Skip to content

Animated Bedtime Stories, Telling Dummies What to Do, and Why Docker Lost, with Viktor Farcic

Episode 128 · September 30th, 2026 · 1 hr 2 mins

About this Episode

Viktor Farcic is a developer advocate at Upbound, runs the DevOps & AI Toolkit YouTube channel, and was last on the show back in episode 89. Whitney and Coté start with the animated bedtime stories he has started making - tech history told as fairy tales with a nightmare at the end - which he began after catching sight of himself in a mirror and deciding people should not have to watch that. From there it is what his working day looks like now, with 20 to 50 agents running at once, a coder and a tester and a reviewer who are not allowed to do each other's jobs and argue until they agree, and a private YouTube channel where the agents post recorded tests for him to review on his phone in the morning. The last stretch is money: why he thinks AI is cheap next to what a person costs, why he expects a bloodbath as industries fail to adapt, and a walk through Docker, Chef, Puppet, Ansible and HashiCorp on whether an open source company can ever turn a profit on its own.

Check it out in YouTube, if you prefer that kind of thing.

Mentions:

Special Guest: Viktor Farcic.

Show notes written by AI with a human's lazy-look.

Transcript

Whitney Lee 00:02

Coté. Okay, so we have AI in the mix. There's so

Coté 00:05

Yeah

Whitney Lee 00:06

mm, uh, you know, uh, they say AI's an amplifier. So you and I, we're content creators. Like, you make videos, you have newsletters, you have articles. It seems like you're making stuff nonstop. And even before AI, it seemed like you were making stuff nonstop.

Coté 00:20

It wa-, it, it is possible. I've proven it

Whitney Lee 00:23

So now that you have AI as an amplifier, what is different? Are you doing more things? Are you doing different things than you ever thought you would?

Coté 00:33

Hmm.

Whitney Lee 00:34

combination thereof

Coté 00:37

I think it's a… You know, I don't know if I do more things. I think, I think the things I do, I do better. And where better is defined by, by, uh, like, like let, let

Read Full Transcript Hide Full Transcript

Whitney Lee 00:54

'cause it's some sort of a- admission that what you made before was kind of bad.

Coté 00:58

Yeah. Well, I'll, I'll, l- I'll… In, in this case, better means it's, um, enhance is the wrong word. It's enriched. It has… So for example, when I make-- when I used to make show notes for a podcast or a video,

Whitney Lee 01:12

Mm-hmm.

Coté 01:12

I would not 'cause, like, I, I would be, like, sick of… I-- By the time that I was to the point where I was making show notes, I would be like, "I wanna be fucking done with this. All I gotta do is press this button." And so I would just, like, not do it. Whereas now, I can, uh, you know, I can have the robot make five versions of show notes, edit it, and I can pick one, and then I can edit it a little bit. It can, like, gather up all the stuff. It can make, like, chapter indexes of stuff, and it's great. So, like, I have, I have more stuff

Viktor Farcic 01:42

You're such a disciplined person. The fact that you're picking up one of those five is amazing to me

Coté 01:48

Well, you know, I, I, I, uh, maybe that's left over from when I, when I manually did stuff. Now, now you're, you are correct. Sometimes I'm also like, "Fuck it," and I, and I just, I do a very minimal read over to make sure it's not… Th- I, I would say I'm doing a read over to make sure it's 30% of, of my, my, my quality control is that it's not terrible,

Whitney Lee 02:11

Uh-huh

Coté 02:13

is that it doesn't say something people will, will hate me for or I'll get in trouble for, right? And so, like,

Whitney Lee 02:18

I

Coté 02:18

it's

Whitney Lee 02:19

realized too that there's a third option that I didn't consider, which says more about me than maybe it does about you. But the third option would just be doing what I did normally, but now I have more free time. Does,

Coté 02:31

Yeah, yeah, yeah

Whitney Lee 02:32

anyone choose that option?

Coté 02:34

I th- I think so

Viktor Farcic 02:36

I have, I have a better one. Kind of to me it's, it's not really about amplifying me or, uh, being better in control and all that stuff. It's about being able to say to other people that I'm dependent on, "Fuck off. don't need you anymore." I'm willing to ac-accept worse results, but I don't need you anymore, right?

Coté 02:58

I think, I think, you know

Viktor Farcic 02:59

you to review my PR. Well, thank you, Greptile. Bye-bye.

Coté 03:04

I think, you know, I think, I think, I think both, both of what, you know, what both of you are saying also applies. And the, and kind of what you were saying, Whitney, is like, is like, I think-- Let me put it this way: I don't, I don't necessarily do more, like you're saying. I kind of do the same amount. I just do other stuff that I wasn't doing previously, which sounds like more, but there's a distinction. Like, I'm not necessarily, like, putting out more episodes or more content. There's just other little side projects and things that, that I do that I didn't used to do, like B and C or C-tier level stuff. And, and then yes, like, like, like our guest Viktor is saying, there are a lot of activities that now I, I guess… Well, I used to tailor my activities already using the, the, the Viktor philosophy, which is: How could I go through my life with as few human dependencies as possible? And, and I think, I think I have-- I think about, uh, almost after 30 years of being in IT, I'm highly optimized for that way of operating. And so now-- But I think there's still more to be stripped out. There's like, you know, less review processes.

There's especially, like, less waiting on people for things, 'cause I can just get the thing, and like, so there's, there's that kind of thing.

Whitney Lee 04:25

Viktor, will you introduce yourself please, first of all, and then tell us how AI has changed the content that you make

Viktor Farcic 04:32

Okay, two questions. Introduction, yes, I'm Viktor. and what was the second?

Whitney Lee 04:38

That, that was not so helpful.

Viktor Farcic 04:43

Uh,

Whitney Lee 04:44

W-

Viktor Farcic 04:44

I don't know. I,

Whitney Lee 04:45

you can look… Yeah

Viktor Farcic 04:47

have difficulty explaining what I do. I do random things really. Can-- it depends what, what I'm interested in, what,

Whitney Lee 04:53

Uh-huh

Viktor Farcic 04:54

to have done and things like that. I could be creating content. I could be working on a project. I could be speaking with a company, uh, or I might be playing games and letting AI do all those things. Uh, depends.

Whitney Lee 05:09

So you

Coté 05:10

Nobody knows

Whitney Lee 05:11

YouTube channel, DevOps Toolkit,

Viktor Farcic 05:13

very popular is a, uh, overstatement,

Whitney Lee 05:16

all relative. Yeah, okay. It's, it's a lot more popular than mine, but that also means nothing.

Viktor Farcic 05:21

my goal was to reach MrBeast numbers, but it's not happening. It seems that less people are interested in cloud and, uh, infrastructure and what so not than in MrBeast. Go

Whitney Lee 05:34

Go figure. So you do that, you run a-- write a ton of blog posts, you speak at lots of conferences, you make a lot of useful software for people, and then you work at Upbound and, uh, talk with customers and advise folks in that capacity, yeah. Um, so one of three things. So you did all those things before AI, and now you have AI. So what do you do? Do you do more of those things? Is-- do you do different things? Like how, how is-- or how-- do you relax more? How has it, your life changed?

Viktor Farcic 06:05

relax. It's not allowed. It's

Whitney Lee 06:07

No.

Viktor Farcic 06:07

Uh, no, no, no relaxing. I, I

Whitney Lee 06:10

Who enforces that?

