Tobi: Hello, friends. This is the Alphas podcast. I am your host, Toby. The goal of the Alphas podcast is to empower CTOs with the info and insight they need to make the best decisions for their company. We do this by hosting top thought leaders and picking their brains for insights into technical leadership and tech trends.
Tobi: If you believe in the power of accumulated knowKusalic to accelerate growth, make sure to subscribe to this podcast. Plus, if you're an experienced CTO. You will love the discussion happening in our slack space where over 600 CTOs are sharing insights or visit one of our events. Just go to alpha list.com to apply. Tobi: Welcome to the podcast today. Meet Ivan Kusalic, CTO of NPAL --
Tobi: I am with Ivan Kusalic. Uh, who's the CTO of NPAL? I don't know. Who knows? NPAL here, maybe. Ivan, do you want to tell us a bit more about NPAL as well? But you are leading 250 engineers at NPAL, which is an like a solar company, right? Like an energy, uh, company basically. And you just recently went back to coding.
Tobi: You told me after seven years of silence because you're so excited about ai. So we wanna talk about navigating ambiguity. Um, like, like navigating in, uh, un unsafe spaces, unknown spaces, and we wanna talk about, about AI and, and what it means to all of us. Mm-hmm. Uh, I mean, it's a, it's a hot topic right now, like everyone talks about it, but, uh, like we still, we still add like a bit of a, um, icing on the cake today. Tobi: Um, welcome to the podcast. Thank you very much and
Ivan: great to be here. And I think you opened already three questions. So where do we start? Do we want to talk about ai? I feel that, you know, ai, part of the conversation is gonna get, uh. It won't age too well, but still we definitely need to do it. Tobi: Yeah. Let's, let's do it at the end. Ivan's Journey into Engineering and Leadership --
Tobi: Let's, let's first start with, uh, like, as I always do with your persoNPAL nerd path mm-hmm. Let's say like, uh, how did you become an engineer ultimately? Like, wh when and, and why? Um. And, um, then like, why did you decide for, for leadership, uh, later on, um, and, and now, like why, why did you decide to jump back and also also code again? Tobi: So, um, yeah, let's, let's start with your nerd prof. Like, like why, why do, why do, why do you use computers?
Ivan: You know, my whole youth is kind of a nerd story. My, my father was mechanical engineer and so he made sure that me and my brother get exposed you not to anything technical. And, you know, I started programming when I was 10-year-old.
Ivan: Uh, there was something called logo and q basic. Not sure how many people know logo, but you could draw with a turtle on a screen, kinda lines and polygons and so on. So, you know, it was quite cool, uh, for a kid. And, uh, you know, I started programming somehow got into, um, algorithms and data structures quite a bit, you know, uh, coded in c plus plus red black trees, played with, uh, dynamic programming and, you know, trying to figure out how the hell all of that works.
Ivan: So, in a sense, uh, a lot of programming, but also, uh, different things like, you know, electronics. So I was, as a kid, uh, etching my own circuit boards with acid from copper plates. You know, at some point I build a. Very trivial lamp that you could, you know, light with lighter and you could, uh. Uh, stop it by blowing on it, because basically it would swing and there was just a light sensor below it, right?
Ivan: So to check that there is light from a lighter, and when you blow on it, it would, uh, you know, drop the light and so it would turn off. But for a kid that was, you know, like a magic. So really, you know, my, my whole youth was kind of all about, uh, all about nerdy stuff. And somehow got really into this, um, you know, both in ma into maths and also into, uh, into computer stuff.
Ivan: So also ended up, you know, a bit competitive, uh, you know, competing on natioNPAL level in creation, math, logic, computer science, so natural event to study computer science and university. You know, there I discovered machine learning. Uh, and kind of fell in love, uh, and, you know, did a lot of project in machine learning.
Ivan: Um, one of the things that I really liked was using support vector machines. Even did with France some projects that, uh, uh, that even Ian government used to, you know, detect and, and, uh, classified traffic signs from on the roads. But, you know. Support vector machines. Nobody talks about them those days. L LMS clearly won.
Ivan: I just wanted to say you're a dinosaur. It's a bit like that, you know, uh, your age completely changes, uh, when we talk about AI because their age just stretches. Right. So, uh, say, but, you know, uh, did kinda a lot of machine learning and, uh, then went to industry, uh, stayed for a year in Croatia for kind of tiny startup, actually American startup, um, that had a dev center in Croatia.
Ivan: Um, and then just wanted to see the world. You know, I wanted to go somewhere where actually software is in, uh, more focal than in reb in Croatia. So, you know, chose Berlin, uh, joined the platform as a service company. So if, uh, some of your listeners still remember Heroku kinda actively for Heroku clone basically. Ivan: Um, and after that, uh, went to do different, you know, things on the backend side.
Tobi: Sorry, but, but Heroku Heroku is still great, right? Uh, like the Rails people in here, like, uh, I think it's great. Many of them are potentially still using it, Ivan: right? I think it's great, but you know, it's, it's less talk those days than it was, you know, in 2010.
Ivan: I think in 2010, it was really the thing. I think it was one of the few things on that level abstraction, right? Kind of proper application layer platform. So it was a really proper pass. And yeah, the somehow ended up then afterwards, um, as, uh, engineer and later architect in here technologies, which is, uh, a kind of mapping company, uh, often, uh, working in automotive space.
Ivan: Um, you know, there I designed some cool multi-regioNPAL, highly distributed, high consistent guarantee systems, and that was cool, but I realized. I mean, I'm designing new systems two, three times per year, but first, it's not just me. There are many people and they also need to like it and challenge it and improve it and take ownership.
Ivan: So that was kind of my gateway to management, you know? So I went to the dark side, wear a bit of black t-shirts periodically, if your listeners could see me now. And, uh. Yeah. One thing led to another, somehow ended up as a CTO in a small startup, uh, responsible for, let's say 30 to 40 engineers, depending on the time.
Ivan: Then ended up there also as a, as a, um, all the managing director. You know, they sold it to me by saying, I. All the fun of being able to go to jail and zero pay rise. So this was a great deal, but I jumped on it because, you know, I could learn a lot. I ended up even being a co there. Um, and, uh, did cool things like, um, you know, negotiating big deals or, uh, um.
Ivan: Redefining sales incentive schemes. So I can tell you that's very different than working with engineers. And yeah, that's, you know, like that was a kind of real position that, that I learned a lot. So after that, went to HelloFresh as VP engineering responsible for 150 people. And then after that, you know, last year January, joined AL as a CTO. Ivan: So that's, you know, kind of the nerd part and also a bit mixed into that, how the management happened along the way.
