AI Can Make You Faster. Chasing Shiny Tech Slows You

Speaker A: I tend to use the right tool for the right job, and it's always good to know what there is out there. We don't always go for the shiny toy just because it's a shiny toy. The point is that sometimes it's all about speed. I tend to find the middle ground between what's great from a technology perspective, but what is great from a business perspective. Because ultimately if you don't achieve the goal from the business point of view, you kind of lost to the point of the project. Because sometimes people tend to over complicate stuff and all you need is a car with four wheels rather than a Ferrari, you know, because I firmly believe that the more you take care of your people, the more they will take care of your customers. I'm sometimes skeptical about certain things where we're becoming very dependent on the AI. AI helps me be more productive, but I still want to know what is happening and even understand the code.

Speaker B: There still needs to be some creativity, which we humans are the best with. Hey, and welcome to a new episode of, um, Masterful Founders. Today, I'm your host, Christian, and I'm joined by my fabulous guest fellow, Fabio. Fabio is founder of a web agency focused on anything, it and online. So he already told me he's a nerd by heart, and I'm looking so much forward to talking to him and with him about anything tech. And what has changed because I also have a huge background in software as a service and, yeah, doing some agency work. So, Fabio, how did you start learning to code? What was it that you like? Okay, this is so fascinating. This weird box on the desk.

Speaker A: When I started out, it was literally the beginning of the Internet. We're still with those really noisy modems which would connect and make like. It's like they're making a whole symphony. I was always a curious person. I love technology. I love to learn new things. I had started out literally with basic HTML, learning literally how to code an HTML page. And back then, obviously it was so new. It was all reading books and YouTube didn't even exist. To answer your question, it was all about curiosity. I am very curious. I love learning new things, and I guess it's what keeps me going.

Speaker B: I read on your agency services you have a broad spectrum of knowledge, which is you're about curiosity, learning how do you come across new technologies? Because there's like every day you open up, uh, hacker news, and you will see there's this thing rewritten in rust again. Why should I care? And then like a few years later, you're like, oh my God, I saw this in Hacker News. I was waving it off and now it's became the next big thing. How do you find new technologies and how is it like, is this relevant or can it be relevant in future?

Speaker A: Like I tend to play the game a bit differently in the sense where I'm always exploring and even including my team, always pushing them to learn new things because ultimately you need to know what there is out there. So in fact at the moment we're venturing out onto AI and then onto blockchain. And those are two things obviously where you need to stay ahead, you know, so go Lang. And the type of technologies, I tend to still stick to the VPS side of things, which is maybe an older way of doing things where other people who would spin up Kubernetes, you know, or use Docker for it, I tend to say, listen, okay, what are we building? How complex is it? What do we really need for this? Is it as simple as a, uh, basic sort of single page app or an application which kind of is purely simple? Laravel, uh, application where you can just spin up a VPS and you're more than happy with it? I tend to use the right tool for the right job and it's always good to know what there is out there. We don't always go for the shiny toy just because it's a shiny toy. If we can build something with Laravel, we're not going to build it in React just for the sake because they would like React or whatever you like. You know, that is something where we've seen a lot. For example, we've seen clients where they build a uh, website using React, then they've got indexation issues and it's a whole mess. From the SEO perspective, it becomes a nightmare. The point is that sometimes it's all about speed. I tend to find the middle ground between what's great from a technology perspective, but what is great from a business perspective. Because ultimately if you don't achieve the goal from the business point of view, you kind of lost to the point of the project.