Viktor Farcic 06:13

I think it's similar to what Coté said. Uh, first of all, there are things that I should have been doing and I'm not. Like a silly example is, okay, if I do a live stream, I should get time codes. I never did because, you know, it's work. kind of okay, uh, then, then you sit down, write an application that is scheduled to run every day to check whether there is a stream published, and if there is, extract time codes, whatever it is it is extracts, uh, and puts into the description is better than nothing. That's my philosophy, right? Because that's-- nothing is the alternative. So I do more of the th- things that I haven't been doing.

Whitney Lee 06:56

Mm-hmm.

Viktor Farcic 06:57

n- and I agree with Coté, not necessarily pr-- So I'm not making more, publishing more videos, it's just that there are things around that, I wasn't doing and now I am, right?

Whitney Lee 07:06

Uh-huh

Viktor Farcic 07:09

and, uh, and there are a bunch of projects or tools or things that I would never have time to do, and now I do,

Whitney Lee 07:21

Uh-huh

Viktor Farcic 07:21

of, oh, you know, there is a list of, you know, all the things that you were thinking about in shower that you, uh, that ended up never doing.

Whitney Lee 07:30

Uh-huh

Viktor Farcic 07:30

it now, right? Uh, so a

Whitney Lee 07:32

When you say things, do you mean like coding projects or do you mean like

Viktor Farcic 07:37

Yeah,

Whitney Lee 07:37

animated bedtime stories?

Viktor Farcic 07:39

not happening yet

Whitney Lee 07:46

Until you can garden some french fries, I don't, I don't see it coming from you.

Viktor Farcic 07:50

I mean, that's already solved. We have delivery in, in, in this, in this country, so it's, it's, it's a solved problem

Whitney Lee 07:57

So

Coté 07:57

you know, I, I think, I think y- like, like I, I don't actually do a lot of, uh… Well, I don't-- I wouldn't say I do any coding myself. I have the robot do coding, but I don't make applications. And kind of like as, as you were talking ab- as… Well, what am I trying to say? There, there is a lot of stuff that I do that is not gardening, uh, but is like non-coding stuff that I wouldn't do otherwise. Like, a lot of it's just like organizing files and finances and stuff. But then there's also like, uh, you know, I'm like trying to learn Dutch, and thanks to you, I use Anki, and Anki is fully programmable. So like, I just will like say like, "Go change this card. Go do this, and then make a po-" Like, it does all this stuff that is, uh, like I guess I could make those flashcards. I could go look up in a reference book, but I would never do that, let alone hire some Dutch person to give me like a 10-minute podcast about the difference between onder and nach and things like this, right? Like there's-- So there's this whole other realm of things

Viktor Farcic 08:58

You know when you said, "But you, you're not doing coding," that, that's wrong. You're maybe not realizing, but behind the scenes it's writing code. I

Coté 09:06

Right, right. And, and I, I, I, I agree. I, I just, it's the, it's… I, I would separate it as, uh, not even intentional coding, but like I… 'Cause yes, you're, you're, you're correct, but I feel like there is something different between I'm making a software application, uh, versus like I'm just getting some shit done, and I don't, I don't really care about the software part. But yes, there's lots of code being written, just I'm not involved

Viktor Farcic 09:32

Oh, who is?

Coté 09:35

Exactly.

Viktor Farcic 09:36

Wait, wait, wait, You

Whitney Lee 09:38

Artisans.

Viktor Farcic 09:39

still

Coté 09:40

maybe,

Viktor Farcic 09:40

type letter by letter

Coté 09:43

he-he-- Well, here, here's, here's an example is like, I know, I know in the agentic or whatever the, in the AI coding world, everyone's interested in writing specs and things like that. And I'm like, I d- I don't care. Like, the only spec I write is like on, on the line here. So I'd have no idea

Viktor Farcic 09:59

that's a spec. Do this

Coté 10:01

Fair, fair, fair enough. All right. All right. I'm, I'm not, I'm not flashing a bunch of symbols at it

Whitney Lee 10:07

So all that, all that line of questioning up top was a roundabout way to get to the fact that Viktor, I, uh, I see a new type of content on your YouTube channel, which seems to be around animated bedtime stories told by Uncle Viktor.

Viktor Farcic 10:24

big motivation for that was to check whether you actually ever saw any of the videos I ever published. okay, experiment successful.

Whitney Lee 10:33

It took me a month and you had to be a guest on my show, but yes. And actually, to be fair, Coté showed me

Viktor Farcic 10:39

thumbnail so that you don't think it's coming from me, kind of, and maybe that's how you will see it. Exactly.

Whitney Lee 10:49

So for real though, is this powered by AI? Like, first of all, it's cool. Like, it's, um, like a, like how the rise and fall of Docker is one of them, right? So it's, you're telling a story as though it's a bedtime story or, you know, and then it's animated kind of in the style of Rick and Morty, then it's, it's cool. But,

Viktor Farcic 11:10

It's,

Whitney Lee 11:12

is this an AI-enabled project?

Viktor Farcic 11:14

It's a painful realization that led me to that. And the realization is that I normally don't, you know, don't spend time in front of the mirror, but accidentally I did recently, and then I saw this, and I thought kind of, "Come on, this is not something people should be watching." Kind of let's be realistic for a, for a moment, right?

Whitney Lee 11:38

That's a testament to how good your content is

Viktor Farcic 11:41

in front of this

Whitney Lee 11:45

So is that, is that AI, uh,

Viktor Farcic 11:48

No,

Whitney Lee 11:49

yeah

Viktor Farcic 11:49

I w- I w- I was hand drawing every single frame of

Whitney Lee 11:52

Well, I know you weren't, but were you, like, do you have a creative director in the background? Like, or is it, like, fully defined by specifications?

Viktor Farcic 12:00

which part of the earlier conversation when I said I don't w-- uh, the biggest benefit is not talking and working with other people you did not understand?

Coté 12:11

Well, well, uh, the, like how, what, what's the, what's the process for doing that? Like, do you,

Whitney Lee 12:15

Yeah

Coté 12:17

how did you fig- I, I imagine you have to like, h- uh, what, what word am I… Narrow down from the world of possibilities to like, "I like this style." And then the oth- the other thing you have is you probably know more, like, but they're each on kind of like a five-second loop so that you can have like the same video playing o- which is great, right? But like how did you, how long did it take you to kind of figure out like, "And we're done." Right? Now I can just tell the thing to do it, and it'll do it.

Viktor Farcic 12:47

So th- there are a couple of things, uh, I, I have to do there or at least I think I have to do. First I figure out kind of, okay, what's the, the ambient, right? Kind of like what's

Coté 12:58

Yeah

Viktor Farcic 12:58

kingdom with a castle or whatever, right? Um, you know, I, I would generate, I don't know, 10, 20, whatever different variations of it and say, "Okay, that's the one." Kind of that's the style. Then I generate, uh, in that style, uh, a few recurring characters,

Coté 13:16

Yeah

Viktor Farcic 13:16

Kind of their, their stills. And then, then, then it starts. Okay, scene number one kind of like create a still, uh, usually a couple of them kind of, okay, that's the one. And then next step is animate it kind of, oh, let, let the chicken pick some- something from the ground and, uh, let this guy move the hand like this or, or whatever it is. and, and it goes I use Hi- Higgsfield for that which is, you know, one of those sites that don't do anything special really just give you access to different models.