Tobi: Okay. The Role of NPAL in the Renewable Energy Sector -- Tobi: Solar industry feels a bit like from the outside, could, could be boring, could be interesting, right? I mean, super regulated space. Uh, many, many providers, many energy providers, many, many, many companies that are like under digitized. Tobi: Um, like what do you do there? Like,
Ivan: so I mean, I think that, uh, you know, people think that it's a simple industry, but, uh. It's, uh, it's actually quite an old industry with, with all comes with challenges, but you know, in nutshell, ample is a Greentech unicorn, uh, you know, around 5,000 people, uh, uh, and 215 Tech.
Ivan: We are headquartered in Berlin and what we do, we basically, I. Uh, sell, install and manage renewable energy systems for households. So we are selling things like, you know, solar panels, batteries, EV chargers, and heat pumps. And then we make it work as a one whole system in a household that is basically our energy consumption and, uh, yeah, basically.
Ivan: Help people actually save energy, also use greener energy and hopefully make a world a bit better place, or at least risk of the climate change decrease very, very slightly by our efforts. Tobi: Yeah. So you're in a head, head to head race with, uh, Anson, right? Best we can say that. Lina, I not Ivan: comment much on that, but I think that numbers say that, you know, the, the race is not completely equal.
Ivan: So there is a party that is a bit leading and, you know, some statistic would clearly say at that temple. But yeah. Uh, we are the, the two most prominent companies, uh, in Germany. Tobi: Oh, okay. Okay, cool. So that means like you are, you're also building, um, something like they do with this, how is it called, um, heartbeat. Tobi: Like you have some device at home that, uh, kind of, uh, controls all your energy flows.
Ivan: Yeah. So basically we also have a, you know, what we call Pul one, which is in the, in the nutshell, uh. Small computer that then coordinates the rest of the household. Uh, and so it's used both for, you know, regular managing of your energy consumption, but it is, uh, also used, uh, you know, to, to interact with something which we called VPP or Virtual Power Plant. Ivan: Um, and so, yeah, um, you know, uh, definitely have such systems that they have
Tobi: some complexity. And what does, uh, the virtual power plan do? Like, does it combine all households to one virtual power plan, or Ivan: what exactly? So basically. You can think about the key problem in energy space is really stabilizing the grid.
Ivan: So a electrical grid first. It's very old and kind of, you know, a lot of additions and complexity in it. And second, it's like, I mean, it's uh, you know, it's, it's energy. So if you have too little of energy, then you know there is a blackout. If you have too much of energy, then some fuse or even versus something else.
Ivan: Flow the summer. And so really the supply and amount of energy needs to be, be very precisely matched. Um, and so there are whole, you know, uh, uh, systems with, uh, energy trading basically, um, energy trading markets, uh, that enable you actually to balance the grid, right? And so what the virtual power plant does, on one hand, we.
Ivan: You know, uh, commits to stabilizing the grid by either absorbing or giving energy back to the grid. Uh, and then, uh, what we do, we execute it within the next 15 minutes by coordinating our households to charge or discharge the energy to the grid. And so, you know, we are in Germany, we are the only one who have, uh, you know, such advanced VPP.
Ivan: Others are trading they ahead. So basically they say today what they're gonna do tomorrow, what we are doing is we are committing, uh, to a trade, you know. In a moment, and then within the next 15 minutes we need to execute it. And the benefit is not, of course, just stabilizing the grid. The benefit is also that basically our, our household then gets the benefit of that price exchange.
Tobi: So basically through you, I could, like, if I have such a device and I have a storage at home, then I. You would automatically manage when the storage is charged or discharged as well. So I could also buy cheap energy for, uh, like minus prices or whatever. Um, and uh, like I don't know if that's really exists. Tobi: I, I, it actually Ivan: does. Tobi: I I would earn money with through consuming energy. And, and, and you would manage that exactly.
Ivan: That that's, that's it in a nutshell. But, you know, this is like a relatively complex setup with many work pieces, but this is really it in a nutshell. Tobi: Okay. Okay. But I, I think like most ho households are way, way too old school to kind of, um, really manage that efficiently, aren't they?
Ivan: But that's the whole point, is that you don't need to manage it yourself, and we manage that in a background for you. What you need to have is, you know, we need to have a particular kind of contract and agreement, and you're willing to do this and we'll inform you what we are doing over the app so you can see what's happening.
Ivan: But in a nutshell, this is all, you know, managed in the background so that that's what the system itself does. Uh, so you don't really need to think about that. I mean, this is, you know, you also would not want to do that, you know, by being awake or needing to watch what's happening with your energy, right? Ivan: So, of course, that's all, uh, that's all
Tobi: ized. Okay. And, and where is the complexity really? Like the, what's, what's the biggest, like what, what, what keeps you busy? Like 90% of your time, I guess management, but where, where's like technology, uh, te technologically, where's, where's the cool stuff? Ivan: So say that really, um, the main kind of complexity that, uh, NPAL has a technical organization is that we have a really widespread value chain.
Ivan: Uh, and so, you know, that's the, the main flavor. I know it's not a real time biding system in EdTech where you have millions of requests, you know, uh, while at the same time. We have significant number of requests, significant volume of data, but that's not really the challenge. The challenge is, you know, how do you manage this complexity with, with a few hundred engineers and you know, you have a value chain, uh, that starts in China because you know and travels the continent.
Ivan: Then you have legislations, uh, and the compliance topics. Then you have. You know, just, uh, just the grid itself. Uh, it has two layers. Uh, one is kind of physical transit layer and there are four operators for that. But then when you go to the next layer, which is kinda a layer that connects households to the grid, and not just the transport across the, the, uh, bigger geographical distances, there you have 950.
Ivan: You know, parties, uh, some of them are very old, some less, uh, some are digitalized, some not. They use different forms, PDF, all kind of stuff, right? So just that is complex. Then, um, you know, if you go to our sales process, I mean, uh, you know, basket size is, you know, uh, 20 plus K. So, you know, the sales process is not just bombarding people, but you know, a lot of, uh, conversations that take time.
Ivan: Uh, then installation is not trivial. You know, you actually need to manage and set up all of this, uh, so that, that it really works in a household. You need to test it, diagnose it, and so on, because although once you set it up, you cannot just, every, you know, few weeks go back to that household. That would be crazy and simply in a cost to manage that.
Ivan: Right? So there is this, uh, thing where you need to be able to update it remotely or, or you're in trouble. There are generations of devices. Um, that, that are, you know, that you need to manage, which have different capabilities, different, uh, you know, uh, just trivial different, for example, battery, uh, you know, battery sizes.
