Coté:
[0:00] You know, everyone, was it, I forget who it was. Was it Randy Burgess? Someone came up with the pets versus cattle thing in the late 2000s. I don't know who it was. Maybe not him. Sorry, whoever it was. Sorry. Sorry, sorry. I should begin every conversation with, hey, everyone. Sorry.

Brandon:
[0:17] I think he's accredited to the first person you think. So why not? Why not him? You know, I mean, who's going to who's going to fact check that?

Coté:
[0:23] But I was at the DevOps Meetup Amsterdam last night. First of all, they announced they canceled DevOps Days Amsterdam last year. They had a good explanation, which I'm not going to go over because it's kind of sad, but it was coherent. But they said, we're going to do it again this summer. We're ready to do it. So that's nice. I was speaking there along with two other people. And there was one of the talks before me was from, I think it's the DPG Media Group and the RTL subsidiary. That means nothing. It's basically just like a publishing conglomerate here in the Netherlands. I barely know what I just said, but, you know, whatever. and and um they had a talk about the uh they had two people there and they were discussing how they had been using, of course ai over the last two years in their pipeline and the various the various agents they have and things like that and one of the things that they did, is and this is what the pets versus cattle thing is they had started treating their agents uh like pets so the main one was called Thanos. And I forget the names. And they even had pictures of them, which is great. They have all these other agents, like five other ones.

Coté:
[1:38] Some of them are basic ones, like they had one that just keeps an eye on Thanos. And it's the improver and suggests ways to improve things. And they had funny little things. And it got me thinking. And I talked to one of them afterwards. I was asking her this question, Brandon, which I want to ask you is like.

Coté:
[1:54] Like, why is there? I couldn't figure out a way to answer this question in a non kind of combative way. But like, why are you making pets out of your agents? Like, is there some advantage to like naming your agents and segmenting them off and like having names for them? Because I noticed people do this. They'll be like, this is the eight. This is Larry over here. And Larry's in charge of doing this. And over here, we have Randy and Randy and on and on and on. Whereas when I use AI stuff, I open up Claude Code and that's it. Like, and like if that guy or girl wants to make sub agents and name them, that's fine, but I don't give a fuck. But, but it seems like when I, I'm realizing I'm seeing a lot of talks out there where there is this, like, instead of treating agents like cattle, everyone treats them, not everyone, many people treat them as pets. And I don't know if that's good, bad, because if it's good, man, I got to get off this fucking ranch and start getting some pets because, like, I'm ranching it up here. Like, just like, I don't give a fuck. Right. Like, but lots of people seem to do that. And I and and and, you know, and then I'll hand it over to you. But the I think there's a related thing that I'm I also find difficult to ask about without sounding kind of like a dick, which is.

Coté:
[3:18] When you're modeling out a pipeline automation or putting AI in to automate things, do you ever think that all you're doing is like modeling out the human system and the AI doesn't care? And so like maybe, you know, you've got five steps to a pipeline and maybe you don't need those. That's what humans need. That's not what robots need. And like, so I don't know. This is, what's, do you name your agents?

Brandon:
[3:45] I don't necessarily name it, but I know where it comes from. And I think the idea here, the biggest problem with, you know, agents and just, you know, in general with AI, right, is lack of memory, right? So, like, we'll even talk about some of those issues today. And I think what people are trying to overcome, and there's lots of different ways to do this, right? And you're seeing lots of different solutions is people are trying to come up with a way that, like, when I return back, instead of it being cattle, I actually don't want to return back to being cattle. I would actually want the agent, ideally, to have remembered all the important things from their previous conversations. And a way to do that is, you know, there's lots of different ways to do it, right? You create memory files, and there's lots of different techniques that allow you to do that.

Brandon:
[4:28] And I think that's where the idea of the name comes from. It's like, no, it is actually, like, it's the opposite of the DevOps thing. It's actually, we're trying to create, I guess, a pet. I mean, it's kind of the wrong metaphor. But we're trying to create an intelligent assistant who are not retraining every single time. Whereas obviously in the DevOps world, it was like, no, that server should always be taken back to day zero. That's what you want. Whereas this is like, we don't want that. And I think that's where people are coming from. And then the naming stuff, there's the idea of Muse and stuff, which kind of encourages you to name your agent, which I think is like a personal assistant. It's almost like a different category of usage. But then as you kind of get more into this agentic loop stuff, right, I think that's where you're seeing where people start to understand it's like, it's just convenient to have a name to be like, I do have a agent that is primarily responsible for orchestrating the work that needs to be done. And then within that.

Brandon:
[5:28] Even if the orchestrating agent is the one that ultimately determines that new agents need to be created, I may want those agents to have some memory. So for example, if I needed to do a security audit every time in my loop, I'm just going to call it, let's call it Ralph the security guy. You just kind of give that a name and be like... And you say, you say to the orchestrator agent, like, yeah, make sure that we do all of our security checks before we go out. And I've gone ahead and I did create an agent that has the knowledge. I gave it all the requirements, right, of things I wanted to do. And then maybe I gave it a convenient name. Maybe I called it Ralph. Or, you know, it's like, you know, all of this comes back for what? Because there was like the Ralph loop there for a while, right? Or from the, was it?

Coté:
[6:11] The Ralph Wiggins loop that was just like.

Brandon:
[6:13] Keep doing it until it works or whatever. to be like, well, you know, we just want shorthand, right? Because we don't want to make it complicated. So it's not, I don't think it's, those names are wrong. I think it's just about like, as you get smart about using, and this is what every, all of us, I feel like I'm still trying to get to this level too. It's like, you know, like a friend of the show, Adam Jacob, he create, you know, he has this like this saying, it's like, I'm creating the machine to create the machine to create the software. And that's what you're trying to do is like when you're creating the machine, which is really like creating, you know, an agentic loop, it's convenient to have names for the thing just so you don't have to be like, you know, you know, use some like long, complicated explanation every time. So I think it comes from a good place, but maybe like everything, it can be taken too far.