Speaker B: I uh, don't know how often you came across customers where like they are, ah, pitching an idea in project, they're reaching out to you to make it happen and then you are gonna challenge them. Like the easiest question, like how much marketing are you gonna pour into this? And then you get like fantasy numbers. And then clients are always like. Because I had this like a few years ago, we set up an E commerce system and we had an ERP Running and everything. And then I got asked like, oh, will it handle 100,000 orders a month? I'm like, no, because this is not what we expected from in the beginning because we were sitting on the table like discussing what is going to be the marketing spend, what's going to be like you wish them. And now you're asking this like moonshots. Have you had such experience, like how do you make sure that your clients setting the expectations straight? What's your experience there? Because I find it super difficult sometimes talking to clients and customers like, hey, let's come down, let's, let's think. Then we can plan a proper uh, project and then go step by step.

Speaker A: It's always about communication. You're asking the client, hey, listen, what are you trying to achieve? Where do you want to go within three to six months? And obviously even whatever we build, it's always scalable. Let's say you're running an E commerce shop and then they run a Christmas campaign. It's not scalable. That is a problem. But I think, to answer your question, it's a lot about where you're going to go in the next three to six months and seeing the application that it's, it's a living thing rather than a, uh, sort of something that you're going to play statically there. And then the flip side is you need to be careful where you don't over engineer stuff because sometimes people tend to overcomplicate stuff and all you need is a car with four wheels rather than a Ferrari, you know.

Speaker B: So, um, yeah, oh yeah, I fully agree. I've seen projects where we haven't had any traffic, but it was supposed to be startup and I joined the project later so they were already building and I was supposed to take it over and bring it to the next thing. And I was like, ok, how's the architecture? And they were like, oh yeah, we got 25 microservices. It's running Kubernetes on AWS and it's like 25 microservices. Do they have processes like how to update like all the dependencies in each microservice? No, we need to think about that. Have you had similar situations?

Speaker A: Yes, for sure. I believe in if you have a nail, you're going to use a hammer, you're not going to use a screwdriver. That is the analogy which I'm trying to push because I suffer from it sometimes. It's all about, I believe, using literally the right tool for the right project.

Speaker B: And sometimes it's also about educating the customer. Like, what is out there? Because sometimes they pick up some terms like, oh, react, react, we need to use react. We need to use typescript. We need to have this and that. And like, wait, let's come back. Let's define the scope. What is the business goal? Like, and then we can think about technology. And then usually startups, they also have to think about hiring. So this is what I experienced firsthand. So many times we made some, frankly, stupid technology decisions, like, m chasing the shiny thing. And the shiny thing is attracting people. And this is also like supply and demand. And then suddenly you're like, okay, now we need someone who is actually really fluent in react and node and aws and oh, my God, they are really expensive. This is also one interesting thing. Maybe you can tell us more about, like, how do you hire the right people? How do you avoid falling into this trap of, oh, oh, my God, I needed a rock star. And then you hired someone really expensive who turns out to be, like, having more an attitude than experience. Happened to me before. Lesson learned. So please share your experience on hiring the right people and making the right decision so that you also keep hiring in mind.

Speaker A: I take culture very seriously. Again, we're a team of 11. Um, we're completely remote, so we're all over the world. The reason why I chose the remote is because it literally allows us to kind of hire the best people from anywhere. So we have literally, uh, two interview stage. The first one is understanding the person and seeing if the person is a culture fit. I firmly believe that the more you take care of your people, the more they will take care of your customers. When you create that culture where people are happy, people enjoy working with you, I think you get more productivity. So, for example, we don't really focus on the hours where you need to be there from nine to six. It's about what needs to be delivered. So deadlines. First stage is literally about getting known to another person where I ask questions about what they want to achieve, where they want to go, and ultimately see if we can really fit that dream. And then the second stage is obviously the technical side of things is literally a live coding session where certain tasks are given and we really sort of get to go to the whole technical thing. Those two sessions are enough for me to get an understanding of who the individual is, what they're trying to achieve. I also think that nurturing a good relationship is also a process. I work a lot on career development training. The process is not just at the hiring. It's Also, after they join the company, we also offer a referral program internally. For example, whenever we need to find someone, I go literally go on Slack, advise everyone and say, hey, just a reminder that we're hiring someone and they know there's a successful referral fee, that if someone is employed with us for more than six months to one year, they get a referral fee. So automatically your own employees are going out there, asking friends or old colleagues that might be looking for a job, and there's that incentive. And the famous English saying, birds of the same feathers flock together. So it's a case where hopefully people are of the same mindset, people of the same culture. So, yeah, that is one thing. And the last thing you before mentioned, the aspect of from the second interview, which is our technical part, is how do they think? Or how do they approach a problem during interview? As in, I'm not shy to tell them, hey, you want to use Stack Overflow, I don't care. Ultimately, it's all about how you use the code. How do you approach the problem? You know, you're totally right.