Ivan: Um, and then maybe the most complex of it all is the household is a physical thing, you know, and you cannot pretend that someone does not have washing machine turned on just because you're telling your sister to do something. You know, the, the system will behave as a hardware system and, uh. You know, uh, on the software side we can give instructions, but, uh, you know, system will do what it doesn't.
Ivan: So you need to kinda, uh, be able to adopt to that, monitor that. And so really, um. It is a lot of moving pieces, a lot of complexity that is, so the value chain is just, just very long. And so, uh, core of my job is in the end, how do I manage that, you know, with, uh, while not having, you know, uh, 10,000 engineers for every single piece of this journey. Tobi: So to sum it up, you're building is here.
Ivan: I think that that's. True for 87% of the companies out there. So of course there is a CRM part to it. Uh, but yeah, it just, uh, you know, it's, think about it like CRM plus automotive, uh, you know, value chain, and then you're kind of somewhere there. Tobi: Okay, okay. Okay. Navigating Ambiguity in Complex Systems -- Tobi: Um, yeah, and, and that like fits quite well to the topic of, uh, like navigating ambiguity, right?
Tobi: I, I can't imagine like, um, if you, if you. Enter that space. There's a lot of unknown, a lot of unclear, um, ideas and then, or unclear topics. And then like, just knowing where to step next is, I mean, yes, you could have, like, I guess your vision is something like becoming the leading, uh, VPP. Um, of the world or Europe or, or Germany at first, I, I guess in Germany already, like they're on par with, with, uh, or a bit ahead of eCom or however you want to say.
Tobi: And, and, and, and I guess that, that, that kind of then can become like quickly your passion, uh, to, to, to learn navigating that, right? Yeah, Ivan: definitely. I mean, I think it's, uh. Something that I've been attracted from the young age. I like understanding complex systems that's not know, that's what I really was interested when I was, you know, dismantling, you know, at radios, you know, when those things were present, uh, and understanding how the pieces fit.
Ivan: And, you know, that's, uh, how I, what I enjoyed as engineer, basically designing system, optimizing them how to make them cohesive. You know, high quality, and I think this is the same in management of the lichen. You know, Al really tickles this, you know, this bone that is basically about actually managing complexity.
Ivan: And, you know, nobody can tell you almost anything what you do there, right? So, so in a sense like you're, you are going navigating uncharted waters and there is a lot of complexity and, uh, you hopefully enjoy the ride. The Importance of Intuition in Leadership -- Tobi: And, uh, do you have any like, interesting stories that, uh, highlight the importance of navigating ambiguity?
Ivan: I do, but I'll go, not go with Temple. I'll actually go with, uh, you know, a bit before maybe one of the almost formative stories from, uh, you know, uh, from earlier. And, you know, on a gloomy winter day, that's where we start. Um, I was sitting in my office staring at a pile of papers and I just didn't want to sign it.
Ivan: I just hesitated. Um, my team has spent countless hours, uh, working on this pile of paper. It represented contract for the biggest deal, will secure that year. And as a CEO. I've spent all the months in negotiation with Dubai tourism. The deal was about bringing travelers to their location. Uh, more than 70 people have been directly involved in this journey to secure the future of a hundred people company, and I was holding back.
Ivan: Why was I holding back? Because the contract was nearly perfect. A small army of lawyers has reviewed it. But what if something went wrong? What if the world changed? Um, so my tuition kick in and, uh, led me to add something called force Major clause. Uh, it stated that if there was an overall 25% drop in travelers to region, we wouldn't be required to meet 70% of our target volume to get paid, nor will we face, you know, quite harsh financial peNPALties if you didn't do so.
Ivan: Instead, we'll be compensated for any delivery we achieved. So both parties signed the contract at the end of January 20, 20, 2 months later. There came covid. So, um, you felt that Tobi: already? No, no, no. Uh, yeah, it absolutely makes sense. Um, so like, also like sometimes you have that feeling in your, in your gut, I guess, right? Tobi: Like also through not only intuition, but, um, but, but, uh, like a lot of learning.
Ivan: Um, no, the way I think about intuition, intuition is really experience in a topic you understand quite well. Uh, that you are not even needing to consciously think about it, right? So I think in this case it was about really, you know, navigating complexity of that contract, but also thinking about all the consequences possible.
Ivan: But really I think intuition is. Something that is actually not illogical. I think it's a deep source of knowKusalic, uh, that, you know, just sometimes we underappreciate western society. So, so yeah, I think it was, um, you know, sometimes intuition just speaks to you and I'm sure you also have stories about intuition, you know, leading you to do something.
Tobi: Absolutely. Um, I sometimes find my intuition also in a way, a bit irratioNPAL. Um, and, uh, I ratioNPALize it later through like numbers. Mm-hmm. Mm-hmm. And then, um, like even if my gut first says no, then, um, if I add numbers and if I, if I add data, then um, often I work against it as well. Mm-hmm. Um, so I, I think intuition can also like.
Tobi: Stop you from doing something you in the real world want to do. Yeah. Mm-hmm. Um, and, and you first have to convince yourself that it's a good thing. Mm-hmm. Right? Um, but, but, but yeah, generally, I, I like, I agree. Ivan: And the way I think about it, you know, intuition is an, uh, impulse to explore. Right. And so then the exploration can be completely in line with your intuition and can be completely against it, right?
Ivan: Because like you said, you know, you do a bit of systems thinking, you aNPALyze the situation, you crunch the numbers, uh, and then you don't let just, you know, the simple lead you somewhere, but it is an impulse to look at something much closer. And so, you know, like you said, it may, it might be that you go directly against it, but often it'll trigger, you know, right thoughts for exploration in a moment that you would otherwise miss.
Tobi: And, um, do you have like other examples of your day to day? How, how, how you deal with it? Ivan: Yeah, definitely. So, I mean, if I were to super abstract my, you know, leadership style, um, I actually rely on intuition quite a lot. So in a sense, what I set is, you know, the direction, vision and direction and milestones and so on.
Ivan: But I use intuition to, uh, uh, kinda seize the opportunity in a small moment to improve something, you know? So I have kinda understanding of the, uh, you know, organization as a, both, as a system, as well as the kind of, you know, organic, uh, uh, entity with, with lot of people. And so I might have just a random conversation with someone on a certain topic, particularly if it's 1 0 1 or something where there is not really.
Ivan: You know, structured approach where you just want to get a certain outcome, but rather if it's more kind of brainstorming or a free flowing conversation. Um, my tuition often spots this moment in conversation to kind of either connect the dots for, for someone or to make a small nudge in a certain direction.