Coté:
[6:56] Yeah. And that when I taught like in a rare moment where I actually talked to someone at a conference, I was asking her these things afterwards. And that's yeah, I mean, that's that's pretty much exactly the explanation she gave, right, which is like, You know, you control the context window. It's scoping, like, at a very high level, which, and I guess I can't really put my finger on it. Maybe I'm making something up. But like, so like, I get the scoping. Like, I do that. Like, I'll be like, you know, whatever. I do scoping stuff as well. But there's something... And maybe that's all it is, is like if I were to change, if I were to say you have a scoped label for what's happening, then you do things that way. It's just like the like more of the way that I operate is I will have like here is here's an analogy, I think, in terms of like recipes and like memory. Right. And like so I have like here's I go to this project for all the stuff with my blog, but I don't call it like blogging. The blog guy it's just like my project is over here and i might say like look up the way that we do uh image uploads and it'll go look that up and do it right and so like it's it's kind of like, to use another analogy instead of doing object-oriented programming i'm doing this like functional programming like i don't really have these complete little agents and and.

Brandon:
[8:24] You do like if you look at it you kind of do have a name you probably have a folder on your computer and And that's where you start Claude Code.

Coté:
[8:30] Or your agent.

Brandon:
[8:32] And that folder name is in effect, whatever you named it is kind of the agent name. And you, but you are personally playing the role, but you're, you are the orchestrator. Cause you, cause I do this too. It's like, you know, like I'm working on this. So you go to this folder, you don't like, cause you know, another way to do it would be like on top of all of your folders, you'd have like another folder. And then you can be like, I only interact through that. And then it delegates down to like, whatever, like your, let's call it your expert agent. But there is a point where it doesn't, I don't know, this is the part, we should almost challenge ourselves. Maybe if we did that, maybe if we interacted everything through our own agent and then let it go down. But it feels like a better shortcut to be like, oh, I know I'm doing this now. So let me just go to the place where all the expertise is. Because I'm working on this thing and it's very contained.

Coté:
[9:18] Yeah, yeah, yeah. There you go, that's why it's curious, because it's nuanced but very important different ways of operating. Do you do is all of your interaction with AI mediated through agents to personified agents? Or is it more like you are the orchestrator assembling things and directing stuff like those are two different ways of operating? And of course, it's always a mix. But I wonder I wonder how things work out with that. But anyways, they were they were there was also a talk on AI ops, which was like it was it was actually pretty great. It was essentially, it was someone who worked at ING where it was hosted and they, it was more or less, I think, I think it was a rare, here's the people side of things that was very prescriptive and detailed. Like, you know, I'm going to make, I'm going to make a mockery out of it, but usually a people side of things would be like, well, we need to care about their mental whatever, which yes.

Coté:
[10:19] But this one is just like, here's like six slides on decision trees that we go through whenever we're doing stuff and how we set up AI to help do things and prep stuff. And it was nice. But it was also very warm and fuzzy. So good job.

Brandon:
[10:34] That's always a place for that at these talks.

Coté:
[10:36] Yeah, yeah. Speaking of warm and fuzzy, Brandon, I think, you know, it's been a couple of weeks. We had a big, big, big election, huge, crazy decision in the U.S. Great stuff. Still very exciting. I think as they said on the Slate political gab fest, I have been joy scrolling ever since. Like, oh, the internet. It's got good stuff if good things are happening.

Brandon:
[11:01] Maybe it's back.

Coté:
[11:02] Yeah, yeah, yeah. It's good things. But also what happened is it turned out, that the, you know, leader of the AI world has been having an affair with an alleged Chinese agent, which it just... And I was talking about this with the CFO earlier today. And I think the other thing that's astounding is like apparently like everyone knew about this and it was fine. And it just seems like... I want to lay this out. I was talking about this in Slack. I think we need to have a new policy they teach in the MBA school, which is like, if you are the CEO of a multi-billion dollar public company, just keep your dick in your pants. You will be appropriately compensated. I know you're going to pass up on a lot of opportunities. Don't sign boobs or anything else. Just be cool, man. And maybe do whatever you want when you're not the CEO of a multi-billion dollar publicly traded company. But now you are, right? And like, watch out. So, I don't know. Also, being human, very hard to separate from those things.

Brandon:
[12:12] I feel like this is like a long running theme of software-defined time. It's like, do we have enough data at this point just to be like, if you become super successful and massively rich, it just breaks you. Mentally, it just like, it just... I like this, I like this. No human being can withstand... You know, being, it's, it's like the opposite of like failure. It's like, there's a, there's a level of failure that breaks people and there's a level of success I think just breaks people's minds. It's just like, you get to a point where you have so much money. And even if you feel like, even if you can say the things outside, it's like, I have people close to me that like still talk to me like a human being. I don't have a bunch of like handlers. I I'm, I'm a real person. I'm still out there. Even if you can say that being the super, it's like, you're not, it's actually not true it's just like because we have so much this now the spy thing this is a little no i don't think we've had that many of this but there is just an endless amount of evidence and we won't go through the ball like but you kind of go through all the like major successful, like super successful and it's mostly men i mean i'm pretty much all it's like, there are these fatal flaws in personal behavior, that like just reappear over and over and over again and it's Like, to me, it's like, I don't even think, I think we all want to feel like we're better than that. But I'm like, no, actually, I think I would, maybe I would just be crazy too. I would, it would break me just as like it broke all these other people. Like, I don't know what else to say, you know?