Speaker B: How are we working these days? We are using ChatGPT. We are using, of course, Google and Stack Overflow because you cannot remember everything. This is what we're doing. And the second stage is where, hey, just work as you would work. Use ChatGPT if you like, use GitHub, Copilot. Just get the challenge done and let us see how you approach the problem. I fully agree with you on that. Maybe you can also share your opinion. Like, how would you structure a remote only team? Because right now I have one client. It's really interesting. They're remote too. Completely remote. They're doing meetups once in a while and it works. And how do you manage, uh, your team?

Speaker A: Yeah, so again, I believe it's all about communication. It's not easy, of course, but it's also about organizing these meetups. So like once a year you're actually meeting up physically. It's also about organizing different events. We do a lot of training events. We're all the time organizing things. For example, about Docker, things about, for example, this week where we had something around blockchain, and that automatically brings the different individuals together. It's all optional, you know, so it's not like one of these things where you have to come, but you're automatically bringing people together and people are sort of meeting, discussing new technologies. Someone is obviously discussing the topic. Last time we had one about AI, how to use, for example, Gemini. There's a New feature in Gemini that is one of the things that we do a lot. Another one which is a bit more on the fun side is we tend to do a games thing every month. Life is also needs to be about fun. It cannot be just about work and just about technology. But I think the most powerful thing that works is the meetups and then we have a weekly standup where we go through everything. Everyone discusses their problems or opportunities because uh, it's always good to share that.

Speaker B: I had one thing you were mentioning with hiring the right people and technology choices. Rust versus Golang for the univaire like Rust, I think first time I came in contact with was 2013. So it's a quite recent language. So is Go and uh, it's highly opinionated so I would love to hear your opinion on it.

Speaker A: So Golang gives me that satisfaction of working with a language where one you have a proper compiler and that is like it's such a nice feeling. And then it's like the fact that it's so strict, that is something that I love. And also the performance, you know I built recently a very small crawler, just a check and uh, it's the way it distributes on multiple threads and it's so beautiful. Again with help of claw, just to be entirely clear here but just for speed. But I think Go is so performant. That is something I really enjoy. And then why I kept on going with Go is because when you look at certain blockchain technologies like for example the Ethereum, the EVM and this kind of thing are built on Golang and this kind of thing, I still have an interest in learning Rust for sure. I saw recently the Linux kernel how they're introducing the aspect of Rust. Ah. And honestly it's the usual excuse it's time, you know, it's just a matter of hey listen, you know, spend some time learning a new language because if it wasn't for a time I'd be learning Rust and I'd be learning so many other things. There are so many different and exciting stuff but you tend to get into your comfort zone and in fact we're on the blockchain side of things. We were focusing a bit on those chains which are more golang focused. But uh, at one point we will definitely pick up Rust because Rust is also being picked up and we're huge Linux fans also. So definitely something to pick up at some point.

Speaker B: This brings me to the next topic. How did you get into all those blockchain?

