Darin
00:00:00.000
Viktor, I'm going to set up a scenario for you. Let's assume that you are a hiring manager in let's say 20 28, 20 29. So three to four years from now. Are you ready for this?
Viktor
00:00:11.643
But nobody knows what will be three years from now, man. But okay, let's go. Let's give it a try.
Darin
00:00:16.799
So you're hiring someone and they're telling you, okay, great, I understand the role, but here's how I actually work. I have. Someone that does all of my research for me, I have someone that does my first draft of writing and I also have someone that does my first pass of coding. it's a non-negotiable for me. I have to have these three people working for me. What are you gonna say to me as a hiring agent?
Viktor
00:00:47.973
you're saying three because you are moderate. Right? You, you don't wanna ask for 30.
Darin
00:00:52.874
Correct. Yeah, I've just got three. 'cause that's just my normal workflow. That's all I really need.
Viktor
00:01:12.836
they, if they, if they will be happy with a hundred, uh, KA year and share between them, then go for it.
Darin
00:01:21.068
Yeah. I mean, look, you're hiring me. You're hiring all of us. You're not having to go out and find these other people because I'm able to do all of these different roles myself because of my team that I have.
Viktor
00:01:30.372
in all fairness, you would probably get hired in that case because companies actually, many, not all, but many companies. Constantly insist, kind of like, hey, telling people within their company, right? Kind of, Hey, if you've worked with people you know that they're good, please tell them to consider joining. That's happening all the time. Recommendations are happening all the time, Because it's hard to find good people. And then if I go to the interview, I never tried that. Uh, now I'm imagining if I would go to the, to the interview and say, Hey, uh, yeah, I'm being hired for this work. But, you know, I've worked with John, he's absolutely amazing at, at this, and I've worked with Michael. He's awesome. Amazing. And there is Nicole. She's doing this kind of like, that's amazing. That was my team in the previous company. I think you should hire the whole team. Now, this slight rephrase of what you said, but do you think that they would always say no?
Darin
00:02:35.091
Yeah. whatever. obviously what we're leaning at here is that we're talking about bringing your own agents to your day job. Now, bringing your own, not bottle, but bringing your own device showed up initially in 2004, but it really didn't start taking a hold. Until 2009. What happened between 2004 and 2009? Viktor?
Darin
00:03:12.538
and there were also devices that came out that were based on the Android operating system. So Intel was the first big company that really spent a lot of time coming up with a standardized way of bringing your own devices to work because people were doing it anyway.
Darin
00:03:30.561
prior to that, bringing your own device, what happened? Usually you were issued a crack barrier or blackberry. Or pagers from the companies, and they were not only trying to run their business, they also had a whole sub-team. Not just managing hardware, but managing mobile devices.
Viktor
00:03:49.027
And that's even still relatively trivial example, because I don't spend much at that time, I wouldn't be spending much time that Blackberry, most of the time would never leave my pocket. But then if you talk about other stuff, like, okay, yeah, I worked with uh, windows, I worked with Quest, I worked with, uh, Linux, kind of I'm productive in one of those operating systems. Then that becomes much more important because then we have real productivity gains.
Darin
00:04:15.875
Well, here's the basis. People got used to being able to bring their own devices to work, sometimes, sometimes it would get subsidized by the company. Sometimes it wouldn't. That's fine. But now what we're trying to posit here is that bring your own agents to work. To become normal.
Viktor
00:04:35.101
Yeah, because, okay. Let me put it this way. Let's, let's say we have two scenarios. One scenario is that every employee in a company works with company design built and approved agents, And there is a obvious benefit to that because hey, build one used by many people, right? But then why are you hiring me? Right? What would they bring new to the table? If I'm going to do the same as others, okay, so maybe I'm slightly better manager, designer, what's or not. I'm slightly better at distracting those agents than somebody else, but still not such a big difference. But if I say, Hey, but mine are fine tuned for this specific work, and I know and you're hiring me because that's where you actually see, A hole in what you're doing. So it depends who you are, man. It depends whether, are we hiring a junior then kind of, what do you mean your agents? I mean, come on, come just get into the grind and do it. But if you're, let's say that before ai, and I think that this will be changing, right before ai, you were. Hiring a lead designer, a lead frontend, or lead operations, right? Not in op ops guy, you're hiring the lead, That person comes with opinions, correct? And you're hiring that person because of those opinions and experiences, so on, so on and so forth. Now, if you're hiring me, because my opinions and my experience and so on and so forth, and if all those. Things I already codified or converted into agents, then their extensions of, of all that.