Coté 13:49

Mm-hmm. Yeah

Viktor Farcic 13:50

Uh, uh, that's it, right? And it's all done with, uh, Claude or OpenCode depending on one I have credits for. Um, that's about it

Coté 14:03

Yeah, it, it's, I, I've, I've never… Well, I don't know if this will come as a shock, but I have never made a TV show or a movie or been involved in it. But it seems like from watching those extra, extra specials on, on things where they show you the making of, it seems like shockingly similar to how you make TV and movies, where you're just like, "We've got a bunch of stills. We pick out our concepts, and then the very last thing we do is the computationally intensive thing of actually filming." And then I imagine there's editing afterwards. But it's, it's, uh, it's kinda like what we were saying earlier. It's like, this isn't really changing what I'm doing or how I operate. I just do it better.

Viktor Farcic 14:45

Yeah, and you know, basically it's, it's kind of conf- conf- uh, reject or confirm type of situation. Uh, I have elaborated instructions to, to Claude that kind of, okay, so can, you know, uh, open the clip, uh, get, uh, every 10th frame, check whether actually that makes sense. If it doesn't, don't

Coté 15:07

Ah, yeah, yeah

Viktor Farcic 15:08

it again. Uh, show the different stills kind of in parallel on a screen so that I can pick one and so on and so forth. It-- There, there was a s- fair amount of work involved to make it running, you know, trial and error, but it works and I'm burning

Whitney Lee 15:25

I,

Viktor Farcic 15:26

like crazy

Whitney Lee 15:27

I was gonna ask, it seems like an expensive endeavor. Yeah

Viktor Farcic 15:32

You know, so whenever I hear expensive word in relation to AI, my answer is always, "No, it's extremely cheap." Right? Compare it like, okay, w- what, would I hire an agency to do it? Would I hire a person

Whitney Lee 15:49

just wouldn't make it. It's extremely cheap to make it, but you wouldn't make it at all, like compared to hiring a human. But

Viktor Farcic 15:57

I

Whitney Lee 15:58

you also would not make it, period. Yeah

Viktor Farcic 16:01

Yeah I, I would either not make it or I would pay much, much, much more. Uh, uh, I mean, I, I don't have numbers, but I think that probably a couple of hundred bucks goes per maybe couple of hundred, maybe 500 per episode all included.

Whitney Lee 16:17

That's, yeah.

Viktor Farcic 16:18

Uh, it's,

Coté 16:19

Well, that

Whitney Lee 16:19

that is.

Viktor Farcic 16:20

but find me a person who would actually it for that amount of money. And when I say do it with all the variations and iterations and rejections involved in it, right?

Coté 16:32

All right. And, and that's, that's probably much cheaper than the, you know, however you rate it, the overall value that, that, that you get out of it

Viktor Farcic 16:39

Yeah

Whitney Lee 16:39

W- when, when you were describing the instructions you created to make these videos, one thing you said is like, um, "Check whether the still compares to the story, and if it doesn't, don't ask me questions, just fix it." that made me realize, oh my God, I need more instructions in my, my interactions with AI that just says, "Don't ask me questions, just do it." Is that literally something you put in there often? Yeah, I need more of that.

Viktor Farcic 17:05

Uh,

Whitney Lee 17:05

It wants to… Yeah

Viktor Farcic 17:07

I'm not there, just to be clear, but the goal kind of the, know, man can dream type of situation is that I just confirm. I ne- I never reject anything. That, that's the goal kind of like,

Whitney Lee 17:19

Then why would you even confirm? Wouldn't the… If, if that's the pie in the sky, why are you in the loop?

Viktor Farcic 17:26

Okay. So yeah, let's go st- after that, yeah, kind of it goes live. Uh, actually now I'm working on, uh, instructions that will, uh, take my narration, uh, together with those things and assemble it even, right? Uh, the, the video, audio clips, everything,

Whitney Lee 17:47

Oh, is your video editor about to lose their job?

Viktor Farcic 17:52

that, that's the-- I mean, uh, yes, hopefully.

Coté 17:59

That, that's the, uh, that's, that's, that's the, the pie on the moon, uh, scenario right there. Yeah, well, you know, I, I found, uh, uh, I, I used Descript and Claude, and it's phenomenal at video editing. Like, it, it-- I, I still am in only the, uh, the pie in the sky sometimes where, where I go look at it.

Whitney Lee 18:20

yet

Coté 18:21

But I was actually thinking before recording this episode, maybe I should just, uh, you know, I'll, I'll have to pay attention, 'cause remember, one of my criteria was I don't wanna get in trouble for anything that's published, right? So

Whitney Lee 18:31

Mm-hmm.

Coté 18:32

as long as I know the, the whole thing being edited, there's nothing bad in it, then who cares how it gets edited? Well, I mean, you care 'cause you want it to be good. But you could, uh, you could probably take a pass at just having it edit it a little bit

Viktor Farcic 18:45

The thing is that, uh, there are, there are a bunch of tools, you know, command line tools like FFmpeg,

Coté 18:54

Yeah, yeah

Viktor Farcic 18:55

are absolutely horrible for you as a person using it, But they're actually pretty good for, for, for agents, for AI,

Coté 19:02

Yeah

Viktor Farcic 19:03

They can do marvels without Descript or, you know, Final Cut Pro or any, any of that kind of… They would get confused with those things. They are not built for them, right?

Coté 19:11

Right, right. And, and to your point, the Descript part is for me, right? So, so like it'll assemble it together, like comp it all together and for me to look at and mess around with. But if I didn't have to, then I wouldn't have to. It would be, uh, much better. So yes, we're on our way to the moon. That's, that's what we want

Whitney Lee 19:31

S- so once we get to the moon or, um, I guess even toward the sun now, uh, so once the-- once good practices in AI settle and, like, people figure out how to use sub-agents and manage AI in such a way that they can maximize the amount of work that gets done with m- and minimize the human oversight needed, what do you think, what do you think job roles will be like? What job roles will exist or not exist? Man, like say more. What-

Viktor Farcic 20:06

I mean, th-there are different types of management in our

Whitney Lee 20:08

Mm-hmm.

Viktor Farcic 20:09

industry. Let's, let's call it technical manager, right?

Whitney Lee 20:13

Okay

Viktor Farcic 20:15

Um, I'm, I'm trying to avoid saying specifically kind of, uh, architect or product man-- uh, technical management of sorts because, um, so the, I'm going to be, be very disrespectful here, but the monkey business that, that, that, that's going away, right? Kind of, "Oh, I can type getters, setters better than you can type getters, setters." I, I, I, I… That, that's agents now. I, I, I really don't, I don't wanna do that anymore. My job now as, uh, whatever I am,

Whitney Lee 20:46

Mm-hmm.

Viktor Farcic 20:47

Goal. Let's say goal, not necessarily job today, is to say what I want in a

Whitney Lee 20:54

Mm-hmm.

Viktor Farcic 20:54

is understood by dummies. have

Whitney Lee 20:57

Mm-hmm.

Viktor Farcic 20:57

dummies running on a machine And validate that that's what I'm getting, right?

Whitney Lee 21:07

Uh-huh.

Viktor Farcic 21:07

Uh, and specifying is a combination of I want this like that and the hardness that we build around it, right? Because I don't wanna

Whitney Lee 21:19

Mm-hmm.

Viktor Farcic 21:19

myself every single time. "Oh yeah, you know, you need to open a pull request, don't push to main," all that stuff, right?