Ivan: And I think I do it probably. At, at minimum 20 times per week. So this is really almost like a reflex at the time. Music very often, which is kind of slight nudge into, into a certain direction.
Tobi: And do you also sometimes have, um, like a topic on your mind that kind of. Is with you for a while, and, uh, whenever you jump in a one-on-one and talk to someone, like you have like a, I don't know, a tool you wanna use, a system, a process you want to introduce, and then like you connect a lot of conversations to actually this, this, this thing that you have in mind Ivan: definitely happens.
Ivan: Uh, I think, uh. The more I have persoNPAL systems, uh, to manage all those thoughts, the, the less this happens, you know, because before I, uh, developed elaborate set of notes and persoNPAL system to manage thoughts and productivity, then it was happening much more because, you know, I, I wanted to get this thought out of my brain.
Ivan: Out in the world now, uh, still happens. Uh, but, and particularly if something is my top priority in focus, then, you know, then I also let it go, right? Because I'm driving to the, to the right priority. Uh, but also sometimes, you know, uh, more kind this exciting thoughts, uh, you know, so I, I share them in, uh, informal settings and so, and kind of connecting to them.
Ivan: Uh, but I often just also note them and put them in my calendar for the appropriate moment. So, you know, depends, depends on the day on a, on a context and so on. How do you manage that? Tobi: I have to admit that like sometimes, uh, like I'm not super structured with that. And if I have something on my mind, I, I, I, I, I keep it with me for a while and leave it ripe, right?
Tobi: Mm-hmm. Like iterate, iterate, iterate. Um, and uh, at a certain point I get it out like maybe sometimes even after. Six months or 12 months, like, it, it, it just rips and it, it just continue iterating, iterating, iterating, and then all of a sudden it, it, it, uh, like I, I also annoy, annoy people with it, I guess.
Tobi: Um, but then all of a sudden it, it, it pops up and, uh, it's the right moment to release it. Um, uh, but, but, but, but yeah, it's not like I don't have like a fixed structure for it. Tools and Techniques for Managing Thoughts and Productivity -- Tobi: How do you, um, like structure your day to day? I mean, you had just mentioned your calendar. Mm-hmm. Uh, what, what, what, what else do you use to, I dunno, um, uh, remind you of, of things and, uh, keep notes. Tobi: So,
Ivan: I mean, I do, let's say two things. One is that I am. Quite a power user of, of Evernote. So probably using it for 10 years. And you know, at some point they were sending Evernote, really Evernote, excellent school, old school, although I think they got acquired maybe, uh, two, three years ago. And since then, they're really also, you know, adding a lot of feature and it's, it's actually getting more and more modernized.
Ivan: So I really. I really actually enjoy that. They are really improving quite a lot. Um, but you know, for me it's just, that's kind of almost like my database of thoughts. Uh, you know, they were sending at one point reminders to people, kind of what's your biggest productivity streak? You know, they would say like, okay, this month you use the.
Ivan: Evernote one day less than the last month, and the next month they would say, used it one day more than the last one. But you know, I was using it every day. So the calendar calendar in a month would be 30 or 31 days and they would tell me it's plus or minus one day. So it was quite funny. Uh, but you know, I generally use Evernote.
Ivan: Um. Really systematically, I have, um, I have, you know, I use it for introspection. I use it for work thoughts. I use it for even those kind of thoughts that take sometimes months or even years to get really precise and polished. Um, so it's my main, uh, let's say just knowKusalic database. You know, I think about it as extension of my brain.
Ivan: My, you know, real memory is not, uh, definitely not, uh, uh, you know, uh, bulletproof and, uh, is quite faulty sometimes. So this is my extended, you know, uh, system for. For actually recording thoughts and iterating. It's my workspace of, you know, mental, uh, mental things. The other thing I use is calendar. Like I said, um, I use calendar both for, um, you know, scale kind of as a literally as agenda of meetings, right?
Ivan: But I also use it as a timely reminder. So I also have a separate calendar where, you know, it's just knows approximately what needs to happen in which amount of time. That part is still not, not perfect. I would like something better to, for time relevant. Uh. But, but Calendar is the best I got so far. Tobi: I'm, I'm using a tool called Reclaim ai, which has been acquired by, by Dropbox recently.
Tobi: Like to also manage, uh, like you can manage your, your streaks, uh, and, and your, your. Habits in, in the calendar as well. Like, I, I don't know, like I want to have like two hours each day for productive work or something like that. And you can, it, it basically tries to defend your calendar in a way. Um, and also puts in like to-dos, et cetera, quite helpful. Tobi: But, but yeah. The Future of AI in Engineering --
Tobi: Also it's problems again, um, and I, I think like the next generation, um, and also maybe like good, good, good step to come to the, to, to the topic of ai, which we also wanted to touch. Um, will, will be like. A persoNPAL assistant, really, right? Mm-hmm. Where you can dictate all your thoughts, put in all your thoughts, and get, get them out again without necessarily, without having, uh, really notes somewhere that you have to pick up again.
Tobi: But just like an extension of your brain, basically, right? Like you just mentioned. Um, yeah, that would for me, like be the next step. Let's, let's wait for that. Um, I, I, I think like one to two years and then we have it. Um, uh, what do you think? Like, um. About, um, AI and the future of engineering. I mean, we had a little pre-discussion and mm-hmm.
Tobi: And already could have al almost recorded it. Um, what, what's like, are you afraid, are you positive? Um, what, what are your thoughts on it after like, you also stepped back to engineering, right? Um, like what's your experience? Ivan: So first, I mean, I think it's a topic that absolutely cannot be ignored. So I think that, uh, it is not clear.
Ivan: You know, how it's actually gonna gonna change our industry, but that it'll change it and that it'll be, you know, fundamental part of it. There is no doubt in my mind about that. Um, so. I'm thinking that, uh, every single developer will be definitely using ai, you know, maybe even within this year. Uh, I think definitely within next two years, anybody who does not start to, uh, using AI will be increasingly outdated.
Ivan: Uh, now what will that mean? You know, for the, for the future of our industry? It's hard to say, uh, you know, uh, let's say more, uh, negative, uh. Kinda our pessimistic point of view would be that, uh, you know, we'll automate a lot of things and produce huge amount of code, but then we'll introduce also consultants, uh, and agents, specialized agents that manage that massive of bloated code.
Ivan: So, you know, that's one possibility. Um, but I think that, uh, engineers will still need definitely to be there to, to guide it. Uh, and that's also my persoNPAL experience with coding, right? So I think that, uh, it is unbelievably. It's quite disbalanced. It is not, uh, you know, uh, far, very far from perfect.