Darin
00:06:28.714
that's true, but what I'm focusing on here is that the companies that aren't going down the AI path. This is the back before Bring your own device was sort of commoditized. That was a big deal. It's like, wait a minute, you can't use your own personal iPhone to connect to exchange so you can check your email. No. You first have to come into the office and do this. That's what I'm saying is now we've got the Shadow AI that we talked about a couple of episodes ago that is changing the company for better or for worse, it doesn't matter. It's going to change the company.
Viktor
00:07:06.698
Yes, but I don't, let's exclude for a second. Companies in that are mono post. Right. Kind of like who I, I have the rights on all, all oil in this country, and it doesn't matter what they do. Right. Exclude those, let's say companies in competitive markets. Now, your previous example, like Blackberry versus whatever else, right? How much that you would say in productivity influence. Productivity.
Viktor
00:07:43.935
Yeah, you'd think that sales would sell twice as much if they use some iPhone instead of Blackberry. Or Blackberry instead of iPhone.
Darin
00:07:52.916
I'm just saying when Blackberry first came on the market, it gave them ability to be connected to their customers and
Viktor
00:07:58.485
Yeah. So when, when there was the alternative was not to do it at all then. Yeah. But that was relatively short-lived.
Darin
00:08:06.345
but lemme continue that on though. Once we started getting the devices in and things like Salesforce started popping up to where we could actually have a mobile app that we could use on a device or on an iPad or whatever, versus we couldn't get it good on a Blackberry. There was, being able to be truly mobile and do our jobs was a big deal.
Darin
00:08:28.408
but that didn't happen until. People started fighting for, Hey, look, I can't use my Blackberry to do business anymore. I need to use this other device.
Viktor
00:08:37.110
But here's where it gets confusing. If you take into account your previous statement is you said, I dunno, in the future, like five years from now or whenever, right. Versus company without ai. Something in that direction right now to make that the same in a way you would say, okay, but five or 10 years because time was moving slower back then. Yeah, you're still not having some kind of mobile connectivity. Five or 10 years after Blackberry became a thing, then you would be in real, real, real, real, real trouble. But the reality is that, yeah, there was initial, not for me, not for me, but that was very short-lived. Eventually everybody started using mobile connectivity, correct. Now we have other examples where companies do not switch at all. Like, uh, no. Kind of, it's been 10 years since Kubernetes. I'm still not using it. Kind of. No, no, not for. the impact is not so big that they're completely outta business. They're almost out business. Right. But AI is such a multiplier. And talking now, five years from now, right? Not necessarily today, that, uh, unless you are in a monopole type of situation, uh, it'll be like mobiles. Eventually you will have to cave in or you'll be completely out because there'll be such a big difference. That you'll be out of the game. Completely. Much bigger difference than I don't use Kubernetes.
Darin
00:10:19.037
well, in that three to four to five years, I'm in complete agreement, but right now we're in this, in this messy middle. Sort of like how it was when people were initially bringing their own devices where shadow, it was a big deal.
Viktor
00:10:33.267
currently I don't think that we, outside of relatively small companies, I don't think that we see productivity gains, from ai. I don't think that they exist yet. They exist on individual levels. I, so if you interview. Specific persons, people, you will see stories that cannot be empirically proven, proven how we are 20% faster, how we are 50% faster or a hundred percent faster or better, and all that stuff. But on a company level, that is not happening because until the whole process, the whole machine is on board, the bottleneck stays If I write code twice as fast, but the pipeline to that delivers it to production is the same speed as before. Am I twice as fast? Yes. Is company releasing twice as fast? No. I don't think that there are many companies, there are, but not significant percentage of companies who can actually truly say, yes, AI makes this company. Um, really more productive individuals. Yes, because company level takes time, right? We need to change the process. We need, we, a lot of things needs to happen, right? And it's been around for a short while, but give it a couple of more years and then you will see real competitive advantages, like massive ones
Darin
00:12:01.643
I'm still thinking about today. Shadow. It was a big deal. So here's the, here's the what. I'm trying to get the comparisons. I'm trying to draw like shadow. It still happens today, but if you think back to a few years ago, shadow, it typically meant, okay, we're gonna test out and use this other tool that hasn't been approved by corporate and we're not gonna tell them we're using it. shadow AI is the same thing. I mean, technically shadow a shadow AI is a subset of shadow It.
Darin
00:12:29.558
But it has a bigger, bigger problem, potentially a bigger security and governance problem than ever before.