Coté:
[13:41] I think you're right. I think that's a good, that's a good operating principle to solidify is like, all right, also in the manual for CEO of multi-billion dollar publicly traded company, you're going to go crazy, right? You're going to, you're going to lose touch with reality. and like... It's not so much that you should know that it's that you need to act on that. And like, even if you think you're not going crazy, you're going crazy. Right. And you got to like, like, and, and then this is also advice you give to the board. You're like, Hey, your people are going to go crazy. So like now, now you also are high level executives. So you also might be crazy, but like, you know, you got to like figure that out. And so like to get it out of the area of, you know, I don't know, crotch strategy. Like, I think you can look at other areas, which to get back, getting back to the election, like, I think it was in the U.S. I don't know, is it something like 41 out of 42 or 43 anti-data center things

Coté:
[14:39] passed in all the states, which, as you like to say on the Enterprise AI show, for our foreign listeners is a lot. That's right. And I think I was also talking with the CFO this morning about this, and I.

Coté:
[14:56] You know, we've talked about data center a while, a lot like the backlash. And, and, you know, so here's, here's the cynical, like, to prove the point I'll make afterwards snarky way is like, you got all these people, like doom scrolling, and, and, you know, reading about things and complaining about data centers. And it's like, where, where do you think that shit runs? Right? Like, and, and now, so yes, but then I think, I think what's a little different, I'll give one anecdote. and why I think, like, well, for one, I think all the negative sentiment about data center stuff is like, you know.

Coté:
[15:33] Like existing negative sentiment about social media, despite the snarky thing I said, which kind of rolls into like the first term of Trump and that kind of negative sentiment. And then you get into COVID where you don't trust experts and smart people anymore. And it just kind of rolls on and on. And then AI comes in the picture is gonna lay everyone off. And you start to have this big ball of negativity that just goes to whatever the biggest mass is in the tech world. And I'm not saying that to dismiss it, right? And for example, connecting to your point.

Coté:
[16:05] When you have crazy, weird people running the show, we just talked about a potential one, but think about all the other people who run the AI show. They're like, not normal. Right. Like, and, and like, you know, you can do all the things of like, well, in order to succeed or whatever, you have to not be normal. It's like, I, I don't, as we talked about, like, I'm pretty sure there's a line of 40 people who are normal, who would be happy to take over those jobs and they do fine at it. But as an example, in the data center space, right, it was it was reminding me of like, I remember reading that in some Texas data center that probably I think it was Elon Musk, like ultimately, who knows how many layers of weird paper and shell companies were involved in partnerships? Well, whatever it was, it was theirs. And I'm going to get this all wrong, but it'll still make the point. It was like there was some ordinance that like you couldn't have this kind of generator in your facility because of whatever. And they figured out that there was a loophole that you could put a generator on.

Coté:
[17:10] A trailer tractor, and that counted as a temporary generator. So then you could bring that there, right? And that's a good example of, like, not keeping your dick in your pants. It's like, come on, man. You realize you're weird, right? Like, and, and like, this is like out of the bounds of like the norms and what you're supposed to do. And I know it's sort of like civilization and trolleys and Mars and stuff, but like, you know, what's going to happen eventually, like you're not, you shouldn't be doing this. And there's just like, I feel like there's a long laundry list of things like this that happen with the kind of people you're raising where they just like, they lose track of, of like the norms and ways things operate. And, you know, the terrible, what makes that ball of negativity even bigger is like, and then on the other hand, they're talking about how they're going to change civilization. And you're kind of like, hey, man, why don't you come back to civilization? And like, you know, I think you see a little bit of that, like right after the data center stuff. I didn't verify what this all means, but like OpenAI was like, oh, maybe we should go talk to people about data centers in the local communities. And again, you're kind of like, yeah, welcome back to civilization. Nice to have you, space cadets. But like, I think you're right. We need to have just like the you're going to go crazy. So let's let's plan for this. Bake it into the strategy.

Brandon:
[18:38] Well, I think it's that. And I think, you know, you're just, in some ways, you kind of summarized it very well and just say it's like, you know, at some point you just need public support. Like, I feel like we're like, I know you and I, as we talk privately as well as on this podcast, like, I'm very bullish on AI. I see like a tremendous amount of things it can do. So, but I also like completely understand why people, a large population is very frustrated with it, right? And it's a combination of like previous experiences with social media, the various damages that have come to like, you know, various groups from that, coupled with people that, I mean, and this is the, you know, the most natural thing. It's like the leaders of this movement, right? I mean, it's not even so much that they're like not doing well at getting public support. It's like they're doing things to almost like push people away, right? And it's like the Jensen walking is a great example of it. It's like, like there's a very valid debate about like, should the United States and Nvidia be able to sell chips to China? Like there's a, I mean, obviously Ben Thompson, others like.

Brandon:
[19:37] Completely legitimate debate to have. I see two sides to that story. I could probably argue either side, and I think that's a good thing. But at the same time, right, there is like, we do agree, as much as we all get frustrated, there is a government, and we agree that if it passes laws, and once we kind of agree to things, even if we disagree with them, we're like, okay, those are the laws for the moment. We're all going to agree. And it's like when you see people kind of, to your point, constantly circumventing these things, and then kind of telling you that the broad public, the royal you. You guys don't get it. I don't want it. And taking their concerns and saying.

