a unicorn. I’m pretty sure that those people are pretty aware that they are building small businesses that will make some money in some cases if they will be successful and if they will have good promotion or some networking effect or some viral effect or whatever. They will make a lot of money, but still, it will never be the product or the company that will be worth $1 billion or something like this.

Dariusz Tarasek: Yeah, probably. But did you know that you actually can buy an existing AI wrapper? Is there any marketplace for that or what?

Wiktor Żołnowski: Yeah, there are a few marketplaces. Actually, those marketplaces help you sell any kind of business, but these days you can find a lot of AI wrappers there. The prices there go from a few thousand to even $2 million, and it all depends if you have already an attraction, if you have SEO on the right level so people can find the website easily, or if it is just code behind it. Probably if it’s just code, it might not be worth it to buy it, unless you really believe that you can do something better with it. That’s really interesting.

Differentiating between good and bad wrappers

Wiktor Żołnowski: So, let’s talk about the difference between a good AI wrapper and a bad AI wrapper. So what is the example of a wrapper that we can honestly call a scam?

Dariusz Tarasek: You might have heard about some scams that basically build an app that looks exactly the same as an existing one. There was a case about Signal apps, for example. Okay. For some reason, the Google Play market allowed them to publish such apps. The difference was in the naming; it was “Death Signal,” not “Signal,” for example. Okay. And they were advertising it, and many people basically registered there, and they were clicking the ads, they were making money on them, and they stole their data. You have to always be prepared that something might be a scam. It’s not easy to differentiate. Do you have any ideas, maybe, how to differentiate the bad application from the good? Do you have one?

Wiktor Żołnowski: Yeah, I think that all of the good wrappers are providing this extra value. They are not just a proxy to ChatGPT or something else behind it. It just provides you more value. And I think that the lowest level of the value is just the structured answer and simplified questions. Yeah. So when you can add a simple question, or even just add a picture and the question is already there behind it somewhere, you don’t see it, but you add the picture, upload it, push the button, you get this structured answer, then there is some value on the top of the AI itself. But if this is just, you know, you add a picture–I’ve even seen some video that someone was showing how to build an AI wrapper, I think it was also for some pet recognition, I believe, or some bug recognition, something like this–he showed that you can just simply use the basic API to ChatGPT. When you’re sending the picture, like in the app, the user is taking the picture, uploading it, and then it’s sent through the API to ChatGPT with some very basic prompt, and the answer as it is from ChatGPT is directly presented in the user’s app, so without any structuring or anything. And I believe this is something that I would still call a scam. Yeah, like this is not enough. This is something that if I would take the same picture and put it into GPT and copy-paste some kind of prompt like, “what is this bug here,” then I will get exactly the same answer. But if this answer would be structured some way by this app before it will be presented to the user, like it will be simplified, or there will be the prompt that will show what the category of bugs this is, or there will be some extra questions about it that will be asked to the GPT by default and that will be presented to the user, then there will be some kind of value added on top of that.

Dariusz Tarasek: Yeah, and I would add to this, maybe, something that is integrated with other apps or tools you use, right? For example, an email writer, right? You need to write a friendly email, you would like some help. Of course, you can ask ChatGPT to do this, but in the end, you have to copy and paste it into the email, right? If you have a plug-in that can do this for you in your mail client, it’s added value, right? So it’s not only better UX, it’s also workflow integration. It’s a very specific, I would say, set of prompts that probably would take you a lot of time to figure it out by yourself, right? But the example of the bad wrapper, let’s say about this, right? So this is just something that rebrands the existing AI with a paywall. Did you ever find maybe such an example?

Wiktor Żołnowski: I do not have anything like this in my mind right now, but I can imagine that someone would create, let’s say, as you mentioned before, the fake Signal app. They may create a fake ChatGPT that they will call GPT Chat, put it into the App Store or Play Store and ask people to pay for it, like a few dollars or something, and then pay a subscription that would be higher than the GPT subscription. So yeah, I can imagine that it’s extremely simple to be done and people can fake it. If not with the Google Play or Google Play Store or App Store, there are other sources, other app stores than those two, that could be used for that and that are not so restricted in terms of uploading or publishing the content there.