Whitney Lee 21:24

Use TDD. Yep

Viktor Farcic 21:26

instructions are there already, and those instructions are-- There are two primary motivations for those instructions, to do the right thing and for me to not have to explain myself, more than absolutely needed, right?

Whitney Lee 21:43

Hmm.

Viktor Farcic 21:43

Um, and then validation at the end. one thing that I'm doing right now, still experimental, that I let my agents work, bunch of them, they work, let's say at night while I'm sleeping.

Whitney Lee 22:01

Mm-hmm.

Viktor Farcic 22:01

And among many others, they have instructions to when they're finished and they create a pull request and it's all validated and audited and all that stuff, they publish, uh, a video on YouTube And then I wake up in the morning, I open my phone and go say, "Feature number one, uh, this looks okay. Yeah, cool. Ship it. Feature number two, uh, uh, no, not, not, not correct. Okay, I need to, I need to correct that," right? So basically, I'm watching videos

Whitney Lee 22:33

Yeah

Viktor Farcic 22:34

um, to check the things. I don't care anymore about functional things. They're, they're all correct more or less, but user-facing, yeah, that's, that's how I validate things or at least that's how I'm trying… That's, that's the place I'm trying to get

Whitney Lee 22:52

You watch a video of like, of what your content will look like based on what they're…

Viktor Farcic 22:56

No, no,

Whitney Lee 22:56

like, no.

Viktor Farcic 22:57

now. No, no,

Whitney Lee 22:58

No. Okay. No

Viktor Farcic 22:59

talking about, "Hey, uh, develop this, uh, uh, uh," let's say working on a webpage, "De-develop this that will do that. There will be this button and this form," whatever, right?

Whitney Lee 23:11

Uh-huh

Viktor Farcic 23:12

I wake up in the morning

Whitney Lee 23:14

And it's published J-

Viktor Farcic 23:15

for me of, uh, recorded test because those are the tests and videos at the same time. Kind of like, okay, run those tests to validate that that works, and while you're running those tests, when

Coté 23:27

Make a video

Viktor Farcic 23:28

generate a video

Whitney Lee 23:30

Uh-huh

Viktor Farcic 23:31

publish it. That's a pri-private channel. You cannot see it, right?

Whitney Lee 23:35

Uh-huh.

Viktor Farcic 23:37

publish it and I'm going to see whether that looks as it should be

Coté 23:41

That's interesting. I, I li- I like that. Yeah, like it sounds like, like there's on, on the, the input side, there's like what to do, which is always valuable. And then on, on the output side, there's, there's almost, uh, like… Well, of course it's verification, but somewhere in there is like, uh, like having taste or like having some opinion about like what it is you want and how you want it to, to go. Not just verifying that like it's red or green, but that this is done in, in the way that I wanted

Viktor Farcic 24:15

That's, that's actually the latter is the, the problem I'm having rather than the former. Uh, I think we are getting closer, uh, and closer to, yeah, this-- you told me this and this is what I did,

Whitney Lee 24:30

Mm-hmm.

Viktor Farcic 24:31

Uh, it's doable. Now, I'm not saying it works 100% now, but it's achievable, right? I'm more talking about… That's why I intentionally say user-facing things. I'm more referring to, okay, uh, it's a square button. It shouldn't be a square button. It should be a round button, right?

Coté 24:48

Uh-huh.

Viktor Farcic 24:49

green is horrible, kind of. That's not the green I like. Uh, things like that. More about the taste than, than functionally working

Whitney Lee 24:59

Do you think when-- Okay, in this future world where everyone's a manager, they're, uh… think the lines in the sand will be similar? Like, like someone's a feature manager and someone's an application level manager, and someone else is a systems level manager. or do you th-

Viktor Farcic 25:19

so.

Whitney Lee 25:20

No, you think people will manage m- Yeah.

Viktor Farcic 25:24

Uh, th-this is me talking. I'm not sure about the world. I don't think, I don't… You know, th-then all those roles become a bottleneck again, again

Whitney Lee 25:36

Mm-hmm.

Viktor Farcic 25:38

I imagine a world where I'm in charge of a project or a feature. Let's say maybe there are multiple people working on a project, cool. But kind of I'm working on this feature, whatever it is, square

Whitney Lee 25:49

Mm-hmm.

Viktor Farcic 25:50

right? Uh, and I deliver, I with my minions deliver that thing, right? I imagine more people being evolved horizontally, right? Kind of, okay, so you work on this, I work on that,

Whitney Lee 26:07

Mm-hmm.

Viktor Farcic 26:09

of th-th-th-that all somehow works rather than, oh, you're in charge of the color of the button, and I'm in charge of the shape of the button, and you're in charge of a tech- text of a button. No. I, I'm not sure that that worked even before AI, to, to be honest, but, uh

Whitney Lee 26:26

So

Coté 26:26

know, may- maybe, maybe an interesting thought exercise for this is like, uh, outside of the domain of programming, like here in my neighborhood, they're, they're, they're about to, uh, let's say refactor the street. And, uh, what, what that mean… You know, whatever. So they, they got plans, and they went through all the, uh, the council meetings and et cetera, and now we've got some cool PDFs with stuff like that. And I would imagine kind- I mean, it's an engineering sort of thing, but there's probably a similar, similar effect that AI would have, and maybe even used of like, we have a street, it's kind of a d- sort of a definable system. And I think the, the issue they had was the, uh, this is like this very Dutch thing to worry about, which is to say a very not necessarily American thing to worry about, is like the trees on the other side of the street, they're not doing very well, so we should figure out how to replant them and have them do better. And there's also some… Whatever. There's some reasons, some other things. But, you know, now they have a plan. They've got a bunch of PDFs.

They've got one of those great uncanny valley pictures of the street that's like very crisp and clear, and it has those pictures of photorealistic people that don't look quite right, and they're also having more fun than anyone has ever had on this street, you know, going about their daily life. But it does, especially with the, the, the horizontal thing, I wonder if… Obviously, there's people that still need to move things around, so until you have robots. But I wonder if the, the, the council doing this could have just done all of that spec work on their own and then handed it over to someone, or if you actually needed to hire whatever you call it, an engineering firm that would like… Like they, they kind of have to know when you dig on streets. I mean, the assumption would be that engineering firm has context that AIs don't have, and so they're pulling on that context to, to make decisions that the, the, the robot couldn't

Viktor Farcic 28:23

Yeah. But I, I feel that that would be more equivalent to buying or renting. So, uh, let me bring back to software. Let's say, um, let's say it's a webpage, right? And I'm really terrible with design. I don't want to have a person designing it for me, uh, but if person can-- I might hire somebody to, to generate a design team for

Coté 28:50

Right, right,

Viktor Farcic 28:50

right. And then I use it and kind of, okay, this is designed for you. They know what they're doing within you want kind of go. Uh, so I'm, I'ma, I'ma get that person one time, as a recurring, um, uh,

Coté 29:09

Interaction?

Viktor Farcic 29:10

interaction. Exactly.

Coté 29:12

May- maybe the, maybe there's, there's a, there's a too cutesy way of putting it is like human's job is to have opinions

Whitney Lee 29:18

Mm-hmm.

Viktor Farcic 29:20

Yeah, and

Whitney Lee 29:21

to taste

Viktor Farcic 29:22

of teach, teach, teach my people those opinions.

Whitney Lee 29:26

Mm-hmm.