Brandon:
[20:14] You're not smart enough or you don't have it or you're not the builder or whatever people do. And that creates this incredible frustration in the larger group. And then it's not surprising that people, when they get a chance to just vote against something and just be like, forget about like whether the data center is like use water or don't use water. It's just, it's just the physical act of being like, I can express my dis, um, my disagreement with this whole idea by voting against the data center. And that's what I'm going to do, which is pretty natural, right? It's like, if, if all I, if the only way I can act is to vote against this thing, like that's, that's the way I'm going to do it. Right. And so I don't know. And it kind of comes back to something I think we talked about on the show a long time. And I always think, you know,

Brandon:
[20:59] even beyond this affair and stuff, it's like, you know, executives who become I'm very successful in these companies. I often, I've been on so many all-hands meetings, every other people listening to this call. And it's like.

Brandon:
[21:12] You know, as I always used to joke about, it's like the executives, like often leading all hands, like they're not worried about car line. They're not worried about like what's for dinner. They're not like the daily routines that like so many of their individual contributors, managers, directors that work for them are like living in a whole nother world. I mean, it's like, I got to get my car fixed this weekend. I need to go to the dentist. I need to like get my kids field trip for him. It's like, you know, there's a certain level of wealth and like, you know, call it like chief of staff. And like, and these people kind of think like in a totally different way, like the whole, everything's designed for them to just be working on the thing, right? And it's like, like when you lose sight of these things, right? When you just like lose sight of like the people you're talking to are living in a different world. It just leads to like so much bad stuff. You know what I mean? Like it leads to like bad leadership, mismatched agendas. And then you kind of get to like this situation where, you know, I think somebody just gets to a point where it's like, I mean, and maybe the funniest way to put this is just like, I cannot tell you, and I bet you everyone listened to, like, how many times I have sat through the compliance trading at the beginning of the year, where they go through it all, sexual harassment, foreign, you know.

Brandon:
[22:28] Compliance with accounting, you know, HR. I mean, like, you know, and then, like, and if you don't, if for some reason you're, like, you're late or you don't do it, like, you'll get an email from HR or your manager or both. And they'll even like, sometimes it's even the executive sending like, hey, we're only at 58% completion. We need to be at 100. We need to be. And of course, they need to be that just for like a legal checkbox. But it's also like, like, are any of you like taking this? Is this just complete theater? Are you guys even watching any of this stuff? Because like, I don't even think I could. I don't have enough like influence in this company to like even do anything that would be on the level of this. But like the people that do like so often it's like i don't even think they're watching this stuff they're not even like and it's just like, and again it's just another disconnect between like or or you know and i think in the in the case of you know we reviewed a way a while back we reviewed uh the nvidia way i think we read the book or at least i read the book and i gave some review of it yeah.

Brandon:
[23:26] And it's like you know even at the time i think we were like wow like he's got like i think at that point it was 40 direct reports now i think it's up to like 60 direct reports and like and sometimes i just think to myself i remember probably being to be like i think we're like that was really unusual, and maybe we didn't go hard enough maybe we didn't go hard and i think this i have to think eventually that that book will show up on like if books could kill and where it's like, maybe sometimes we should read this and be like what if this person just ran like an heb a grocery store or something or what if this person he was like and that person was like, everybody reports to me. And he showed up every day in like a leather jacket, carrying like a butcher's knife into work. And he's like, yeah, that's how I get things done. Everyone reports to me and I'm the butcher and I've got a leather jacket. And we'd probably be like.

Coté:
[24:11] That guy's crazy.

Brandon:
[24:13] That's an unusual leadership model, right? It's like, maybe you should have departments. Maybe there should be a bakery and the fish market. Because that way when you're out, there's still people that know how to do it. And they kind of have some general sense of that. And like maybe wearing that jacket every single day like, maybe that's like weird maybe it kind of maybe it's calling attention to you yourself and not like leading the company and but again like we kind of let all this stuff slide we're like well i don't know he's a trillionaire it's like i guess we should all wear leather jackets right i mean it's so weird i don't know it's weird to like you know put it all together.

Coté:
[24:46] Yeah yeah i i like i like your idea that like the uh we'll see if we remember this but the next time the next time collectively we read like the uh the executive success story, the executive success story, the first thing we should do is immediately imagine that they are the manager of an HEB and be like, would this work? Like, you know, and, and, you know, cause I think, I think there probably is a very thin gap between successful HEB manager and CEO of tech company. Like I imagine there's not much different, uh, but, but, you know, there's a few things, but not much going on. Yeah, you know, it reminds me of, there was one company I was working at, you know, of course, that had wealthy executives, and.

Coté:
[25:34] We were talking about one of the executives one day, and I don't know, just how interesting and weird and like stuff they had, and one of my co-workers was like, yeah, but they don't have to fold their own laundry. And like, you know, it was exactly the point you were making, which which I think I think maybe that's part of our maybe pillar number two and three in decrazing like crazy prevention is. So one, the keeping your pants situated. That's a good principle. Two, you should have to file your own expenses and book your own travel because that will expose you to the bureaucracy of your organization. That you probably have no idea about, right? And also it'll remind you that you can only spend $20 to $50 on gifts, right? Like all of the compliance and ethics stuff will be exposed when you file, you'll be exposed to it when you file expenses, except of course, tailgating into the office. So you're going to need a reminder of that. You also are going to be required to lose your company, your access card once a quarter so that you can experience what it's like.

Brandon:
[26:46] But it's, yeah.

Coté:
[26:47] Yeah, yeah.