Dariusz Tarasek: Okay, but what do you think about the platforms that offer you image or video generation and actually use models that sometimes are not available yet in certain regions of the world, right? For example, Artlist.io uses Sora 2. It hasn’t been available in many countries before, but if you paid for it through their paywall, you could use it. I believe it was Italy at some point, where ChatGPT was banned at the beginning, and I don’t know what the status is right now, but I remember the numbers that showed how much money VPNs were making in Italy were like skyrocketing in that month. So people showed that there is a demand for that, and they were ready to pay for a VPN to actually use the same technology that was available in other countries around. But again, this kind of solution has a probably very short period of life because sooner or later it will change, right? And you’ll be able to use it directly at the source.

The short lifespan of many wrappers

Wiktor Żołnowski: The lifetime of these products in most cases will be or is already pretty short for one reason: because some of the wrappers are just allowing access to some tools that are not available in some regions. And as soon as they will be available in this region, those wrappers may not be as useful as they used to be. Another thing is the competition. Since it’s pretty easy to be built, there will be more and more people, more and more companies that will be investing into it and building more and more wrappers. So, even if someone has the traction at the beginning, their users will migrate to someone who is offering better prices or a better product. So that’s something that will shorten the lifetime of this kind of platform. Also, the progress of the models behind those platforms are also implementing it inside them, like for example, picture generation or video generation or some other things. Even for us right now, it’s obvious that you can upload the picture to ChatGPT and you can ask the chat what is there, like multimodality. I remember like a year ago I was recording the episode here when we were discussing that multimodality is something like, “Oh, it will be awesome, it will be soon here, it will be great.” But for now, it’s like, it’s obvious that you can do this. Yeah.

Dariusz Tarasek: Yeah, you know, there are over 35,000 AI wrapper companies around the world right now.

Wiktor Żołnowski: Not so many. I thought… not so many. Yeah, but hundreds of them are being created on a probably weekly basis.

Finding unsolved problems for new wrappers

Wiktor Żołnowski: The good question is: are there still any problems, any ideas that haven’t been really addressed by wrappers? If there are only like 35,000, then I think still there is like a bunch of problems that could be solved by this kind of solution. I’m not sure if all of the problems that could be solved are worth being solved and are worth being paid for. But still, like I think that our imagination is the only limit right now for anyone like us who are using AI on a daily basis. Like whatever you are using ChatGPT for or Gemini, let’s say. I recently use a lot of GPT for talking about some legal documentation that I had to sign or I just needed to understand or I just needed to check what’s in there and what I should pay attention to the most. Even building a wrapper that will be doing exactly the same, that will allow a user to upload the document and will point out the most important things that the user should pay attention to, like with some quotes and other stuff. This is actually pretty simple prompting to the AI that could be built into this kind of wrapper, and that could be a huge thing for many, many users. Like, still remember, only 9 or 10% of people are using AI nowadays and 90% of internet users are not using AI at all. So it’s like a huge number of people, like a few billion people who are not aware of the AI capabilities, who have these exact problems that you are solving with AI. So whatever problem you are solving with AI, that might be an opportunity to build an AI wrapper for people who are not using AI but do have exactly the same problem.

Dariusz Tarasek: Yeah, but I believe the key thing is the question if the users are already searching for the solution, right? They might not be using AI for it, but they are searching for the solution, and you can find it out. And the question is if they will be willing to pay a fee, if the problem they are dealing with is painful enough to pay for it, to pay for a solution, right? And if that can’t be solved with very simple prompt engineering in general GPT chats, right?

Wiktor Żołnowski: I wouldn’t completely agree with the last one. As I said, still 90% of people are not using GPT, and they’re not prompting, they are not aware what prompting is. So, like, a five, six, seven billion people are your potential user base for anything that you are using prompting for. Yes. So still, at least for some time, that might be an opportunity to make some money. Yeah, to make some money now, right? But if you think about a longer future, you need to have a plan in place how to transform this simple AI wrapper into something bigger, into a platform or maybe some different kind of service that will help people differently than only using the, utilizing the capabilities of existing AI. Or you just can build another wrapper. Yeah, that can be another way of life, right? Just building wrappers, at least for some time. Yeah. I’m pretty sure that this gold rush will sooner or later end. Yeah. But for the next couple of months or years–rather months than years–I believe that there’s a lot of opportunity for people to make money on very simple software that is utilizing AI in the background.