Ivan: Requires a lot of guidance, but I'm really impressed, um, what it can do. That has been changing, you know, sometimes, uh, sometimes the, the jumps happen within a week. So really I think it's, it's there to stay, uh, how exactly in which form? That's, uh, that's a bit harder to predict from my side. How do you think about it?
Tobi: So, um, just to be a bit more precise, like, do you use, I dunno, za or, or something like that? Or, or, or other tools or what, what do you, what do you, what do you do? Ivan: So, I mean, for years I'm using now, you know, uh, uh, chat interfaces, right? So I use bot chat, GPT, uh, and O Cloud. Um, I often actually ping pong between them, right?
Ivan: Uh, so, you know, like particularly when I'm writing a complex piece of text, you know, I create my bullet points, then give it to both of them and then ask to improve the version. So, so that works well. Experimenting with AI in Coding -- Ivan: But, uh, basically in November last year, I've, uh, really. Wanted to understand actually for myself, um, what's possible.
Ivan: Uh, that's why, you know, did an experiment where I was really first time coding after seven years. Um, and, uh, there I just used charge GPT, you know, literally copy pasting, uh, you know, uh, to charge GPT. Um, and it was already super helpful. But since then, I know now I'm in, um, last two months, uh, uh, kind of playing on a more complex persoNPAL project.
Ivan: I wanted to see how it behaves on non-trivial example. You know, in my first experiment it was just a persoNPAL blog, you know, and managing subscriptions for newsletter and things like that. And now I'm working, working, you know, on a, on a persoNPAL system, um, that, uh, you know, has both front end backend and a mobile app. Ivan: Um, uh, and so, um, yeah, uh, there I'm using, uh. Cursor. Productivity Gains with AI Tools --
Ivan: Um, and, uh, you know, mostly with, uh, with, uh, Claude and actually two weeks ago I believe, or something was, you know, the new version of Claude was released. And that was, you know, I opened it one evening and then I opened it next evening, uh, cursor. And, uh, the increase in productivity there already was. Ivan: Those things are hard to judge, but I would say at least it was 20% jumpy productivity, at least my persoNPAL productivity in one week.
Tobi: And, um, would you differentiate between, let's say, one off things like. Small like bootstraps. Mm-hmm. Uh, like I often use, I don't know, she PT to just, I don't know, connect Bamboo HR to Slack or something like that. Tobi: Right. Like, it's super helpful. It's super quick. Mm-hmm. Mm-hmm. Mm-hmm. Challenges with AI in Complex Projects --
Tobi: You get like meaningful results, but as, as soon as you have legacy, it gets harder. Right. Also, context window is a problem. Um, how do you manage that? Like do you tell all your engineers to use ZA in the day to day and mm-hmm. Uh, what's your experience with like bigger projects when you really have. Tobi: Uh, like some legacy.
Ivan: So, you know, I think it's, uh, it's actually also surprisingly good in, uh, in also modifying the existing code basis. It does have some. Particularly bad habits. So, you know, for example, you need to ask explicitly sometimes not to delete existing functioNPALity, which is a bit funny. You know, it flips on and off, does it do comments and so on.
Ivan: But generally it can modify existing, uh, existing code base, uh, at least the last few months. It's quite effective of that. Um, I think that the bigger, bigger challenge is more on the. It's that it's this balance. AI's Randomness and Managing It -- Ivan: You know, I think about it as a, she's a phrenic senior. That's actually my mental model of, you know, of ai.
Ivan: So, you know, it is quite random. Uh, if you, you know, it, it has this randomness baked in. I mean, that's, that's how it's done. It's, it's, uh, samples, uh, distribution, right? So, so it, uh, actually can take with a sanc where it can go in quite a different directions. Um, it can provide you a. Um, you know, a solution that misses on something important.
Ivan: Um, and, uh, at the same time, it'll do whatever you ask it to. And so, you know, it can do completely crazy bloated code. And so, so I think about it much more as, um, you know, as, as a highly competent and knowKusalicable. Senior who is super inconsistent and does a bit of, you know, a bit of random things and you need to manage that, uh, you know, that, that randomness. Ivan: Hmm. Real-World Examples of AI Limitations --
Ivan: Lemme give a concrete example of something that, you know, it wasted 10 hours of my persoNPAL time. Um, so I was trying to do, you know, the, the messaging kind of notification part, you know, between my iPhone and, and, uh, backend server and, uh, you know, it, I was using for the app, uh, expo with, uh. How is it called?
Ivan: Manage Workflow? And it, you know, claimed actually both, you know, both AI that I use, which is, uh, Claude and Charge gt, uh, claimed very confidently that it cannot be done. I cannot use the, you know, Apple's main system app, uh, AP. S uh, something like Apple Network message, whatever. Uh, I cannot use it directly from my setup, and they were both claiming it.
Ivan: And so I started implementing, you know, instead of using, uh, AWS with uh, SNS and, uh, you know, directly to connected to the Apple Network, I started, uh, adding a completely different managing, uh, messaging system, which they suggest both APIs, both AI suggested, and um. And it, they were really sure I cannot do what I wanted to do.
Ivan: And then 10 now is seen by randomly reading on a blog post about something else that I was checking. I found it's possible. Uh, and then when I said it, you know, I said, it's, uh, possible according to this and this, and then, you know, uh, I said like, oh, sorry. Of course. Uh, and, uh, you know, but it's interesting to understand what's happening.
Ivan: So basically. This feature that what I was trying to do is relatively new. I don't remember how much, but it's relatively new. And so there was probably a lot of content in the internet where it was claiming this is not possible, but it's possible as of recently, you know? And so it was super confident that it is not possible, but of recently possible.
Ivan: And so I think it is really highlights, one of the examples of the, of the challenges of actually using AI because it'll be super confident in a. Wrong direction. You know, another example is, uh, with security. Now, when I was working on this, uh, tiny block thing, you know, I asked it to look for, uh, security improvements, and it suggested things like, you know, um, uh, from, uh, you know, encryption, uh, you know, enforcing HTPS, even rate limiting and things like that.
Ivan: But it completely missed, you know, uh, user input sensation. And so you can see that the other thing made sense, right? I mean, okay, rate limiting for my blog, you know, questionable. But the other thing is definitely made sense. Uh, but, uh, but then it missed user input sensation, right? So this is why I think about it as a kind of sheso phrenic senior.