Viktor
00:12:35.172
yes, definitely. Think of shadow IT for a second. Forget about AI for a second. And there are two questions I have for you. First, why does Shadow it exist? And I'll give you the second question after you answer that one,
Darin
00:12:49.422
The reason why it exists is because I'm needing to get my job done and this thing exists that can get me. To the point to
Viktor
00:12:57.736
I am more productive if I just circumvent your processes and what's not, and I just go to AWS. would you agree if I rephrase what you said and say that shadow it exists because individuals or maybe even whole teams are more productive by doing things differently.
Viktor
00:13:16.453
Now, why does it exist instead of companies saying, okay, it's not shadow it anymore. We are doing that. Why does it keep being shadow it? Because the productivity gains on the company level are not so obvious as on individual levels. You are obviously more productive in AWS, but if we adopt that and given that we didn't change some other things, we will not benefit as a company that much. You benefit company doesn't.
Viktor
00:13:52.022
Correct. Except that you can watch more YouTube because you're more productive, right? Kind of. You benefit. That's why you do things outside of prescribed methods, but companies are always into more profit and what's not. If companies would clearly see that this is clearly a better way for all of us to do it like that, it would stop being shadow. It would become. Part of what company does. Now, I'm not saying that companies are always right, they're most of the time they're wrong and they should maybe switch to AWS altogether. But the point is that the fact that you can create TC two instance much faster than going through your internal IT department or infrastructure department to get a VM for your development environment does not make application end up in production faster. Since that application is still going to go to your data center instead of AWS, right? You are just simplifying your part of the process.
Darin
00:14:49.910
I'm thinking back to. Individual contributors. Let's go back to sort of the top fake interview process we were talking about and as an individual contributor, which is what I'm talking about. It's like I've heard some people talking about instead of getting the typical questions, Hey, can you code in Russ Java? What? Go? Whatever. Doesn't matter. They're now saying in some places, can you reason with the ai, in order to code, effectively pair programming.
Viktor
00:15:17.758
Makes perfect sense. back in the day, and this is like, I dunno, 20 years ago when I was hiring. I would, let people, I, I would not say anything against them using internet, and this is like 20 years ago or something like that, maybe 15, to actually figure out how to solve whatever I gave them to solve. Only one person used it and that person was hired. I'm not saying that that's a correct way to, to hire people, but the point is that you're looking for a, for a person who can solve problems for you or persons who can be productive. Hey, if you're Googling, it'll get you there faster. Why wouldn't you do it in an interview One might say, yeah, but you don't know whether they know something or don't. And I would say that that's wrong because knowing what to Google is a differentiator as well. An experienced person will find the solution faster on internet than inexperienced. Would you agree?
Darin
00:16:22.413
Completely agree because what we are doing, most of us, I'm not gonna say all of us, but most of us is not brain surgery. We are not flying a jet airplane. We are not launching rockets or landing rockets. In this strange world of 2026 where we can actually land rockets, right? Things that could impact human life, right? Most of us are not doing that.
Viktor
00:16:49.162
Yeah, but even if you are kind of, so let's say, yeah, you're flying a rocket and let's say that we solve the speed of flight problem and actually you're an astronaut in a rocket and you get almost instant access to internet. And there is a problem. should you know what to Google? Yeah, but shouldn't you Google instead of trying random solutions? Because nobody knows everything. Of course you should. I would even say that engineers who rely only on their knowledge to do the work instead of Googling for solutions and then applying the knowledge to pick the right one and modify it, and so on and forth. Our bread engineers.
Darin
00:17:30.427
I've also heard some really, I'm gonna use two s words, and they're not that s word. strange and stupid scenarios that are starting to play out in the corporate space to where they are treating quote unquote digital talent, the same as human talent and putting 'em through onboarding, performance reviews, training. But okay, we laugh. Isn't that what we want out of things that are doing work for us? And I'm saying things as a broad thing, whether it's human or digital.
Viktor
00:18:05.467
course, I mean, I see AI in certain aspects as a junior developer, very, very skilled. The junior developer who just joined the company, maybe not even join. It doesn't matter. You just join the company. You know nothing about that company. What? What is the first thing we do? We train you, man. We teach you everything. We do everything that is specific to us, and there is no way that you could possibly know those things before you join this company, right? Now, of course, if it's junior, kind of, we teach you, teach you Java as well, right? But let's say it's not junior, it's kind of somebody above junior, okay? So you know Java, you know Kubernetes, you know, cloud or whatever the job description is. Cool. What should you do for any near term, first future, maybe even long term? You learn everything else that is specific to this company, right? That I feel that that should be the correct approach, especially in bigger systems. When somebody joins, we assume that you know what you know, and now let us teach you things that you don't know that are specific to us. We will have to apply that to ai.
Darin
00:19:11.929
Do you think we can actually teach AI faster than we can teach humans? I don't know that we can.