Viktor Farcic 29:26

that's the thing, right? Teach them. Cool. My p- my people are dummy in that regard. Cool. Bring me your people or teach my people or whatever needs to happen, I don't know, so

Whitney Lee 29:38

So

Viktor Farcic 29:40

I can use them

Whitney Lee 29:41

You say our future jobs are to be managers and specifically to be able to tell a bunch of dummies what to do. what, what advice do you have for telling dummies what to do?

Viktor Farcic 29:56

Um,

Whitney Lee 29:57

Like, like don't ask questions is good advice, and I'll be using that, telling my dummies that

Viktor Farcic 30:03

I mean, I, I don't have a singular advice. Simply, uh, my instructions are growing with time.

Whitney Lee 30:09

Mm-hmm.

Viktor Farcic 30:10

I, I probably spend, let's say, one third of my time on, uh, brainstorming and deciding what I want to do. Uh, and one third of the time while it's doing, I'm often there, but not to validate its work, more to provide-- to improve instructions, Kinda, "Oh, I s- I saw you doing this three times over. Um, maybe I should extend the instructions," type of situation

Whitney Lee 30:43

when you're extending instructions, are you doing that at a specification level or within a sub-agent or in your rules files?

Viktor Farcic 30:51

Uh, the, I'm doing that at, uh, skills for the specialized agent that I have, right? So I have agents for reviews, agent for, don't know, coding, for testing, and so on and so forth, and then depending on, uh, which one is messing it up, um, I extend its instructions

Whitney Lee 31:13

I understand that you have a, a challenger agent whose whole job is to the other, challenge the other agents

Viktor Farcic 31:22

Actually, they're all challenging each other. So

Whitney Lee 31:24

Yeah

Viktor Farcic 31:26

uh, when I, when I bring a swarm of agents to work on something, let's say there is a coder, there is a tester, reviewer, and then the coder is not allowed to, to write tests, tester

Whitney Lee 31:37

Uh-huh

Viktor Farcic 31:38

but he's not allowed to write code, so tester writes tests and says, "I'm done. You figure out how to make those tests pass."

Whitney Lee 31:47

Uh-huh

Viktor Farcic 31:47

Then coder would sometimes say, "I cannot make those tests pass because those tests are stupid." Uh, then it goes back to tester, right? The reviewer comes and says, the heck are you br- doing, man?" Kind of like, "Stop." Uh, it-- I, I would say that the way I'm working right now takes order of magnitude more time to finish something than I was doing it before,

Coté 32:14

Mm-hmm.

Viktor Farcic 32:15

they go into endless fights, right? Um, it, it takes time until… I mean, it's like people, right? Kind of if you enable people to really speak their mind, not only follow the leader like lemmings, um, then you end up with a very long process because everybody has a different opinion, and eventually over time you will probably all agree on something,

Whitney Lee 32:41

So,

Viktor Farcic 32:41

time

Whitney Lee 32:43

so then do, are you optimistic that you'll be able to streamline this process, or are the results that you're getting from this process worth the extra time watching the dummies argue?

Viktor Farcic 32:55

already doing that.

Whitney Lee 32:57

Which?

Viktor Farcic 32:58

is always improvement, but at any given moment I have, I don't know, anything between 20 and 50 agents doing something

Whitney Lee 33:05

Right. So, so the quality of what you're getting from using this multi-agent system is worth the extra time of having them quarrel with each other

Viktor Farcic 33:15

Uh, that's the thing

Whitney Lee 33:16

Yeah

Viktor Farcic 33:17

uh, time, how long it takes for agents to do something matters much more when I'm sitting and watching as, as if that's entertainment, uh, than when they're doing, right? So, uh, if a single agent would do it in an hour and the s- my swarm of agents takes five hours, uh, doesn't matter that it's five hours if I'm not really watching it,

Whitney Lee 33:45

Yeah. The single agent did it in an hour, but with you manually approving and changing and doing stuff. The swarm does it in five hours. Okay.

Viktor Farcic 33:57

Yes

Whitney Lee 33:57

The swarm does it in five hours, but you can walk away and, uh, order, order your takeout and watch TV with your family. Yeah

Viktor Farcic 34:05

That's, that's the whole point of it, right?

Coté 34:10

I, you know, I, I, I, I did a, uh, you know, I-- to call this an analysis is more of a fanciful joke, but I, I had, uh, I ha- uh, for, for my playing D&D with AI, I had to look over all of my stuff since 2023 and calculate how much time I spent actually playing versus like harness crafting. And it's kind of a, uh, it's kind of a, a delightfully depressing chart because way back in 2023, of course, the only harness was copy-paste. There was nothing else that, that you could do. But then slowly but surely, the-- I'm making this picture with my finger from the top, more and more of the harness comes in and eats away at the actual playing, right? As you like build things out and have MCP servers and skills and like all this stuff to arrange stuff. And then in, um, in a, in a good/bad way, as you're like, "You know what I would like is blah," whatever. And then you spend time talking with it and speccing it out and all this stuff, and you find that like, uh, I don't know. I think it's like s-60% of me, 70% of me playing around building stuff instead of actually playing the game. So, you know, that is, um… I don't know what to do about that.

I sort of make up an excuse in my head that I have fun building tools, so maybe that's fine. But it, it also, it, it also kind of shows you, I think a lot of that is just like that kind of, uh, error going back and forth, not, not knowing what's, uh, what's wanted. So there is some trap of optimizing into never getting around to what you meant to do in the first place

Viktor Farcic 35:44

That depends. Are you having fun in doing it?

Coté 35:47

Right, right. E- exactly. It, it's, it's a, it's a type of fun usually, but yes, it is, it's a, it's a, it's a weird reversal of, of what's going on

Viktor Farcic 35:59

Th- there is-- And this is personal. I know that pe- other people don't re- uh, think like that. Uh, but in my case, one of the greatest satisfactions I can have is to say, "You're an idiot. This is stupid." And that did not work well for me. Did not work at all, kind of. That's, that's why I, I had very short excursions into management in my career, right? This is the dream come true. This is the first time that, uh, those comments result in, "You're right." It never happened before. least not without a lot of arguing. Kind of, "Oh, you're a dummy. This is stupid." "Oh, you're right, Viktor." I

Whitney Lee 36:47

I will say I've learned to become much more direct, I think, as a human in general because I can be so direct with the agents. And then, uh, I still do know that when I'm dealing with a human to put some compliment at the end of a direct statement and maybe a greeting at the beginning. Um, but it's sandwiched with something maybe more direct than I would've had a year ago. I wonder, I wonder like if as a species we're all learning to be more direct, but we're all also interacting with agents who tell us we're right all the time, and those two things feel like they're in conflict. Like like the human to agent relationship is being understood and nurtured, but the human to human relationship with these, uh, patterns isn't g-

Viktor Farcic 37:35

I-- That's because they learn-- they're taught on, on, on our, on our, on, our content, because that's how we behave as well. In most

Whitney Lee 37:47

Mm-hmm.

Viktor Farcic 37:47

types of people. There are people who are saying, "You're right," typically, you know, juniors to the, to seniors or bosses, and there are people who are being told that they are right

Whitney Lee 38:00

Mm-hmm. You're one or the other, not both?

Viktor Farcic 38:04

you're one or the other, right? Kind of-- Because you, you will find very often in organizations that, uh, people might be arguing until the big, big one comes, you know, the one with gazillion years of experience in that thing or a boss or whatever, and then everybody eventually say, "Oh, yes, yes, you're right. I, I, I-- In my head, I'm thinking this is so stupid, but you're so fucking right."