Ivan: It is really capable, but you really need to guide it and you cannot just, uh, you know, trust, uh, anything that's done, but productivity. If you're guiding it well, I think it does definitely spike very, very significantly, Tobi: but very, very significantly means like by 30% or something. Right? Like not, Ivan: but it, uh, so that really depends what you're doing. Setting Up AWS Infrastructure with AI --
Ivan: So Yeah, for example, um, so as part of this persoNPAL project, I needed to set up, uh, also, you know, my AWS infrastructure, I chose to use, you know, AWS as a cloud, right? And, uh, you know, and there are many pieces there. You know, um, first, uh, VPC and, you know, route 53, uh, then, you know, I'm also having cognito for, for identification.
Ivan: You know, there is API gateway, I. Then there is ECS, uh, and with, and fargate then, uh, SNS, uh, DynamoDB, you know, so it's, it's not a super, super trivial set of services that I'm using, right? And, uh, I, you know, given this persoNPAL project and I actually wanted to code, I, I set up all of this just in a console, uh, not in infrastructure as a code.
Ivan: And I did this in less than 10 hours. Now, if I do this as an engineer in Consul, it would've taken me much more but worse. I would have almost certainly ended up in some rabbit hole about not understanding how is something in v PC happening, right? Yeah, yeah. And, you know, you spent 10 hours reading articles or 20 to get out of it.
Ivan: And so in that particular case, I think the productivity gain was, you know, between three and 500% Now, you know, so I think it, it varies about very, really about the topic. Tobi: But that's, that's like the bootstrapping moment, right? When you really wanna shoot something out and you use, I don't know, bold, bold def, or you use Chachi to generate, use something, some bridge that would've normally taken your days, right?
Tobi: Mm-hmm. Um, like I agree, but then when it's like constant, it's, maybe it's, I I think it's, it's a bit less. Yeah, I agree. I I don't wanna talk it down. Like I, I think it's like, I, I use it daily. I think it's, it's, it's super helpful, but, um, it's not. Like some people right now misinterpret it like they think it, it's always on that. Tobi: High that you have at the beginning. Yeah, but it, it, it really degrades, right? Um,
Ivan: it definitely degrades and that's why, I mean, we are also often super simplifying what is software engineering, right? I mean, you know, in the, I do not measure my, uh, you know, engineers' productivity about the speed of typing, right?
Ivan: I mean, that's, that's not really the criteria. So, of course we write code, uh, or, you know, actually engineers in our, uh, in management, I write actually quite little code. And until this exploration, not at all. But you know, in the end, uh, that is just a part of it. I think bigger part is, or, you know, the most complex part is managing system complexity, designing architecture, figuring out how feature fits together, not building the wrong thing, aligning with the stakeholders, qual clarifying, uh, you know, requirements and so on and so on.
Ivan: Of course, that, uh, all of that stuff. AI can help you a bit with, you know, writing some documents or something. But in general, you know, we are talking here about productivity on purely, um, writing code, and I think that they're, I would say it's by now probably more than 30% on average. But if you look at the total role, you know, of course then the, there are all this other stuff where, where it's not about ai. The Evolution of Tech Stacks --
Tobi: I think the key question is right now, from my perspective, like does it, um, create like a new layer of complexity, right? If we mm-hmm. Look back in, in, in, in time and what, what happened in the, in the last years? I mean, you, I don't know. You, you never look back and, uh, interpolate the future from the past, but, um.
Tobi: I think it's, there are some patterns that are repeating. Mm-hmm. And I think like 1, 1, 1 example would be like when front end and backend got divided, right? Mm-hmm. Like when all of a sudden, like big companies told us, and also the user experience was a bit better. I. Um, hey, you have to use a single page web app.
Tobi: You have to build your front end and, and in, in, in JavaScript and everything has to be, uh, single page. Um, and it's different front, front end frameworks from backend frameworks, right? Like I'm, I'm a dinosaur as well. Like I, I use Ruby and Rails a lot. That was really like really, really super productive. I still think it is super productive.
Tobi: Yeah. Um, and, and then like all of a pa all of a sudden it got destroyed by. There's like, especially like big companies that, um, that that pushed it a lot. Uh, and people follow big companies a lot. Um, and it also has some reason to exist. Mm-hmm. So that exploded really like, um, every week there was a new build system, new dependency management tools, um, new.
Tobi: New, new frontend frameworks, like a lot, um, like everything that was there was like almost broken and no longer you, you felt you can't, you cannot use dec anymore. Right. Um, and, and is it a bit the same here, I think? Is it in Ivan: just on a different level? I think it might be, but if I zoom out, I think this is how we behave as industry.
Ivan: I think we have a terrible memory and we go through hypes, uh, you know, circles and cycles all the time. You know what I mean? We invent the same problems and then invent. So same solutions, like we discovered a new thing, you know, so package management, I dunno. Like for example, you know, Golans journey through package management.
Ivan: I'm like, Pearl solved, you know, solve some of those problems. I think before nineties seven, right? So it was in know, 1980s instead of 1980s, if I remember something like that. And then we rediscovered that, you know, uh, just, uh, committing your dependence is purely in, uh. In a code base, you know, is, uh, is maybe not the best again.
Ivan: So, so I'd say that we as industry, generally we do this. Uh, I mean, another thing is that, uh, depends what you're optimizing for, right? So I think that, uh, tech stack is a very complex trade off and so I think we go from one extreme to the other. Exactly. Because both things make sense. They both have pros and cons and so that's why I think we as industry quite a bit ping pong.
Ivan: You know, if I think about Ruby, I mean it's one of my favorite programming languages. I mean. Ruby is a beautiful programming language. You know, my two favorite programming languages that I use very extensively are Scala and Ruby. Um, and I mean. If I compare JavaScript to either of them, I mean, like, I feel sad for, for the US as industry that that's where we ended up as a, you know, as a main tool. Ivan: Uh, right,
Tobi: right. I, I mean, if I, if I look at it today, like, uh, I don't know the, the whole no, no share as ecosystem, like everything feels broken still, um, or a lot feels broken still. And I would still prefer to use it like if I, if I was not coding myself, because I just think like you find. Better people.
Tobi: It's gonna be easier. You don't have the like, two different languages. Right. Um, and ultimately, uh, you want to have a different framework for, for front end. Um, and, and, um, yeah, that, that like is really interesting. Right? Uh, like when you think like, oh, uh, ah, database structure, how you manage, uh, migrations, how do you do this or that. Tobi: Like, it's just shitty. Right.
Ivan: Exactly. You know, and it's, uh, it's quite funny that actually. In a sense, uh, and, you know, connected back to ai, I mean, the AI is, uh, the best at this is anecdotal at this. I didn't try myself, but everybody seems to agree is the best of the most commonly used languages. And that also makes sense because of, you know, training, uh, uh, training set size.