Viktor
00:19:16.934
Actually right now, we can teach AI faster than we can teach humans without a doubt. I mean, give, give it, give it any source, give it a hundred page document and it's, and, uh, a test with questions to see whether it understood who's gonna learn it faster, human or ai. Now the AI has a different problem today, and that problem is that it is forgetful. It'll forget everything you taught it. The moment you reach the, the, the, the context limit, either you clear it and then everything is forgotten or it's compacted, meaning that most of it is forgotten. And that's a problem we haven't solved yet. We don't know how to teach it. Teach models to be more precise, uh, without generating new models.
Darin
00:20:08.001
Which we may have to do, but if we have good foundation models that we can massage without doing rag, I mean do a, do a true tune maybe.
Viktor
00:20:18.025
I think that's, that's a problem that really inevitably need to be solved one way or another because, the whole thing about, you know, what you know, and then, kind of I train you and what's not, I don't think that that works. We will have to reach the point where LMS keep learning. Assuming that it'll be LMS in the future, we might figure out better ways to do this, but whatever it is, LMS is something better than you. The point is that it'll, we will have to reach the point that actually has memory, like infinite memory.
Viktor
00:21:05.232
Exactly, but we have actually. Uh, very, very big memory. It doesn't look like that. And you, uh, you know, you forgetting things is part of the process of how we actually retain memory. there are things that you'll never forget, right? like how to drive the car, right? The moment you learn it, you know it forever. Uh, but that's separate subject. The point is that one way or another. AI or S or models or whatever you call them, they'll need to start learning from interactions. Yes.
Darin
00:21:37.811
So let's assume for a moment we would actually hire somebody that brought their own team of agents along with them. Just assume that for a moment, the counter to that is gonna be, well, no. In our enterprise, in our organization, security and governance win all the time, period. End of story.
Darin
00:22:05.499
you'll never see AI in that type of organization unless it's completely locked down, which we'll see in probably what, 15, 20 years.
Viktor
00:22:16.716
I, I don't think that's why, you know, my first. Suggestion in that situation would be that, uh, we should open the communication between my people and your people.
Darin
00:22:26.913
Gee, it almost sounds like an API between two places to where we could communicate with people. It's, I, it sounds so strange to me.
Viktor
00:22:42.502
Yeah, kind of like, you know, innovation. I, I can get in contact with you and consult with you whenever I, I need. That's absolutely amazing.
Darin
00:22:51.044
I think another. Potential Counter argument to this is personal AI really isn't personal ai. Personal AI can't be done at a personal level. it could be a counter argument. I haven't heard it yet, but I could see that because I do believe in personal ai. I believe it can be extremely powerful, but do I want it an organization?
Viktor
00:23:12.104
you hire a secretary. I mean, not many people do that, but you know, we are old people. We, we, we remember secretaries, right? We, and that secretary can be personalized, right? You, you can. You can explain to that person that Yeah, if you, those are the types of emails you can answer in my name, right? and those, you know, if you see those types kind of just tell 'em to, to never come back politely. Right. and you can answer those calls and, uh, you can say that I'm not in the office, uh, in those cases. Right. So you, you, you, you can instruct. A secretary, how to work based on, uh, the way you would like that work to be done. Correct?
Darin
00:23:59.684
How about this one, the intellectual property problem. You've hired me with my, my team of agents. I've been there for five, eight years. Everything's dialed in, so my agents are still my agents, but they have been tuned and trained to the company way of doing and thinking, and now I decide to leave. Now what?
Viktor
00:24:26.348
I'll let, I'll let, I'll, if I make an analogy to, to human team, kind of like, I mean, I don't know, do they have the right to leave those agents? Do we give them the choice, kind of like to get them to be employed in a different company? Right. Of choice.
Darin
00:24:46.278
neither of us are, are lawyers, right? We, we can't answer this, but to me that that is a sticky situation. It's like, okay, I had my models that were trained in my way of working. Excuse me. My agents trained in my way of working,
Darin
00:25:00.348
taken my agents and I have massaged 'em. Over time, as I was learning to work within an organization, I was teaching my agents to work in that organization. Now I'm leaving. Whether I'm leaving for on, on purpose, for, for myself or I'm being released, what do we do? It's like how do you untrain something that's been trained for
Viktor
00:25:19.672
I can argue that you know that the property that I bring to this company that I had before I joined, that stays my property, right. If I come with, and I can work with my laptop when I, if I quit, my laptop, stays my laptop. Now the question is. Were those agents extended while I was there through the work in that company, then it becomes tricky, right? Because, hey, I brought my laptop, but I installed this company software on that laptop, so if I want to leave with that laptop, I would need to remove those things, or probably I was not allowed to install it in the first place.