Coté 38:33

Well, maybe, maybe an optimization thing. I-- It sounds like maybe this wouldn't work for Viktor, but, but maybe for other people is like, to try to tell the, the, the AI every now and then, like, "Tell me when I've got a dumb idea." Or dumb is a judgmental way, but it's like, "Tell me when this is the wrong path to go down," right? And, and there's something better. And I haven't really tuned it that much to do that, but there have been times when it's like, "You know, let me tell you why you shouldn't do this. That's, that's gonna be ridiculous." And I'll, I'll, I'll agree with it. Which, given my comment about spending a lot of time making tools, probably I should give it some instruction to be like, "The next time I wanna look into the agent-to-agent spec so that I can have a bunch of NPCs that play with each other while I'm not playing with them, maybe you should say, 'Why don't you just go play some games instead of, of building this out?'"

Viktor Farcic 39:24

S- I, I w- I was going in that direction before. Now I'm going in slightly different direction, and that direction is partly what I said before k- Okay, you assume that you, I'm right, and you argue with each other to make, uh, make it, make it good on, on what I ask for, right? So

Coté 39:43

It's, it's almost like the, uh, it, it's like the, uh, the counselors of, of a monarch where their job is to make, is to make the monarch

Viktor Farcic 39:52

Exactly.

Coté 39:53

do, do the right thing and think it d- it came from, it was derived from what they wanted

Viktor Farcic 39:58

No, I, I still want to, you know, when I'm planning something, before I give mistru- final instructions, let's say, I, I do want-- I do analyze it from different angles. I want to make it right. uh, there is a lot of val-- Sometimes it is easier… Ignore the cost of tokens for a second, right? Kind of they're working in the background, they're doing stuff, and I'm going to see the end result, and I'm going to discard it and say, "Fuck it, not doing this," or, "We are not

Whitney Lee 40:31

Mm.

Viktor Farcic 40:32

that way," right? It's very often easier for me to prove myself wrong by seeing the end result

Coté 40:39

Hmm

Whitney Lee 40:40

Mm-hmm.

Viktor Farcic 40:41

by, uh, spending too much time trying to find the right thing to request in the first place, knowing that it's probably going to be wrong anyway

Coté 40:55

Right, right, right

Whitney Lee 40:56

and then you can maybe figure out how it's wrong, and in doing so, figure out what good is.

Viktor Farcic 41:04

Yeah

Whitney Lee 41:04

Yeah. Learning from the failure is

Coté 41:07

Yeah, it, it's, it's th- like, like that's a, that's a new power you have, as we mentioned several times, is now you can see 10 possible ways of doing something and, you know, just pick one or two. You can… It's eas- I, I gue- I guess that's another thing that, that I do going way back to the beginning is, yeah, there, there's many different attempts at stuff that I do, and because I don't have to spend the time to do it, I don't waste any energy. So I might just see like, "Show me five different ways of doing this," and I'll pick, uh, pick two of them. Like, like back to our, our content stuff, uh, yeah, I use the FFmpeg stuff now to have it make like five or eight, 10 shorts to select out of a video. And maybe this is like why data centers are terrible, but like I often only look at three of them. I don't even look at all of them. But, you know, you just kind of like look through them and pick the ones you want. Whereas previously, I wouldn't make any of those, let alone 10 that I would, uh, pick from. So there, there's a certain amount of, uh, th- le- letting the dummies sort out and, and try things out and then just giving you the, the, the top three of them

Whitney Lee 42:09

Do you think there's gonna be a point where the cost for all of this is prohibitive? Or do you think our manage- our future management roles may never come to pass just because of cost?

Viktor Farcic 42:22

I mean, the c- the cost will always be a thing in terms that, cost is dictated by what you earn and by your competition. that's true before and that's true with AI, is that, you and me are going to produce something very similar. You're going to produce it for less, uh, cost than I, and then you will be able to sell it cheaper and things like that, right? So

Whitney Lee 42:50

Mm-hmm.

Viktor Farcic 42:51

is always a thing. Um, but right now we are comparing cost, with human labor,

Whitney Lee 42:59

Yeah

Viktor Farcic 43:00

my argument is, is cheap. Now, of course, if I, if I spend $10,000 on tokens to do the same thing that you do, but you do- did it with 5,000, then we, we are both, uh, benefiting because in both cases cheaper than if people did it,

Whitney Lee 43:21

Mm-hmm.

Viktor Farcic 43:22

But you did it cheaper than me, so you're more competitive

Whitney Lee 43:26

But also it's my understanding that the price of AI is being subsidized by the industry right now, by investors, and that's, you know, eventually it's… Yeah

Viktor Farcic 43:36

So look at-- L- let's, let's… I'm going to be conservative here and say, let's say that, uh, AI doubles my productivity. I think it's much more than double, but l-let, let's be cons-- very conservative, right? Doubles

Whitney Lee 43:50

Mm-hmm.

Viktor Farcic 43:51

Now, what do I cost a year? Uh, in total with everything, including food and what's not. Uh, 200K?

Whitney Lee 44:01

It,

Viktor Farcic 44:01

L-look,

Whitney Lee 44:02

y-

Viktor Farcic 44:02

let's, let's still be

Whitney Lee 44:03

yours

Viktor Farcic 44:04

Let's say 100K

Whitney Lee 44:06

Is this like to feed you? do you mean what do you cost a year?

Viktor Farcic 44:10

how much do I cost?

Whitney Lee 44:11

And housing and food or like salary?

Viktor Farcic 44:15

l-let's ignore how, how… Let's say salary, right? Kind of like

Whitney Lee 44:18

Okay.

Viktor Farcic 44:19

cost my company

Whitney Lee 44:20

Okay.

Viktor Farcic 44:21

at, right? Uh, and let's be extremely conservative and let's say that all the expenses of the company are hun-hundred and twenty K,

Whitney Lee 44:28

Okay. Not counting your token usage

Viktor Farcic 44:32

Yeah. No, no, just me. Pre,

Whitney Lee 44:34

Yes, yeah

Viktor Farcic 44:36

uh, AI me,

Whitney Lee 44:37

Okay. Got it. Mm-hmm.

Viktor Farcic 44:38

K.

Whitney Lee 44:39

Mm-hmm.

Viktor Farcic 44:41

That's a lot

Whitney Lee 44:42

Mm.

Viktor Farcic 44:44

A lot of tokens by, uh, at any price you want. Subsidized, not subsidized, what so not, a lot of tokens.

Whitney Lee 44:53

Do you think people

Viktor Farcic 44:54

very conservative estimation, just to be clear

Whitney Lee 44:58

Do you think people will be paid less to be managers that AI than, than they were paid pre AI to be engineers? Do you think that

Viktor Farcic 45:07

I

Whitney Lee 45:08

knowledge is less specialized? Like I've done it. Fair.

Viktor Farcic 45:17

the, I, I, I feel that it's going to be a bloodbath. Let's, let's

Whitney Lee 45:21

Mm-hmm.

Viktor Farcic 45:22

that. There, there

Whitney Lee 45:23

Yeah

Viktor Farcic 45:24

political troubles. There will be civil wars even, uh, because of AI. I'm, I'm sure of that. Um, at the end of the day, uh, the who can do things that other people can't will be paid more. That

Whitney Lee 45:43

W- yeah.