Security, quality, and investment potential

Dariusz Tarasek: You know, one thing actually scares me, like the question is: so if we won’t end up with just AI wrappers everywhere, so people won’t be building something really innovative, if they will be just using a wrapper around a wrapper built-in a wrapper? So what about the quality of features, the security as well, and stuff that is around AI wrappers? What do you think?

Wiktor Żołnowski: Quality is one thing, and security is another thing. Like, I’m pretty sure that most of the apps that are out there are not secure enough, especially those apps that people are uploading their documents, PDFs, other stuff that very often are things that shouldn’t leak anywhere. That’s scary, that’s crazy. So for you as a user, please like consider it a few times before you upload any documents to any kind of wrapper, especially when you don’t know the company that is behind it, or there is no company behind it, etc. Even if there is and the company is registered somewhere in a country that you cannot simply reach with your lawyer, it’s probably not the best tool to use. So from a security reason, definitely, this is something that is a huge risk for everyone. In terms of the quality, because of the competition on this market, if the quality is not the issue right now, is not the main thing right now, it soon will be. Like when there will be–and maybe this is another opportunity–if you see any AI wrapper that is there and has a lot of downloads and it’s like a paid app and it’s not working well, and you think that you can create it easily on your own and create it better, if you can do this in a week or two, just do it. Just do it. And you know, even start promoting yourself as the alternative for XYZ and try to promote yourself based on the low quality of the other solution. That’s a healthy competition, I believe. Another thing that came to my mind is the localization. Yeah, there are like a bunch of things that are, let’s say, only available in English, and if you are living in another country, even if you know a bunch of users are using such an app, but they’re using it only in English, there’s a lot of people who don’t know this language in this country who can use exactly the same app but translated or just rebuilt in their native language or an app that will be localized, like, for example, to the local regulations or, I don’t know, local culture, whatever is there that you can localize to.

Dariusz Tarasek: Besides that what you said, like different language, different specific set of features that your AI wrapper potentially can have, I’m wondering what can really differentiate you on the market, right? Because if you build something that doesn’t have technology that you can patent or, you know, it’s so sophisticated nobody can copy it, how you can potentially raise money from a VC or find an investment if you don’t have, potentially, you don’t have anything that adds value to the tech world, let’s say, somehow?

Wiktor Żołnowski: So, I’m basically also an angel investor, and I’m from time to time investing in startups, so maybe I will answer what I would consider as a good investment in terms of an AI wrapper. And before I answer that, I would say that like 99% of this kind of solutions probably is not a good investment idea because their lifespan is too short. Like only the ideas that have potential to be long-term successful products, which are, let’s say, those platforms that are utilizing AI, so they are way more complicated than just simple AI wrappers. Those platforms have some potential to be investable by the VCs or by angel investors or anyone else. Technology is not an advantage. The only thing that is an advantage is their user base and the user engagement. So if this kind of startup has a lot of users, the users are very engaged into the platform, they’re using it on a daily basis, then this kind of startup could monetize that user base. And by user base, I am right now thinking about hundreds of thousands, if not millions, of users. Anything below that is not a traction, it’s not something that is attractive to, especially to big businesses. For angel investors, maybe at some early stage, but only if there is a proven traction and if there is the growth strategy that makes sense and that is already proved that with some extra money for marketing will actually grow the user base. Actually, it’s not only about the current revenue, let’s say, and the current traction, but also the potential to grow. Yeah, like I think it’s like in any other rising, like in any other startup, not only the wrappers. Like this is the crucial metric that investors are looking at.

Dariusz Tarasek: Do you remember those apps that were popular for a very short period of time, and they literally had like a kind of AI filter that, let’s say, changed your photo into the comic version of it, stuff like that, right? I believe they don’t exist anymore, right?

Wiktor Żołnowski: Probably they do. I think that you think that they are not existing anymore because you have like a new smartphone, and your smartphone has it already built into it. Mine too, but I’m pretty sure that there are people who are still using it because they are using some older phones or some, let’s say, cheaper phones, and that software does not have this kind of solution. So those are there, but not in the mainstream as they used to be. I bet they still have millions of users base.

Dariusz Tarasek: Yeah, we don’t know what they do with this data. That’s crazy, that’s another thing.