Darin
00:25:58.804
Which leads me to the final counter argument, at least that we're gonna talk about in the Bring your Own device. Okay, great. We have a policy in place. You can bring your own device, however you need to install these 15 applications on that device so that we can lock down your device that you are paying for, that we are not subsidizing for you.
Viktor
00:26:20.404
That's why bring your own device was never a good idea, man. The much better idea was yeah, give you, gimme your company, uh, device and I will not use it. Then I bring my device and fill it crap that, uh, I didn't want to have that in the first place.
Darin
00:26:40.652
Again, the counters all have their places. I believe that last one was, oh, you won't put stuff on the thing I'm paying for. Then you need to give me a new device. Or you need to gimme money to go get another device that we can put it on there and then I'll give it to you after we're done. But it's, yeah, you, you're not getting stuff. Also, it goes on a separate vlan, which is why it's important if you're working remotely, don't have just a single land at your home. That's important. That's another episode we can talk about for another hour. We won't do that. So let's get to the hard questions as if these weren't hard questions already.
Darin
00:27:15.594
Those were the softball questions. Yeah. Uh, if a candidate shows up with their own custom AI set, much like I would probably do today, Viktor, I would imagine you would probably
Viktor
00:27:28.138
I would definitely be doing that. That would be my requirement. Simply, I dunno how to work without my people. I'm useless without my people. Now
Darin
00:27:37.644
So as a company, are you hiring one person or going back to the earlier, are you hiring four people or 10 people?
Darin
00:27:46.014
You're hiring the team. So how do you compensate? Because you don't compensate the rest of the team, you're just compensating the individual.
Darin
00:28:01.525
your team is cheap. Okay. That's an interesting question. So let's stay there for a second 'cause that comes up a little bit later. Do I say, okay, you can get me for 50,000 a year. You can't get me for 50,000 a year, but you can get me for 50,000 a year plus benefits, and then you need to give me an extra $350 a month in stipend to pay my team or. do I say, no, you can get me for 60,000,
Viktor
00:28:29.594
I think it's easier to say my value 60,000. Yeah. Kind of like, you know, you buy my team. For 60, or you rent my team to be more precise for 60 KA month, and, uh, I figure out how I do the accounts with, with my people. That's, that's my problem.
Darin
00:28:51.728
Should companies pay a premium for those types of people that actually bring a full team with them?
Viktor
00:28:58.497
I mean it, that doesn't matter which team you're, it is not a question of which team you're bringing, right? The question is how value you value you. You are to that company, Not everybody's paid the same. There are people who earn 50 K in that company. There are people who were earn 500 K in that company, right? Oh, was it pre, before saying 50 KA month? Sorry if that's, that was lost a year. Uh, that would be nice though. Um, anyways, there are different people that get paid differently because their value is different. Sometimes incorrectly calculated in all the stuff. There are many highly paid, useless people, but that would be a separate subject. So yeah, you're, you're paying my team. What, what? Whatever. It's a negotiation between you and me. I think that I worked a hundred a year and you say you worth 30. And I say, well, if the difference was not so big, maybe we could negotiate. But thank you so much.
Darin
00:29:52.128
Here's one that's gonna be a little bit harder, I think maybe, maybe not. Let's say your AI Stack passed. All of the interview things, and now it's time to do the live interview and you bomb the live interview. Was that fraud or were you just really teaching your Stack? Does your Stack really know how to do it because you taught it how to do it?
Viktor
00:30:24.066
Well, that depends. I failed many interviews. Sometimes I consider that I fail. Sometimes I consider that they're insane. Depends.
Darin
00:30:35.826
If you were hiring, how would you evaluate a candidate's AI Stack? Do you get a demo? What, what is it that you see? What do you, what is it that you do?
Viktor
00:30:46.765
I mean, you know, these days it's for a while now, I don't know whether that's still the case. It's very popular. All those exams that they give you, you know, kind of like, oh, here and you need to have special software, so you. Get disconnected from anything else. Or you go to a special room for two hours and you need to solve the problem on your own and you cannot, uh, consult internet and you cannot, I dunno what and this and that. I think that those, you know, those, those problem solving interviews, I think they're just silly. I think they're health pointless, right? They're trying to find people that are not as valuable as that they think they are. I, I feel that that goes back to what we discussed in our previous episode, kind of are we trying to find somebody who is really good at typing code? Well, if that's what you're looking for, well done.
Darin
00:31:43.309
Here's an odd question. Most of these have been odd, thinking in, 'cause in 2026 we had the Winter Olympics coming up in Italy, if I remember correctly. do you think we'll see any kind of AI doping scandals to where somebody was using AI in a forbidden context? And it gets found out.