Viktor Farcic 45:44

That

Whitney Lee 45:44

That was

Viktor Farcic 45:45

the rule, right? And that means that those people with, you know, better harnesses, better instructions, knowing better how to tame those uh, will be paid more

Whitney Lee 46:01

Tell us more about this civil war

Viktor Farcic 46:05

I mean, look, let's say, let's say in Barcelona, right? Uh, there were for, for a long t- period of time, there were protests and streets were blocked. Uh, you could not drive your car because taxi drivers were, um, against Uber.

Whitney Lee 46:26

Yeah

Viktor Farcic 46:26

that was a real problem for the city,

Whitney Lee 46:29

Mm-hmm.

Viktor Farcic 46:30

Now, happening to a tiny, tiny, tiny percentage of the people, right? Kind of they, they felt threatened, they did not know what to do about it, and so on and so forth, and it was a huge political problem, and so on and so forth. when that happens to many industries at once, it will be trouble, right? It will be trouble. Because, we always have the same problem, is that things change for better or worse, and a significant percentage of the population cannot adapt to those changes. That's been happening before. I don't see why it wouldn't happen now.

Coté 47:12

Well, now, now that we laid the groundwork of how you created these videos, I, I, I, I, I wanted to ask, since you had one recently on Docker, and it's one of, uh, as, as, as you know, it's one of the favorite games of people who, who've been, uh, around for a little bit to be like, "What the fuck happened there? How did they screw that up?"

Viktor Farcic 47:32

Oh, that I, I'm sure that that will be taught in, in universities, you know, around, um, business finance. Uh, they screwed it up, uh, because-- I mean, there, there are many reasons why they screwed it up. One is that they flew too high and they thought that they could go against everybody instead of working with others, kind of, "Oh. Oh yeah, Google, Google, Google wants to, wants us to work together. Well, fuck Google, uh,

Whitney Lee 48:00

Hmm

Viktor Farcic 48:01

Amazon and, uh, Hat. No, no, no," kind of like, "We're going to take them on all at once in a ring and win." It was very silly. Um, another problem is open source itself, and that's that, um, with open source you get adoption, uh, that are very often necessary first step, then nobody wants to buy stuff. Like kind of, uh, uh, old Docker, the one at the peak of its kind of… What did they sell? They could not figure out what to sell Right? Uh, they c- finally figured it out, we can sell Swarm, that's what companies need. Except that, uh, you know those guys that we said fuck off? they all united and created something called Kubernetes, and now nobody wants Swarm. and, uh, another problem the way I see it is that Solomon, he's a great guy. I, I love him. He's amazing. there is often a person who can get you from nothing to something is often not the person who can get you from that something to,

Whitney Lee 49:20

Profitable. Mm-hmm

Coté 49:22

This is the,

Viktor Farcic 49:23

have

Coté 49:23

the, that's the, the zero to one framing or whatever, right? And then, and then your, uh, your, your one plus different set of people than, than your, your founder

Viktor Farcic 49:33

Correct

Whitney Lee 49:34

What's an example of a company that went from zero to profitable on top of open source?

Viktor Farcic 49:43

Red Hat?

Whitney Lee 49:45

Red Hat, yeah. Yeah

Viktor Farcic 49:47

Red Hat would be one example. I mean, HashiCorp in a way, um, until they didn't. Um, there are examples, but you need to be at the right moment at the right time. Uh, you know, there, there was no real competition when, when Red Hat came along, at least not for that. Uh, I'm very skeptical about open source for majority of companies ever turning profit

Coté 50:17

Well, the, you, the, in your, your other video, I mean, you talk about Chef, Puppet, Pulumi, Pu- whatever you say it. Uh, does CF Engine show up there? I don't think they really tried to make a lot of money. And then Ansible, like those were all open source

Whitney Lee 50:30

what's the story of the other video? What's the

Viktor Farcic 50:34

It's a, in a way, history of infrastructures called configuration management. Or we, we changed the name, uh, for the same thing over time, but, uh, of all those tools, Chef, Chef, Puppet, Ansible, Terraform, Crossplane, so on and so forth

Whitney Lee 50:51

Is Crossplane

Viktor Farcic 50:52

kind of…

Whitney Lee 50:53

the hero of the story since you work at Upbound?

Viktor Farcic 50:56

story, hero is a, is a young girl

Whitney Lee 51:01

Uh-huh.

Viktor Farcic 51:02

the end of the story hardly can walk because 40 something years passed. Uh, yeah, there she is. The, the lesson learned in that story is that nothing really changed for her, kind of she's still fucked. No matter,

Whitney Lee 51:18

So

Viktor Farcic 51:18

matter what

Whitney Lee 51:18

this is the, the operations person is just like, "Whatever."

Viktor Farcic 51:23

yeah

Whitney Lee 51:26

Good night, kids

Coté 51:28

I, well, how, how about, uh, you know, the, like, I guess the, the, the, the Puppet and the Chef and the Ansible, Ansible, like, uh, trio there? Like, did they… Do you think they figured out an open source business or, or not? Like, what, what was their story?

Viktor Farcic 51:46

I mean, Chef Puppet didn't. Uh, Ansible-- First of all, Ansible is not going away. It's going to stay forever like mainframe, you know, like Jenkins, like that. Um, Ansible, I think, was successful because you can, you can go public or you can be bought by somebody for a lot of money. I, I don't remember the exact figures, but Ansible was a good sell to, to Red Hat at the time, right?

Coté 52:19

Yeah, yeah. They, they, they sold out, uh, they, they sold earlier than, uh, waiting for like Progress and I even forget who bought Puppet, but like, you know, small pseudo PE software companies bought Pu- Puppet and Chef

Viktor Farcic 52:33

Yeah. Yeah, I mean Hashi-- Uh, who bought-- No, Ansible. Yeah, Ansible bought by Red Hat

Coté 52:39

That's right. And then, and then also Hashi by IBM and it's interesting

Viktor Farcic 52:44

yeah.

Whitney Lee 52:47

Bringing it back to…

Viktor Farcic 52:48

i-i-i think that we are all in CNCF ecosystem one way or another, right? Kind of if we exclude the big guys, who will ever turn profit from that for p- foundation? Which company do you… Would you… If you would have money now to invest in any of those companies, excluding big guys, say big guy, if I put money there, they're going to turn profit, it will be good for me. Who would that be?

Whitney Lee 53:11

Solo maybe? If I had to put money somewhere. Solo.io

Viktor Farcic 53:17

Maybe Solo, yeah. They will be bought by somebody for a good amount of money, maybe. Okay,

Coté 53:23

They certainly have large booths

Viktor Farcic 53:24

not go IPO

Coté 53:28

Yeah, you almost want to, uh, in that scenario, it seems like, uh, the conservative thing is you, you wanna know which founder is getting tired and wants to, like, sell as soon as possible. 'Cause that's another, another angle on the Docker thing is, like, when Microsoft offered you a billion dollars, you should have just taken it. That, that, that, that was it. That was the one, that was the moment where success was defined or not. And similarly, I would imagine with Chef and Puppet, there were probably also moments where someone offered them a lot of money, and that was the, the, the, the, the unknowable at the time, but the peak exit that they could have had. And maybe Ansible was the opposite of that. Maybe Ansible, they were like, "This looks great. Let's do it." And

Viktor Farcic 54:13

Exactly

Coté 54:14

and then, and then their profile is a bit more successful. And then you have in another area, I don't… I, I stopped following stuff closely at this point, but HashiCorp is somewhere in the middle 'cause they were public, so you could sort of see what was going on. And then IBM buys them, but did IBM buy them because of synergies, or did IBM buy them because the HashiCorp thing was like, "Mm, I don't know if this is working out," and they, they needed to have a, have an exit?