Wiktor Żołnowski: Yeah, like and as I mentioned before in terms of the security, like when we are talking about security, whenever you’re using this kind of app, like be very careful and consider it many times before you upload your photos, photos of your kids, or any documents or anything, any information, medical information, whatever, into this kind of apps, because you never know who is using this data and what those data are used for. Be careful with anything that we do these days because AI has such capabilities to use your data in the wrong way, you know, to steal from you something and stuff like that, right? And I’m sure there are AI wrappers who also help people scam other people, but we don’t know, we don’t want to bring examples here. Yeah, there’s like, you need to be careful with any apps that you’re using right now since, like I’ve mentioned it a couple of times in the previous episode, that software, building software is cheaper than it used to be. So right now, creating a new app, a new software, is not an obstacle for people who have bad intentions to actually cheat on you or scam you or whatever. Software is not a limitation right now for them. So you need to be way more aware than you used to be even a year or two years ago because there is no this cost barrier for building something that will help those kind of criminals to actually make you hurt in some one way or another. Yeah.

Final thoughts and advice

Dariusz Tarasek: Maybe just switching topics a little bit, I’m wondering if for people who would like to start developing something, you know, they maybe are not professional engineers, but they have an idea, do you think that building an AI wrapper would be a good starting point for somebody who hasn’t built anything yet, maybe didn’t even have hands-on experience with coding, but can use tools that help you code things basically without that knowledge?

Wiktor Żołnowski: This is a tough question, yeah, because I would say that with the, even basic web coding skills and the tools that you mentioned that are there, people can build something that works. People can build something that probably could be even distributed, published, and other people can use and pay for. For people who have no experience in programming, that will take more time than for, let’s say, people who have at least a little bit of experience in coding. But people who don’t know what they are doing when building this kind of app, let’s call it by the name, they need to be very careful because it might be very easy to release something that will be very insecure. And until you are doing something that is, you know, like a, let’s say plant recognition, not mushroom recognition, but plant recognition stuff or bug recognition, this is something that will be pretty hard to hurt anyone, even if their plant pictures or their bug pictures will, I don’t know, leak to the internet or something.

Dariusz Tarasek: But it theoretically can be painful for your budget, right? If you won’t secure it right as a developer and people will be taking an advantage of it, you can lose a lot of money, yeah, on the API calls. And I think I’ve heard such stories. Yeah, that could also happen. So I think that yeah, right now that the technology is brought to the masses, like people can use it, can use some kind of things, but I still believe that people who have some experience in coding, in programming, they have a huge advantage over those who don’t have. Maybe let’s wrap it up. You don’t need to build an LLM; you just need to solve a tiny job and do it better than open LLMs. Ship fast, try ideas, and if it doesn’t bring you money, probably kill it fast as well.

Wiktor Żołnowski: Yeah, and you know, like be brave and don’t hesitate to test new things and try to figure out the ways to test it as fast as possible. But still, if you are thinking or considering building something that will be a long-lasting business, not just a one-time shot that will bring you some money, or maybe you can do a few one-time shots, collect some money, and then you should think about building a platform, not just an AI wrapper. Yeah. But for a platform, you need more budget, you probably need to hire more people who will help you with this. Of course, you can hire us for that as well. Maybe not for AI wrappers, because this is something that’s, let’s say, too small for us and too quick, but for building a platform that is utilizing AI, definitely this is something that we have a bunch of experience with. So if you need any help with that, don’t hesitate to contact us. Yeah, actually that whole conversation made me think that maybe I would like to build an AI wrapper in 24 hours and just check if it’s true. Well, you know, me too. Yeah, like I think that maybe the next episode will be already about our experiences and how much money we made on building it. Yeah.

Dariusz Tarasek: Or we will be broadcasting from jail because we did something wrong. Whoops. Hopefully not. Hopefully not. And people call this app a scam. Hopefully not as well.

Wiktor Żołnowski: Okay, so with these good thoughts at the end of this episode, thank you for watching us, thank you for being with us. Of course, subscribe if you like what you’ve seen. Of course, thumbs up. Tell us in the comments if you are using any AI wrappers or if you already built any AI wrappers on your own. We would love to listen to your stories and, of course, I would like to listen to your thoughts on what you think about what’s happening right now in the IT. Make sure to share it with somebody who needs to build a wrapper today. Definitely. Thank you and see you in the next episodes.