Speaker A: A couple of years ago we were entrusted to get into what we call ICOs. ICOs is essentially selling. You have a particular coin and you want to sell the coin to different people and people can buy it. At an early stage, around three, four years ago, this was massive. We were entrusted to get into an ICO project. So we were asked to build where essentially the JavaScript, as in connects to the different smart contracts. That's where the whole thing kicked off. So it was a very, very successful project. We started getting more involved and uh, we had done another ico, so we had to build again a website using different web 3js, using metamask and all these different wallets. And then we started to sort of say, hey, okay, we've done the front end, now it's time to sort of bring in. We do the smart contracts ourselves because before it was a kind of a split team, so one agency was doing the smart contracts, we were doing the UI and the interfacing. That is then, then when we started to upscale everyone sort of on, um, smart contracts. And now we're also toying around with whole infrastructure side of things.

Speaker B: I just had this morning a conversation with a friend. During the run we were also talking a lot about AI. And then at some point I was like, yeah, what actually happened to everything Blockchain? Because he mentioned one of his friends, he is a Web3 developer, was like, oh, I haven't heard Web3 in a while. Is it like my bubble that I'm no longer getting any information from that corner of the inner webs? Is it still ongoing and how is it going forward? Like, sure, it was hype at some point and now we are like in some quiet pit. Maybe you can share some insights on what is going under their hood.

Speaker A: Back around 10, 11 years ago, I was studying on big data, so my master's was on Hadoop. But in reality, does everyone qualify for big data? No. You know, sometimes you just need MariaDB server or an Oracle DB and a couple of million rows, which does not constitute as big data. Uh, blockchain is the same thing where, yes, we will see a lot of things moving onto the blockchain perspective, like finance and then sort of other aspects. Like for example, we're seeing a lot of online casinos moving on to the crypto side of things. There are quite a few use cases like different government services when it comes to authentication, validation. I think those things will be moving over time. I also think that Web2 is going to still be around for a long time because I don't believe that every single use case should be on the blockchain. It's a kind of a situation where it's repeating itself then a different facet where AI is going to be rolled out on a much more broader scale in comparison to for example the blockchain. AI and blockchain are going to fuse together and then we're sort of one complements the other and that will help a lot.

Speaker B: Have you come across solutions where AI is being more on prem and completely decoupled and under your control? And uh, uh, have you looked into this or what is your general thought on that?

Speaker A: I think AI on prem it's the investment needs to be made when it comes to exposing your data. Uh, like for example, you know, this whole thing with AI and mcp for those who don't know mcp ultimately having your services like Slack and whatnot sort of connects to ChatGPT and Claude. So literally these different AIs can consume your data. Uh, I am a bit of a security conscious person and I don't like the idea it's about a fine balance between the business side of things because obviously going on prem buying all those GPUs and all of that can get very expensive very quickly. And as you know well enough, you have to buy your hardware and then the hardware can easily get outdated after X amount of time. So on prem is something you need to consider based on the data you're exposing. So it's like a government, I would think twice about exposing the data.

Speaker B: Right now everyone is jumping on this AI topic and everyone is building something on top of it. And we just got started with LLM and I'm pretty sure there's going to be some hardware change which is doing something in hardware way better than it is right now. Or like AI agents where you can have your specialization which is then narrowing down the models and the memory requirement and the working data set to work with. So what does it mean now for you as an agency? Is it like giving you the opportunity to be more productive using such AI agents to do almost like boring tasks but like produce more code? What do you think about that?

Speaker A: We have a 10% part of the team that don't use AI. They refuse to use, they want to write everything themselves, which I really appreciate and I can really understand. The person is very keen and uh, interested on learning, on developing their skills and knowing the language really well. So that is just one uh, side which I just want to put out there and I really appreciate those people. And then there are the people which are like me, you know, where hey Claude, you know, how do I go about this? And obviously you have to monitor. Sometimes it gets it completely wrong, completely wrong. Sometimes you have to revisit. I'm sometimes skeptical about certain things where we're becoming very dependent on the AI and even from a vulnerability. So last time I was talking to uh, someone who specializes on penetrating testing and he was telling me how my job is ballooning because everyone depending on AI, but AI is not handling, for example cross site scripting is not handling SQL injection. Again these are just examples and it's not because AI doesn't know about it, but it's because sometimes it needs to be instructed to handle those elements. I am uh, very much on a believer that AI helps me be more productive. But I still want to know what is happening and even understand the code because AI for me is more about saving that time of having to search in the different docs and looking for pieces of code and doing the stuff which is rudimentary, that stuff which is boring. Like for example I had to literally last time build a small react application to display some test samples of what I was working on. That is an application which I wouldn't build myself because it's just a waste of my time.