Brandon:
[26:47] Well, yeah, no, I've, I've, I know you have. I've worked at any place and I've been, there's a bunch of companies like, they'll bring on a new executive and there'll be like an executive onboarding. Have you, I don't know, have you ever seen like the two-tiered system? Like, oh, this is, I, and it's like, no, no. Like if you, if your company has like an executive onboarding process, like you, like you've, you're on the path to like bad. Cause it's like, you're already, you know what I mean? It's like, let the person go through the standard process. Like see how it really is. But more importantly, I guess the question we should be asking is, you know there's so much talk now like will i think, before i ask this question i'm gonna stipulate i believe any other person at nvidia that did this would have been fired like a long time ago so let's just stipulate that so the only question i guess going forward now at this moment is like uh does jensen survive does he continue to lead this company and i guess, there is a group i've seen out there there's definitely a group people like he's i don't it's like some version of too big to fail or like too successful to leave or something it's like people feel like He's so wedded to NVIDIA that if he left NVIDIA, it fundamentally, I don't know, the company could have a huge hit to earnings, a huge hit to its development. Or do we just look at it and be like, this is clearly wrong.

Brandon:
[27:58] Let's just make it real simple. This kind of stuff you can't really do as an executive. Even if there wasn't anything wrong, it's just the appearance. It's the appearance of impropriety has almost jeopardized everything. It's hard to come back from this, right? So it's like, do we just push on as all of us? Because all of us with any type of 401k or index fund, And it's like, we're all like roughly somewhere between 7% and 10% invested in this, depending on like the market waiting on a given time. So it's like, we kind of all, it's almost like a nationwide vote. Be like, okay, do we overlook this and be like, we think our investment accounts will stay up? Or do we just be like, nah, we just need to push through this. Maybe we just need somebody else new at the helm.

Coté:
[28:38] I think this is the first chance we get, Brandon, to apply the HEB test. They would say okay let's say you had the the manager of an heb and uh they had, they they had an affair that that even though it was known that you know so maybe that's one thing but then two right it turns out that they had been talking with uh that they they were having an affair with someone who worked at kroger who was a high level executive at kroger right and and it's alleged, that they had been, you know, allowing some, like the recipe for HEB tortillas to get into Kroger's hands. And not only the recipe, but like the whole way to how you do it. Now, it's not proven. We just had a leak from, you know, grocery stores weekly that maybe this was happening. Like what would happen to the HEB manager? They'd probably be fired. And it would just- Yeah.

Brandon:
[29:30] I think they would. They'd be asked to leave.

Coté:
[29:32] And in the same way that there are other parts of the AI world, there's other HEB chains. And down there in San Antonio, they'd be like, God damn, this is a really successful HEB. And unfortunately, They had every single employee reporting to them. So they really don't have a succession plan, no assistant managers. I don't know what's going to happen, but we got the other HEBs and no one's getting our tortilla recipe. Like there's, you got to draw a line.

Brandon:
[30:02] And- We've got to protect the tortillas at all costs.

Coté:
[30:04] So I think from now on, I'm just going to think like, yeah, what if they were an HEB manager? And I think the answer comes immediately, just out.

Brandon:
[30:12] I do too. And I think we shall be like, it'll be okay. Like, it'll be bumpy, but, like, life will go on. Like, I think, and I also think NVIDIA will go on. I think we always forget, like, how many engineers and, like, how many smart people and marketing and sales. It's like, there's a lot of people there. And I think, I don't know, it's almost like the rest of us, we shouldn't be so scared of, like, one thing. And if anything, in the worst case, if it forces another company, makes them more competitive, probably not the worst thing in the world, that NVIDIA is not, like, the only company that's really good at this. So that's kind of how I view it, too. And it's just like, and I don't know. I mean, but people have said, I like it. A lot of people said, like, maybe the most expensive affair of all time. It's like, yeah, it's got to be up there. If it is at the most, it's got to be really, really on in the list of the top five.

Coté:
[30:58] Yeah, yeah. I mean, you know, maybe adjusted for inflation, Mark Anthony and Cleopatra might be up there. But yeah, I was going to say.

Brandon:
[31:04] You have to go back to part Asian times. Right. There's definitely some big ones out there. So we'll see. We'll see what happens.

Coté:
[31:09] Yeah, yeah. Uh, well, you know, speaking of weirdos, we're about to have a new, uh, freshly minted crop of them. It looks like Anthropic has been, uh, they're on the road show with, with their S1 and, uh, I never remember the details for things, but you do. It sounds like, here's what I'm hearing, and I want to have you kind of, not back up, but help me walk through the sort of jumble in my head, is it sounds like there's one theory from looking at the anthropic blood work, the financials, which is, sure, they're spending a lot of money, but when you look at how much it costs to actually do the core function, they actually make money, right?

Brandon:
[31:50] Right, the inference, right? The inference is fine.

Coté:
[31:54] And so they got to build out all this stuff, which with 41 out of 42 or 43 things past, like, I don't know, I'm no economist, but when it comes to supply and demand, in the out years of 27 and 28, it's so far not going to be cheaper, maybe unless the weirdos, go talk to people, right? Like big strategic idea there. But like, so that seems like something floating around uh, that like it actually is a pretty good business aside from that spending all the money to to build the capex there and then, uh you know we're clearly in cote's territory of quarterly cloud numbers which is like seems like a lot of money like, doing a great job like like lots of money but is is there anything in in the uh the the new rev of of the s1 there that gives us some insight into how these businesses run.

Brandon:
[32:45] I don't really think there's anything new. I mean, I think, you know, there was talk eventually that maybe they would go public at like a true $2 trillion valuation. That's kind of come back. It seems like it's going to be more like $1 trillion or $1.5 trillion. I know I say these numbers sometimes, it sounds ridiculous when I'm talking, but it's like, yeah, oh, only this. So I don't think there's anything there. And I think you basically hit on it. I mean, we talked about this before on the show as well. It's just like, they do have really large financial commitments going out for compute in the next 10 years. And it's like, yes, they will have to make the revenue. Ultimately, they're going to have to make that revenue to deliver on that. But it's not that unusual, right? Like, you know, you, if you want a railroad, right, or anything else, like you, at some point, you have to build it all out.