Viktor
00:32:13.661
so. I am pretty sure, I, I'm pretty sure that AI is used, but not in that sense. Right. Uh, I, I would be very, very surprised that Formula One is not being analyzed by AI to improve the car and then tips to the driver and all that stuff. Uh, it would very surprising for me that that's not already happening. Right. now in Winter Olympics maybe for ski jumps. You know, if you, if you just kind of like half a second later to kind of lift off, I dunno, something, but that's not different from what was happening. Always. Just now better, faster. What's not, probably not, doesn't count. It's dope doping, right.
Darin
00:32:57.969
Okay. Let me ask this question. We've sort of talked about it, but I wanna ask it in a little bit different way. Let's assume your, one of your star employees decides to leave.
Darin
00:33:11.229
well that brings up the next point, but we find out the actual star power of them was actually their AI Stack. And not them themselves. Then does the company get to keep the AI Stack again? We were talking about this
Viktor
00:33:26.323
and nobody. Excluding the future that we don't know what it is. AI is not the differentiator. I mean it is compared to those not using it, right? But kind of everybody has the same models. If I was a star truly compared to somebody else and we both have access to tropic like Opus model for example, right? Or whatever we're using, then there must be. That I'm doing something different. And that's, you know, again, the orchestra conductor type of work. So if I'm a star and we both have the same tools, then I'm doing something different. That's what makes me, makes me a star.
Darin
00:34:09.684
you think this time where we can actually use AI to do interviews or actually do use it as part of our day job, do you think this is the biggest shift in hiring since maybe, I don't know, the printing press? I mean, what is, where does it fall into this? Because every new piece of technology has helped the human get things done. But thinking over my past 40 plus years working,
Viktor
00:35:06.460
I am not sure it is bigger because. We know where Internet's got us and we know where computers got us. We don't know what, where AI will get us. I probably wouldn't say bigger. Probably a similar order of magnitude, because you know, if you compare it with past. Improvements, then it's easy to see that this one is the biggest one. The most important ones, but simply because it's built on others, You can say, Hey, discovery of fire is insignificant compared to internet, but if you put it within the context of when it was made, then absolutely not.
Darin
00:35:52.109
and I come to you and I say, Hey, look, I created all this by myself without the help of any ai. Is that a, a badge of honor? Is it BA red flag, or C? So what?
Viktor
00:36:07.362
It's a matter of honor, like, Hey, I constructed this. This cabinet would solve by myself. Great man. Kind of. It's amazing because that obviously made you happy. It's pointless because you spend infinitely more time and, and probably even more money and what's not doing it right. So kind of, it's not something I want to do. I can just kinda, I can get the same house made in a, in a, in a month. So, but it made you happy, so Yeah. It's a badge of honor or you construct your own car, kind of like, man, that's amazing. You, you make your own, you made your own car. That's so amazing. Right? Doesn't mean much more than that. You'll not be making cars for everybody, anybody else?
Darin
00:37:00.961
I like how you said that doesn't mean much more than that. Right? I, I'm doing something for myself. That's great. But if I'm saying, if I'm working as a software developer in a company
Viktor
00:37:13.925
Oh, no. Because kind of So it took you a year to do what, what everybody else did in a week? No, let me connect to the. My previous comment about internet being just as important, what's not? Would you imagine the same saying kind of, oh, but I did this without actually ever going online?
Viktor
00:37:38.729
No. I kind of, I, I didn't download any dependencies. I didn't download any binaries. I didn't do anything kind of like I did it all offline. Kind of great for you, man. I mean, admire your hobby, but we don't, you're not. You cannot recollect this in this company?
Darin
00:37:55.130
thinking ahead to the companies, just even short and maybe midterm, like even up to five years, is AI profic proficiency going to be a thing? Are companies going to expect you to be proficient when you come in? Even if you don't have a Stack yourself?
Viktor
00:38:10.084
Yeah. Oh, yeah, a hundred percent. That, that time, I'm willing, I, I cannot say the date, but not far from now. Yes, that a hundred percent.
Darin
00:38:19.014
Do you think it would be to the point to where. Hey, you need to be able to use these just like today we see in jds that it's must know Java and C and Erlang and Assembly and cobol. By the way, you need to know all these languages and you need to, to have worked with, uh, what's the new language now? Um, oh, you needed to have worked with bun. That's the JavaScript runtime. You need to have worked with BUN for the past 27 years. these kinds of stupid things. Are we gonna say the same thing of you must know how to use Claude or Cursor, or whatever the new
Viktor
00:38:54.018
But that, that's the thing, you know, that's a wrong phrase that you did. Kind of like, you must know how to use clo. Of course you do. Everybody does. And if you don't, you just fake it and, and 'cause are you proficient then that's different. everybody knows how to use Linux kind of because everybody can log in into Tu and uh, click file Explorer. Now, that's not the same thing as being proficient in it.