Speaker B: So what about the security implications? It's not just all those nitty gritty hands on things like cross site scripting and SQL injection, whatnot. The other day I was hearing about that ChatGPT was giving you like the full blown product or like the solution for your problem and was like hey, use this dependency, use that dependency. And then turns out like some of them were completely made up of. I was like oh uh, this is so stupid and you have to see what was it thinking about and get the right library. But then the second thought was like someone could just create this library and sneak in code. So this is like a new world of supply chain attacks. You could actually.

Speaker A: Yeah, true, very very true. There are definitely going to be those situations where we might not be building the full thing and someone approaches you with that. But I still think that these complex topics, I think AI is still not there yet, at least for now. Junior devs I think are to have it the hardest where they need to upscale quicker because AI is kind of replacing a bit that layer. In my opinion the junior to medium

Speaker B: junior devs is going to be tough companies. They will hire more senior people and they were like yeah, we can scale up the development M velocity with AI because AI is taking care of the boring stuff now. And seniors, they're going to double check and proofread, to pull requests the AI agent created. But I'm not sure now if this is discouraging people from entering it and that uh, we might have people who are getting more and more senior, also like senior in age and they're like, okay, they will die eventually. Right? ChatGPT is not perfect and you still need to sharpen your skills by doing something. So this is going to be a really interesting development I guess for the next couple of years. I hope that no one gets discouraged. If you, you get above this junior line and you get more senior, more experienced, there will always be some room for you because nothing is perfect. There still needs to be some creativity, which we humans are the best with. Because everything ChatGpt LLM is doing, it's just learning. Sure. It's creative by hallucinating stuff. Is it the right kind of creativity?

Speaker A: Yes, exactly. Creativity comes in different forms, I guess.

Speaker B: I'm pretty sure that the humans won't disappear from processes anytime soon. How did that happen? How did you get from writing software, cursing at your computer and compiler to okay, hardware, Give it to me.

Speaker A: When I was very young I used to enjoy building PCs. It's something which I love home labs, so super passionate about that in fact I have my own small home lab, uh, which I put together. M again obviously we have our own cloud hosting ourselves, uh, so we manage different VPs for different clients and it's something I guess which is super challenging because you have to always be on the ball so you have to always be learning the security monitoring and so many different facets. So I think it also satisfies that hunger to always learn more. You know, that mix of curiosity and that aspect of dismantling something and not knowing how to put it back together, I think that is what's kicked off the hardware side of things. So yeah, nowadays even with all whole Arduino and Raspberry PI, the Pico and all of those and ESP32, it's such a huge. I would also suggest anyone interested Reddit. Reddit is fantastic. Obviously YouTube is fantastic because it's all videos, but Reddit's the community is so inspiring.

Speaker B: So if you have one good advice to give out to our uh, maybe completely checked out audience because it was too technical for the ones who are still with us.

Speaker A: Yeah, I think stay curious and stay consistent. And I say stay consistent more for the entrepreneurship side of things. So if you have anyone who wants to freelance or start their own agency, I would say stay consistent because that shiny syndrome can detract you very, very easily from getting to where you want to be. So staying consistent, I think, is super important.

Speaker B: So, yeah, stay consistent, stay curious, and there's going to be huge work with things. Thank you so much, Fabio. It was a pleasure, and thanks so much for your time.