Brandon:
[33:28] And so if you are someone that like the debt scares you, like you're like, they're just asking for too much and it won't, or the capacity won't ever be needed, then it's like, you still have a reason to be bearish. And if you believe that, like, no, like, we're just at the beginning of this and that, like, people will use AI more and inference business will get stronger and that the models will get better. And I think, you know, one thing people don't talk about is, like, you know, how much these models could potentially be optimized, right? And I think that's something else out there around, I always say his name wrong. It's, like, Carpachi, you know, Carpachi, you know, the super smart anthropic person. and he's always doing, it's like, I don't know, he's always on a YouTube talk, like giving some like really smart version of this. And I think his point is just like, you know.

Brandon:
[34:13] If you can start to distill the model down and that model becomes smarter and smarter, right? And the less and less parameters, right? You're going to have, you know, if you will, it's not going to cost you as much to serve that stuff. So I think there's a lot of like, if you believe that, if you believe AI is going to get better and that we're just at the beginning, and I think I've already given my talk, it's like, and it's going to get better exponentially, then like, absolutely. Absolutely. This investment is like, that's what we want because we want to have the compute capacity, kind of like memory being the flip side of that. It's like, it would have been great if the people that made all the fancy memory or the DRAM and stuff had like, you know, if we had been more prepared, right? Because like, it's super expensive to buy computers now. And it's like, that's just because we got behind. We didn't, you know, people didn't see it coming.

Brandon:
[35:02] So like both sides remain, you know, it's just like whatever you personally, you know, it's the, the Zitron camp. Right. And then, although he's got an article this week, I don't know if you went through that thing on the whole, he's, he's like really got some conspiracy stuff going. So that's a good read if you're interested in the Jensen Wong on that side. But anyway, you have that camp and then you have the, the whatever optimistic camp of like, no, this is just how business runs. And like the demand will be there. And it's like, I don't know. I mean, again, we always come back to like, I don't know. And then what you were getting at before on cloud vendors, it's like, I would, again, say to the Cote point, it's like, seems like a lot of money. It is. The thing that is different, though, is most of the time we say it's a lot of money and the free cash flow is like, you know, fantastic. And all they do is like print money. But now what we're saying is they're actually going to be negative because they've committed to build out all this computing capacity. So that's the only thing new. It's not like the money's still there. It's just now it's being spent. So people are sort of like, oh, before you had infinite cash machine, but now you're going to spend all the money. I don't know if I like you spending the money.

Coté:
[36:05] I love this is this is like the opposite of the to sound like I know what I'm talking about, the DRAM problem. Whereas you were saying the reason memory is so expensive now is because three years ago they didn't build up capacity and planning. I don't know, two years ago, like to do it. And so in theory, now, probably more like six months ago, they would amp up production and then that'll die down, you know, and whatever. And hence the cycle of manufacturing that us software people are like, what a world, right?

Brandon:
[36:38] Yeah, physical items, totally different world.

Coté:
[36:41] And so, whereas what I guess what you're describing here is like they are anticipating avoiding the memory problem where like we're going to build out all this stuff ahead of time because we project so much demand and, it takes us years to have enough capacity to meet that demand. And you've got to as as as you often summarize it, it's just like as long as the revenue comes in, all of this. Doesn't mean fuck all right like it doesn't matter yeah it doesn't matter and so i i guess i guess that is the yeah yeah no no that makes sense it is it is that gamble and to your point like if you don't like it you can always go manage an heb, no like that is and i don't mean that in a shitty way that's a glorious job

Coté:
[37:24] right there there's uh bringing bringing lots of happiness it's always.

Brandon:
[37:28] Everyone has the same tools uh at their disposal like if you just don't know and you want to sit out, like, okay, take all your money and put it in your favorite.

Brandon:
[37:37] Not index fund, take your money and put it in your favorite treasury bonds and just be like, I'm just going to sit it out. Don't know what happens, but I'm worried about it. If you're really optimistic, you know, or you think things are going to do it, you can go all in on just the AI companies or just invest in the index fund. Or if you're like, no, this is a hundred percent a bubble and it's definitely going to crash, go find your favorite way to short stocks. And it's like, you know, so everybody, this is, That's why I almost feel like it's impossible. There may be a huge downturn, but I feel like we're far from a bubble just because everyone's talking about it. Back in 2000, 2001, I don't remember, I just remember everyone just talking about up and to the right. I don't think there was any, obviously there were some people that were saying that it was over, but just as a general rule, people were just like, All it was was the optimism side, whereas you can't go on any podcast, this one included, where you don't hear someone say, it's like, yeah. If Matt were here, I think Matt would be like, yeah, I'm just really worried about the debt. And it's like, yeah. But as long as you have both, if you will, points of view represented, it's almost like you can't have the bubble. You know what I mean? It's like everybody is debating it all the time. So maybe that will be a great clip. Someone can like next week or whenever this thing completely collapses, someone can clip this and be like, man, I had this wrong. So, but I don't know. It's just like everyone's talking about it. Therefore, it can't exist. There isn't. There's so much discussion of it that it makes me feel like we're aware of the possible downside.