Darin
00:39:24.874
we're used to seeing. GitHub portfolios, like you look at the, the heat map, if you will, and, and the deeper green the better in theory, depending on who you're talking to. are we gonna have the same thing with an AI portfolio?
Viktor
00:39:38.942
I don't know, man. My head is boiling. Now. I think that I need to redirect you. Can I bring my team? Team?
Darin
00:39:56.563
okay. My team will get with your team and we will, we'll shake it out. This gets us back almost to, what was that, uh, Disney movie, Wally. It was nice, cute robots for the first, however long part of the movie, and then all of a sudden it's just fat people laying around a ship. And I was like, I don't wanna be a fat person laying around a ship. Are you as it is? I don't need to be there any longer. let's go even one step further
Darin
00:40:21.283
in like, let's say three years. So let's, we've said one, two, and three years. Somewhere in that ballpark. This is the third year we're gonna see employment contracts, not only between a company and an individual, but between a company and that individual's agents.
Darin
00:40:40.383
It's, again, it's, it's bring your, if you're allowing a person to bring their team with them, then I
Viktor
00:40:47.797
would it be very different from paying for services, right. Kind of like Exactly. I, I can get, uh, AWS engineer. And services, and I can say, I don't need support anymore. You, you can go back to where you came from. I'm not paying for you anymore, but I'm still, using EC2 instances, right? And I'm paying AWS monthly. It feels similar, right? Kind of. Okay. So we part our ways, but uh, the service that I brought myself. I did not build it here. I came up, came with the service to here. You can continue using it. Kind of just, we need to agree on, uh, compensation. Why not?
Darin
00:41:31.805
that's one of the things that, that if we're right about some of these ideas, that's one of those things. It's like, okay, you can keep it, but I need to have a licensing fee for that thing from here until.
Viktor
00:41:42.811
What I see in myself, and I cannot say for others, but I would be surprised, time alone is that I spend more and more time creating, and I'm going to use generic term agents that can have many form from simple prompts to real agents and so on and so forth. Let's say. I see spending more and more time on working on those than actual work. Uh, because I see patterns in what I'm doing and I want to optimize how I do it better so that I save my own time. Sometimes I just do the stuff. Sometimes I work on an agent that will do that stuff, so, because I know that I will be doing this stuff in the future and so on and so forth, so that becomes part of my intellectual property just as knowledge, right? I'm just taking knowledge from here and putting it there somewhere else. And I would, the things I have now, and I'm pretty sure that will grow exponentially over time. I consider that my property. I, I can let you use it for free. That's fine. Some kind of think of it like open source model or it can be not open source. It can be kind of mine. but it's definitely something that feels very similar. Right. I, I can say, yeah, kind of. I never open source the things that I've worked over years, and I come here and I will be much more productive with it, but when I leave, I leave with it.
Darin
00:43:05.846
Well, again, I, I wanna restate, and I'm gonna read some of this 'cause I had some notes written down. Uh, if we're right in this, we have a fundamental rethinking of what talent really means.
Darin
00:43:16.286
Number, that's the first part. Secondly, compensation models will change. We're talking about paying a, an ongoing licensing fee, unified leave. if my team stays behind.
Viktor
00:43:27.405
Yes. J, sorry, I need to interrupt here. These are not predictions for this year
Viktor
00:43:48.390
no to yourself, to your calendar that we need to revisit this conversation five years
Darin
00:43:52.691
Five years from now, correct? Yeah. The other thing we're gonna need is we're gonna need new forms of intellectual property and employment laws. We just wouldn't have to if this is how things are gonna be working. Now that being said, if we're wrong, AI is nothing more than Excel. It's important, but it's invisible. It's, it's just used and who cares.
Viktor
00:44:17.700
wrong. Uh, but actually we do. Whatever AI tells us to do, we are not in control anymore.
Darin
00:44:30.191
that's five to 10. We're not there in five to 10, I don't think. one other thing that if we're wrong, corporate standardization wins. So all the standards, all the typical stuff wins and personal AI stacks are just a little niche thing that little small boutique companies will use.
Viktor
00:44:45.381
I bet that they will be both, just as there are both now, right? There are companies who hire you and because of your ideas, and they want you to actually work on those ideas, and there are companies that hire you just to do whatever you're told.