Ivan: But, uh, you know, is that the best choice? Uh, that's completely different ticket. I think it just now. Uh, direction where more and more people with Japanese, because actually AI will be pushing them towards that as well, because that's where you're gonna gain more productivity. Um, and you know, in a sense, I mean we discussed this a bit before the call, but, uh, is the, will ai right, the most engaging novels with the most beautiful and outstanding writing styles?
Ivan: Probably not. I mean, it's. Trained on internet, and of course a bit curated internet and everything. But, uh, you know, it's not trained on post revealing, uh, you know, uh, books only, right? It is, uh, trained across the whole spectrum. And so in a sense it does have a bit of this, you know, regression to the mean in a sense, right?
Ivan: And so that we see also in text tech choices. Uh, and yeah, for better or worse, that's the world we are heading. And, you know, I'm just happy. I'm not, uh, in some aspects in some of those things. I'm happy I'm not, uh, engineer anymore. You know, because as a manager I would, you know, I would, uh, uh, not easily introduce Scala in my organization.
Ivan: Let's go that way, right? 'cause the trade offs are so crazily bad in some aspects, you know, from, uh, uh, talent, uh, you know, available talent from the onboarding, retraining, all kind of stuff, right? On the other hand, as an engineer. I mean, you know, the, what I could do in s Scala compared to what I could do just in Java or in even in Ruby, but definitely something like Python.
Ivan: I mean, uh, and my productivity there was definitely two, three times, uh, higher. So, you know, we talk about 30% or 50 or whatever in ai. I mean, I think that, uh, in a, you know, with, um, really advanced tech stack, you can get that just by the tech, tech choices. But, you know, the, the trade-off is that, like, as a manager, I mean, like, I would, I would not easily, again, aside from some special application, said, okay, now I want to have 200 engineers, everybody using s Scala.
Ivan: I mean, that's, uh, that would be a fun transition in Berlin. Tobi: Absolutely. Absolutely. Or Alexia or something, right? Yeah, exactly. Ivan: It's gaining of attraction. Right. But, uh, but yeah. Tobi: Yeah, rust. Rust. Like is the new, the new thing. Right. Uh, but, but, um, um, stepping back to ai, like, um, you have 250 engineers, which by the way is, uh, very little for, uh, a 5,000 people, um, organization.
Tobi: But I guess you also have lot, like a lot of solar installation, et cetera. Mm-hmm. Um. Adopting AI in Large Organizations -- Tobi: So, um, what are the challenges in, in, in AI adoption, like mm-hmm. And how do you actually manage AI adoption? Is there some strategy like, hey, do you have AI ambassadors somewhere like that, feed it back to the teams? Or like, how do you, how do you manage that, like in your engineering org?
Ivan: So, I mean, I can tell you where we are because we're currently in the middle of the journey. Uh, so, you know, where we are is we have run pilot. I've run pilot in, uh, since December with two teams. Uh, you know, and basically I selected two teams where I don't have any, you know, compliance risks and so on.
Ivan: They don't manage any PI data and so on. So I kind of selected two tips where I can. Really experiment and let the team select whatever they want. We would just pay for it, you know, whichever tool they wanted. Some people use one LM, some others, some use it with editor, some use it in a chat form. But generally, everybody really, um, really enjoyed it.
Ivan: And almost the, maybe the main learning there was that it changed the, you know, many people were already using it. I mean, I selected a team that was enthusiastic about this. But that now, they were not hiding it in a sense, you know? Now this became a thing where people would talk and share, share tricks rather than okay doing it on the side kind, almost feeling guilty that they're not developing themselves, right?
Ivan: So I think that that's a, that was a big change. Uh, but yeah, uh, after that we are, you know, now in, um, kind, um, I've, uh, identify, you know, uh, additioNPAL outset of those teams, uh, group of, uh, enthusiasts. We have, you know. Uh, uh, some informal, uh, channels where we chat about each shared ticks and trips. You know, how do you set your cursor rules to have a better, you know, uh, uh, be battery cursor, sharing the trip to, uh, what tool they use, who for prototype and so on.
Ivan: So that's kinda, let's say, you know, enthusiastic group. And really what we are discussing is. How do we do this on, uh, organization wide? What is clear is that everybody should have access to the tools. You do need to have some, you know, compliance, uh, boxes checked off. But really for me, the main question and naturally is like, how do you not let your code base growth tumors all over the place because you have not checked, you know, in the end, where are you doing what in ai, because AI will do whatever you asked it, right?
Ivan: So. If you ask it to duplicate your code somewhere because you, they are, you are not checking somewhere else where it exists and manage dependency, it'll happily do that for you. If you ask it to, you know, uh, uh, rediscover some dependency or implemented, uh, the different it's gonna do. So if you don't, you know, I think people in industry call it stabilizing ai.
Ivan: So if you don't, uh, stabilize it, for example, and narrow it down to your set of libraries. Right. It's just going to randomly choose an alternative library and you're gonna end up with three times more dependencies and so on and so on and so on. So the really, the way I'm thinking about this is that the key challenge to solve on, uh, a larger organization's, what I call structural intent.
Ivan: So basically, where do you want to make a change? And under rich conditions. And so, and I'm thinking about that as a, currently what we are discussing is kind of very simple, you know, basically a DR, so architecture, uh, decision, record, kind, implementation, and using it a bit more often than normally to signify, okay, in this part of the system, the change should happen.
Ivan: So kind of to narrow it very, very, it happens. Then additioNPALly to that, I mean. Um, it's also kind of providing a context that's specific to the project. Uh, think about it as kind of exactly, now here narrowing the standard library's choices, you know, programming style, you could even encode here, definition of done.
Ivan: And so I think that this kind, um, two things. Structural intent plus, you know, shared context of how we do things to, to stabilize the ai. Um, plus code reviews is what, is what's, uh, gonna be critical. The piece I understand the least at the moment is really code review. Um, it's clear that AI can also help there.
Ivan: You know, one simple trick is, you know, you ask to summarize the change from diff and you know, you can basically even see how, you know. How meaningful is the summary can tell you a bit about code. Uh, it's a good proxy because if the summary is complex all over the place, you also, you know, are having, uh, divergent concerns there.
Ivan: But really the question is how do you make. Sure that each part of the code base is actively reviewed. And so each change is actively reviewed both for the people to understand what's happening on a code base, but also to know guide AI and fix issues. And so in a sense, you're getting a magical tool that can produce more code for you.