Coté:
[39:02] Yeah, well, I mean, so there's there's a there's an interesting thread to pull out of a few of these things here. One, the elections, the U.S. elections last week. And then to a lot of what you're talking about. And there's an interesting connection between them. And I'm going to call it last Thursday. Where like, you know, there's a certain publicly traded company whose share price I watch very closely. It's an exercise to the listeners to think about what it is. And so I have a very, an almost hourly sense of how the AI economy is doing. And in addition to a whole basket of other things, and as they say, the market as a whole, for some, probably because of data center freaking out on Thursday, the market was way down, like, you know, a lot. And then I was looking earlier today and it's back up to what it was on Monday. And and like so this says a couple of things to me. One, I don't know about the smartness of the market. Or you could say, But it did recover. So it just is smart. But I think what that what that gets to is the other thing. It gets to this notion of you have to be really careful about the time period you're doing an analysis with, like what that what that bar in the scrubber is of the time that you have. So as an example related to the elections, there's a lot of reaction. This is like the snotty nosed, like, what's the word I'm looking for?

Coté:
[40:32] What's the word when someone like stares down their nose at you because they were right about something and they're like arrogant about it? Whatever that word is. Like, right. There's some analysis from there of like everyone was worried about democracy, but it clearly works. Look, look what happened. Right. And, you know, that's an annoying thing to hear from someone. But there is there is a perspective in there that's true is like it sure sucks to wait two years. Right. To like do something. And so if if you're worried about things being screwed up, you're not thinking in a two year term period like you're thinking about here and now and you don't have that perspective. So in the same way that like on Thursday, everyone's freaked out. But if you had a longer time period, apparently everything's fine. And then also, as we were talking about, if you're worried about like all of this debt and spending, as you were just saying, it's like, yeah, but you need to think on like a five year term, like term. Right. Like the point is we spend a lot of money now and then, yes, we're making this assumption, but we have a huge amount of demand that if we didn't have the build out, it wouldn't work. And then it all rests on like, are we right about demand being generated, which I think is is the. I don't know. We say this all the time, but like people like you and me who are like, you know, do AI stuff all the time. We're just like, yeah, build fucking more data centers. What's the problem?

Brandon:
[41:59] We want more tokens. We want cheaper tokens.

Coté:
[42:01] Yeah, yeah. Whatever it takes, give us the tokens, right?

Brandon:
[42:04] Well, that's the real, you know, I mean, in some ways, what we talk about in that question, I think is the underlying, because it's really going back to like the demand question. And it's like, does, I mean, I don't know how to say it other than someone's like, does the mass population, right, adopt AI in a way that we think of AI and using it? Like, like the first, like, if I'm given anything new to do, almost, almost anything new related to like work or technology or, and I mean, even personal stuff, like I'm reaching for AI first. Like I am, you know, it's like the first thing I'm going to do is like go in, talk about, or, you know, talk about what I'm trying to achieve. And then from that, build out a plan. And the plan may be build the software, but it also just may give me some ways to be like, you know, I was trying to like this morning, I was trying to figure out like the best credit card, right? For some changes, you know, things I'm going to do. So again, like I'm using, it's like the first tool I reach for, for all kinds of things. right and.

Coté:
[43:04] And i and i bet with you you're not like downloading a pdf and uploading it somewhere you're just like hey go get the shit and then you go get.

Brandon:
[43:11] Lunch and that's where you kind of get to where you just sort of ask that you just start to like push more at it be like this is what i'm thinking about what can you do for me how could you do it and then it will go out and be like i searched the whole web here's what i found and you're like okay well go look for this and then go or build this or show me this comparison what if i did this and like, Like, but like, it's very ingrained. And then of course, in the world of like software, it's like this whole idea of like, you know, yeah, you know, you're just, you know, quoting Adam J.K. Again, we're building the machine to build the machine, right? Like that's how we do things. And, you know, to all the software developers that miss writing code, but like, I feel like it's more satisfying to like do it through, you know, agents, right? Because like, I don't know, it's almost like writing, I know you're like the writer and stuff, but it's like, I don't know, does anyone like proofreading? Does anyone like the mechanics of like, you know, especially when you're writing code and you've got it mostly working and your model and.

Brandon:
[44:06] Your business problems are mostly working, but it's like, oh man, now I got to do the testing and I got to do the usability and I got to do the security. And then it's like, I don't know, this thing's not quite working right. I got to like go in there and like figure out exactly what's wrong with this array. It's like, I don't know, that painstaking stuff, that toil. It's like, I don't know. I never liked it. So it's like being at a level above it and something that can do it better than I can do it. That's why the whole writing thing is always crazy to me. It's like, I don't know. I mean, I'm not, I mean, I'm definitely on the spectrum of like bad at proofreading to begin with. So it's like, to me, it's like having this tool that's like great at it and like correcting any, all my grammar mistakes is like a miracle. And so when people are like but I really want to be in there and be like this is a comma slice and I messed up using the semicolon it's like I'm so far from taking the.

Brandon:
[44:58] You know I'm so far from taking the SATs and like all that stuff like I don't know man I just like it's like yeah just like it'd be great if you could make this sound better and so, You know, but again, like, I know, I just don't know. So this is, I mean, the reason we got on this topic is like, this is all comes back to demand. Do people start to view, to take it in and use it that way? Do they, do they feel that way? If so, you're going to see the demand. And if I'm totally wrong, if people are like, no, like I enjoy the process of proofreading. Like I, I want to do that. Or like, I really find joy in solving the database problems at a low level. And I don't want to do it. Then the demand won't show up, right? Like that people will actually keep it out. So, I mean, that's the ultimate question for all of this.

Coté:
[45:37] Well, you know, in one of our many divisions, the Software Defined Talk Management Consulting, LLC, LP, GVB, we definitely don't do pro bono work. But I think we might be up for some creative compensation. We may or may not accept a low double digit portion of our compensation and tokens. We'll see.

Brandon:
[46:00] That's right.

Coté:
[46:01] You know, we've been talking to the airline industry and we're like, if you give us lifetime platinum status, we could talk about discounts.