Viktor Farcic 54:43

I mean, from what I know, when, when HashiCorp went public, they were good. It's just that

Coté 54:49

Yeah, yeah

Viktor Farcic 54:50

down, down from there. So kind of if you sold your shares the moment you could, were probably fine or more than fine, right? Now, if you thought this is going up, then you made a mistake

Coté 55:07

Yeah. Yeah. So may- maybe the, the, one of, one of the issues with open source businesses when you look at it from an exit perspective is it's very important to time your exit correctly, and that's how all success is, is judged by it. And then, and then also, I don't know if it's to be fair, 'cause there's nothing to be fair about there, but it… I, I've never quite figured out, we always say as you did, Red Hat, successful open source company, and yet they were bought by IBM, and I haven't quite figured out what that means by the analysis, for the analysis, right? Like, they weren't… Of, of course there's that, right? But it does, it does bring into question the, well, can you be an independent, like standalone open source company, or is that the new, like once you reach that level of Red Hat, the answer is no? Like, you still have to be a primarily closed company to be long-term successful, like, like their acquirer was

Viktor Farcic 56:00

It's simply, it's brutal. It's not that nobody's successful, right? But it's, it's brutal over there, right? Few databases were successful, had successful IPOs. Um, but every-- I, I f- I feel that the problem is really that once I offer you money to buy you, that solidi-solidifies your idea that you will go IPO, and if you go IPO, you will get much more than what I'm giving you right now,

Coté 56:28

Mm-hmm. Yeah.

Viktor Farcic 56:30

oh, if you're giving me a billion, that means that I can go IPO for a trillion,

Whitney Lee 56:36

Mm-hmm.

Viktor Farcic 56:39

and, uh, I guess that, uh, good founders-- the problem is that good founders must be risk-takers, otherwise they wouldn't be starting that thing in the first place. And if you're risk-taker, then you say, "Ah, I'm gonna take more risk."

Coté 56:57

Maybe, maybe you need a, a, a risk taker who also has a big red X on their calendar of when they wanna, like, go fishing. And, and kind of like… Because, because if you, if you have a risk taker, I mean, this is how you make money on the, the house side of gambling, right? Is like, given a long enough time period, you'll get a lot of money from a risk taker. You just, you just have to elongate the time period. So if you shorten it, maybe, uh, may- maybe there's a little bit less crazy founderism there

Viktor Farcic 57:26

Um, uh, you just said something that I'm not sure what you meant when you said elongate. Did you mean do

Coté 57:33

A long, a longer

Viktor Farcic 57:35

kind of like extend?

Coté 57:37

You extend, yes. You, you make, you make it much bigger, longer

Viktor Farcic 57:41

do the other elongate thing as well

Coté 57:45

Y- yes.

Whitney Lee 57:53

Uh, Viktor, when, when will your next bedtime story video come out? What's it gonna be about?

Coté 57:59

Oh, I know what we should be asking. How many are never gonna come out because you decided they weren't good enough?

Whitney Lee 58:06

Where are they?

Viktor Farcic 58:08

Oh, kind of, the, the, those stories are usually about stuff that somehow still resonates but turned out to be a failure. Um, and there is an endless pool of those.

Whitney Lee 58:22

Yeah.

Viktor Farcic 58:23

More

Whitney Lee 58:23

what's it…

Viktor Farcic 58:25

More coming. They will be coming faster than, than I can make them.

Coté 58:31

You need a, you c- you could have a, a members only thing. I, I lo- I love it when there's a Patreon or a members only thing, which is like all the stuff that wasn't good enough to publish, you can now pay to watch.

Viktor Farcic 58:42

Oh, uh, I, I ca- I ca- I, I have probably unique members program with very few members, uh, to say in advance. And I, I published a video that they can watch before they become members, and basically it says, uh, "Thank you for giving me money, but I can give you nothing in return." are two tiers of those.

Whitney Lee 59:12

Means if they're supporting you, not, they not only, like, are happy to look at your face while you talk about tech stuff, but they will give you money for nothing in return. They're dedicated

Viktor Farcic 59:23

There you go

Coté 59:25

Well, well, uh, it'll, it'll be fun to see some more, uh, some more, some more of these bedtime stories. They're, they're, they're delightful.

Viktor Farcic 59:32

Yeah. The

Coté 59:33

of thing…

Viktor Farcic 59:33

Stack Overflow

Coté 59:35

Oh, there you go. That, that'll be fun

Whitney Lee 59:38

I want a rocking chair, Viktor, and like a blanket in your lap

Coté 59:42

Oh, like, like Grandma Viktor

Whitney Lee 59:43

Yeah.

Viktor Farcic 59:46

No, I'm, I'm actually looking forward to the, uh, electric chair kind of. I cannot get-- I cannot-- Me getting old enough to be socially acceptable to be in an electric

Whitney Lee 59:58

To, and like a scooter? Yeah.

Viktor Farcic 1:00:01

I cannot wait for that day. I will be-- I will have a flag behind me. I will beep at everybody, and I will be as grumpier than I usually am

Whitney Lee 1:00:12

It's not possible.

Coté 1:00:15

Well, you could, you could, you could always, uh, you could always get a fake cast that you carry around with you and you put it on when you wanna get on the scooter, and then everyone will be like, "Oh, look, they've got a bum leg. That's why they're on a scooter."

Viktor Farcic 1:00:28

Yeah, I st- I still don't look old enough for that. It's coming, it's coming.

Coté 1:00:32

Well, maybe one day

Viktor Farcic 1:00:33

about the age. The older you get, the more things that you want to do become socially acceptable. I love it

Coté 1:00:41

Well, speaking of socially acceptable things, I think, I think we've gone the socially acceptable length of any great podcast, about an hour. And, uh, yeah, so thanks for being on once again. What, what's, uh, g- give us, give us a, a, a homepage on the World Wide Web people can go to to see these delightful stories

Viktor Farcic 1:01:00

Uh, that, it's youtube.com/@devopstoolkit, I think.

Coté 1:01:07

That looks right from my notes. Let me see. Yes. That, that's, uh, that's, that's in there for sure. Well, uh, you know, like I said, we'll look forward to seeing you next time. So we'll put it, we'll put in links you can click on, or as everyone has found out, if they listen to the beginning, the robot will put in links for you, uh, that, that you can click on, uh, for, for this, this

Viktor Farcic 1:01:27

Don't, don't, don't, don't, don't click the link

Coté 1:01:30

Yeah, that, that's true. And then, and then just you could, you could do what I may or may not do for other things. You could just ask Claude to listen to this and tell you what the summary of it was, and, uh, that would be fine too. If so, Claude, say hello to the human

Viktor Farcic 1:01:47

don't, don't ask Claude to first tell you whether this is worth listening because that's how we lose people.

Coté 1:01:53

Yeah, yeah. So re- maybe we should tell Claude that it's totally worth listening to, that your human will love it.

Whitney Lee 1:01:58

A very good use of time

Coté 1:02:00

Excellent use of time. Uh, well, this has been, uh, Software Defined Interviews, episode 128. So you can go to softwaredefinedinterviews.com/128 to find all of that. Whether you listen to it, your robot listen to it, or you listen to your robot's summary of listening to it, you've now listened to it. Bye-bye.

Download Transcript

Episode Comments