Ivan: But on the other hand, now that means more code for review. Uh, and in code reviews are tricky anyway, right? So, I don't know. Do you face challenges with code reviews? Actually, I think it's overall a challenging topic. Tobi: It is, maybe there should be like some layer, I mean, you have the bottom up layer with, with cursor, basically, right?
Tobi: Like every engineer can use it and can produce way more code and the output is maximized. Um, and, and maybe you want to have like a top down layer as well, which runs on top of your gi, which, uh, yeah, just like. Is, is a, is a control layer around like dependencies, uh, uh, around, uh, like code reviews generally. Tobi: Like who reviewed this, who did this? Like, um, maybe that's still like a missing, missing piece right now, or exactly. I don't know if it's missing, but
Ivan: No, I think this is exactly the way to think about it. It's both bottom number to down VI, but then. There is completely this human factor, right? So, you know, Tobi: yeah.
Ivan: Uh, you know, we can all agree that, uh, you should write tests, right? But then there are, it's not super easy to diligently in a hundred percent of the cases write these tests, right? Yeah. Actually, AI can help a bit with that, uh, although it has its trade offs. But, so, so in a sense, I think the, the human part here is like, how do you make this, uh, uh, you know, adoption, exciting.
Ivan: Productive, uh, with some guardrails, but you also don't lose control. Uh, and so David, this is, uh, you know, I have a feeling this part of conversation will really not age well. Uh, but, you know, uh, currently, uh, currently that's thinking about it, right? Yeah. Tobi: Maybe I call you in three months and Ivan: you know, if you call me in three months, this is, uh.
Ivan: This is not working perfectly. That means that we have actually solved it a bit as industry. So that's, that will be actually, uh, good news. Tobi: So, um, was was a great discussion. Um, and, and, um, I think we slowly have to come to the end and I think like, um, we, we, we now, um, we now touched the, like if I call you in three months in the future mm-hmm. Tobi: Mm-hmm. Reflections and Advice for Engineers --
Tobi: Um, but, but what if I call you in the year 2011? Like, what if, what if, um. Gave me like a small tip, like, uh, like we both know him, uh, he's, he's also on the Board of Empire that your little iot device that you're deploying to all the households has like a time machine feature. And I could now call you in the year 2011 when you just started as a software engineer.
Tobi: On your first, I think your first job at Reed Cube, like, and, and you could now like whisper something into younger self ears. What, what would you, what would you whisper yourself? Um, you know, I'm too greedy to Ivan: do only one whisper, so I know, I would say first, like, heart take is cool, so keep doing it. So I know, keep focusing on it.
Ivan: Um, but the faster you learn. And the quicker you learn that people really matter to be taking along the way and that that matters more than being right. I think that that would be the, the key message. So really, you know, take people along the, uh, along the journey, uh, matters more than factually being right.
Ivan: It took me, unfortunately, embarrassingly, uh, years in mid and late twenties to, to really learn that properly and so, you know, if I could speed the journey. But a few years, I would definitely be super happy about it. Hey, I'm right, I'm right, I'm right. Let, let's just say that, you know, if even as architect, uh, you know, like was a picky on correctness of the, the designs, right. Ivan: But that's, I can imagine,
Tobi: yeah. Something meant that, uh, you know, can you imagine just as fiNPAL words, um, like if you could, could give everyone here like one eye-opening mindset shift. Idea, like what would it be? So it's actually the part which between India didn't Ivan: discuss at all, which is actually about ambiguity.
Ivan: But I would say, um, you know, really there is no escaping ambiguity. The world is, uh, speeding up and it's already extremely complex. And so, you know, don't try to avoid it. Don't try to escape it, embrace it. That would be really the fundamental things, you know, develop the tools. For managing ambiguity, you know, like since you mentioned even as architect, let me tell you what, what that person at that point of time learned, you know, is like, uh, if you don't know something, for example, you can just throw it at people.
Ivan: So it's, you know, so one of my favorite ways to manage to, to, to solve a problem with ambiguity is like I make the best po best assumptions, which I understand. I write it down and I throw it at people, and now they know. Now three things can happen. Basically someone will tell you. Okay, you are right. Cool.
Ivan: We have reduced ambiguity and we have a way forward. They can tell you, no, you are wrong. But then they also need to tell you how you're wrong. So you've learned and you've reduced ambiguity and you can proceed, or they can tell you, I actually don't know. But now you have a group of people who are trying to tackle the same problem, you know?
Ivan: So that's just the one super simple way, how to manage ambiguity. But I would say really. Um, jump on every, every opportunity you have to deal with ambiguity and develop the tools to manage it. Uh, that is the fundamental shift. And I think it's gonna be, you know, in this AI age that's coming, you know, after the, all the changes that are happening in the last few years, and I know macro economical and geopolitical situation and so on, uh, change is not gonna go away.
Ivan: Ambiguity is gonna be there. Embrace it and you know you're gonna be better off. Tobi: Thank you. Conclusion and Contact Information -- Tobi: Mm. Then, um, I think like after this episode, many people will search for your blog and mm-hmm. Try SQL injections on it. Um, um, but, but where can, where can people who wanna learn more about your work, um, and the stuff you do, where can they follow you? Tobi: Where can they. Read about you. So,
Ivan: so I generally, amis used to find on LinkedIn, so just, you know, look for my name, it's, you know, Ian name, that's not super common, so you'll probably find me. And then also my name, name surname combination.com. So ivan ku.com. You know, that's my blog that exists as of AI experiment, you know, three months ago.
Ivan: And there are something like four articles there, but I'm actively writing also about my current explorations and some other topics. And so, yeah, uh, there you can also subscribe for the, you know, upcoming newsletter. Again, ivan.com and yeah, I mean you can try the, the injection actually wouldn't mind. Ivan: Let's see. Let's
Tobi: hack him. Go hack him. Cool. Thanks a lot for the discussion. Have a great day. Take care much. It was pleasure. Have a great one. Bye. Chacha, thank you for listening to the List podcast. If you like this episode, share it with friends. I'm sure they love it too. Make sure to subscribe so you can hear deep insights into technical leadership and technology trends as they become available.
Tobi: Also, please tell us if there is a topic you would like to hear more about, or a technical leader whose brain you would like us to pick. Alpha List is all about helping CTOs getting access to the insights they need to make the best decisions for their company. Please send us suggestions to cto@alphalist.com.
Tobi: Sent me a message on LinkedIn or Twitter. After all, the more knowKusalic we bring to CTOs the more growth we see in tech or as we say on Alpha List. Accumulated knowKusalic to accelerate growth. See you in the next episode. These are podcast, but put it, see it. Font podcast stars,