Darin
00:44:59.426
How important is it for somebody looking for a job today is to start building out their own AI team?
Viktor
00:45:04.896
I think it's, to me, at least personally, it's very helpful that I'm actually, I'm moving my experience and knowledge into some, in, into a different form that, that can help me. It's, I think it's important. in the past I would say that if you do repetitive process that is exactly the same multiple times, create a script. Always. now I feel it's the same except that with much greater vari variations. So doing the same, what I'm doing for scripts is just that the scripts still stay, but now the scope is, is is bigger. It goes beyond scripts, but for the same, same, same situation. If I'm going to do this once, probably never again, uh, I'm not doing it. I'm just typing there over there and we are talking and, and, and it is done. Right. But you see that you might wanna do it multiple times. Yeah. You're creating. Equivalent of a script.
Darin
00:46:03.846
Let's talk to hiring managers for a second. Are you taking into consideration that some of your AI regulations within your company are driving away potential top performance for your, or performers for your organization? I, I think that's gonna be for sure, right?
Viktor
00:46:21.458
It is much going to be much, much, much more complicated than that because I have a feeling that we are referring to tech companies most of the time in the, in this, in this conversation. Correct.
Viktor
00:46:36.308
Now imagine what happens to other industries, tax drivers, supermarkets. Parts of healthcare. I can even, so let, let's say we are talking five to 10 years, right? I don't imagine surgeons not doing what they're doing today. They will be doing the same thing. But let's see, what, what's the in, in, in us? The word when you go to general doctor, kind of like, I dunno what's going on. Huh?
Viktor
00:47:10.142
Primary care physician that if there is anything except giving you some pills, will redirect you to a specialist, right? I guess you have those
Viktor
00:47:20.205
that's gone five to 10 years from now, that's gone. That job does not exist. I'm not saying that people in healthcare do not exist. Just to be a hundred percent clear that specific job does not exist, and so on and so forth. So, we are actually very vested and involved and understand how those things work. And actually the impact on our profession like software is going to be actually much smaller than impact almost everywhere else. Where actually people working in those professions are not even remotely involved in shaping the whole technology.
Viktor
00:48:06.684
It is gonna be very tough. And here's the thing, all those things were happening before. Like Uber disrupted taxi industry, and you can say it, it dis disrupted it in a big way, but it's a, it's one thing that happened at a time, internet disrupted a few industries. Uh, computers disrupted a few industry, but over a long period of time. And one, one at a time. This is going to hit everybody at the same time, more or less, and that's something that never happened before. Maybe Industrial Revolution is actually the only previous example I can think of, Where actually we saw complete shifts in demographics from non-urban living to urban living factories instead of fields. People changed or had to change completely what they do and how they do it and stuff like that. It's that scale of change. It affects almost everything does not make us obsolete in any form of way. Just to be a hundred percent clear. It just, the changes everything at once. Was the movie, movie, all, everything, all at Twins, something like that.
Darin
00:49:20.150
And to the organizations, your AI policy right now is impacting your recruiting period.
Viktor
00:49:28.657
Today, it's impacting your recruiting. Soon it'll impact your business. And if, if, if you recruiting is impacted today, you're not getting the right people to deal with the impact on the business that it'll have soon after.
Darin
00:49:44.509
And that soon after we're talking five to 10 years in everything we've talked about here, probably some can be a little bit closer. But if you think about big companies anyway, they're thinking five to 10 years ahead, if not longer.
Viktor
00:49:58.373
Yeah, a bit. Let's say I made a mistake in a past that I'm more optimistic than I should be, and I say two years for something that eventually ends up taking 10. But if you're talking 10 years, if I look at, in similar cases, I cannot imagine a world in which, transportation industry is not completely changed. 10 years.
Darin
00:50:20.845
And right now the question really isn't, is AI changing how we're hiring people? It is, whether it's because of governance or people bringing in their own team with them. That's what we want. But at some point, you've got to make a decision and which way are you gonna fall on it? I don't think there's a right answer today.
Viktor
00:50:41.117
But you know, decisions are not made by us, rarely are made by a single person in isolation. Maybe you're the one who signs the decision, but decision is, result of company policies and structures and, uh, people you consult and people that those people consult and so on and so forth, right? The decision you will make, it will be affected by the hiring decisions you have today. That's what I'm trying to say, right? 'cause depending on the people you're hiring today, the decisions you'll be making on this subject soon afterwards will be impacted as well. It all depends which people are surrounding you when you're making those decisions.
Darin
00:51:25.600
What do you think? Head over to the slack workspace. Go over to the podcast channel. Look for episode number 3, 3 6, and leave your comments there.