Brandon:
[46:08] But before we go, before we get on, like one quick thing, you know, there was somebody at Cisco. Did you, I think you heard the story.

Coté:
[46:14] No, no.

Brandon:
[46:15] The Cisco, like one of the vice presidents or executives at Cisco gave somebody like unlimited tokens to rewrite 98% of the software.

Coté:
[46:24] Yeah, yeah, yeah.

Brandon:
[46:24] I don't know why it's 98%. We talked about it a little bit in the Enterprise AI show, but I wanted to get your take real quick. Like, I mean, so, so like, I mean, cause even, even you and I, I feel like we're very pro AI, even I was like, that seems a little ambitious, like everything. So, um, but I was like, but I also like it. I'm also like, go for it. Like, just like, I mean, I don't know how many agent loops you'd have going, but I'm like, like check in with me in a year and tell me how that went. So I don't, what was your thought when you heard that?

Coté:
[46:50] Oh, well, first of all, that's a great episode to recommend. Uh, it's, it's, uh, it's, uh, Brian really gets into it. And I think in getting into it, he says a lot of smart things. So listeners, if you've listened this far to this episode, you should go listen to that episode. You'll love it.

Brandon:
[47:07] We'll put a link in the show notes.

Coté:
[47:08] Yeah, yeah. And I think... What do I think of that? I mean, I agree with the analysis Brandon and Brandon, Brian had, but that's not what you're asking. I think I think, yeah, like unlimited needs an asterisk by it, which is like, I'm going to say less than two million dollars. Right. Like, like, but that's but like, once you get to two million, just let's come talk and see see what's going on. But I think I think for the project they were talking about, which was if I remember modernizing or rewriting some piece of Cisco software that had been they'd been trying to modernize since.

Brandon:
[47:48] Then they were very precise. Ninety eight percent of the internal software.

Coté:
[47:50] Oh, OK. OK, that's right. Right. Yeah. So that's that's a lot of software. And I think, yeah, at this at this point, I think what you do is, yeah, you're just like, just go for it because we don't actually know. And the issue is we don't actually know how much we should give you. It's impossible to know what the budget should be and how we should evaluate the ROI. This is a secret to all you buyers out there about ROI of enterprise software. No one knows, right? Like you just, you're just told everyone the secret. You're just like making stuff up to make it fit, right? Like it's just like pricing is crazy no matter what, right? In all domains. So, yeah, I think, I think that's, again, separate from what Brian was complaining about. I think that's the approach that you would take minus, I'm going to, I'll put words in his mouth, minus the, the arrogance angle. But like.

Brandon:
[48:47] Well, I think maybe that's the, if we were to like tweak it in our proposal, it'd be like, we'd like the enthusiasm. I would just say it's like, maybe you can involve more than two people. Like it doesn't have to just be two people. Like maybe there are some people here that actually want to get on board with you.

Coté:
[49:00] Absolutely.

Brandon:
[49:00] And that they could help, even if you have only one agent, if you have one super, super agent running all the other agents, maybe having another couple people just like, like helping you i'm not saying it has to be 100 but maybe like 10.

Coté:
[49:14] Right and maybe maybe to reprise a little bit of the cut to preview some more of the conversation it's a um it's almost like a, uh uh a a what word am i looking for it's it's it's like it's like a pre halo effect it's like yeah i think if you gave, a team of people unlimited tokens they would also succeed right like there's There's nothing special about like there's nothing special about the situation other than unlimited budget, which, is actually number one constraint in any corporate endeavor. So maybe the reason nothing has happened is because they don't have unlimited budget.

Brandon:
[49:53] Right. That's or maybe they just don't feel empowered. Right.

Coté:
[49:57] And that's that as well.

Brandon:
[49:58] And so there's a lot to do again.

Coté:
[50:01] So what I was getting at, back to software-defined talk management consulting, LLC, is I want to, this is my initial pitch.

Coté:
[50:11] As always, I think, you know, we always have, we'll use our reference customer, Apple's VP of cables. Lots of work we've done there. Great results, right? But I think, once again, Brandon, you got to get us a meeting, get on the schedule for the CMO of AI, because we've got something to offer. We've got to be like, listen, we've looked at the case, and I think what you need is you need some DevRel. And not DevRel for developers. You need what we're calling PeopleRel.

Coté:
[50:44] And what you need to do is, I'm sure you've got developer relations people who are like, this is how a skill works. Look how we can, we wrote our own version of Git. That's great, love it. You're winning, totally dominated the developer, the DevRel stuff. But when it comes to, I don't know, we haven't figured out what to call it yet. Peeprel doesn't sound good. Prel, definitely don't want that.

Coté:
[51:05] We might just call it normirel, kind of insulting, but whatever, peoplerel. What they're going to do is just go out and do the same kind of thing a DevRel person thought leadership and just show people the wonders of AI. They got to go to the normal events that they're going to go to the county fairs. They're going to have a booth, free corn dogs for anyone. And they're just going to do demos like the Slap Chop guy. You know, this is this is what they're doing. They're doing demos of it and they're going to give out little little cards, give you like a free month of Claude on the, you know, we'll call it the half max. They get a $50 plan. All the tokens they want for a month, they can try it out. And their job, just to show people what Brandon here is talking about, The, uh, the ultimate thing is like, they go out and they're like, here's what you need to do. I want you to install the, uh, the AI thing. And for the next, for the next four hours, I'm going to give you all the corn dogs you want and anything you want to do. All you can do is ask the agent to do it. And I just want you to see how it works. No, you can't go download a PDF and drag it over to there. Not allowed. All you can do is type into this box and let's, We're going to have a workshop and see.

Brandon:
[52:15] I like this. I like this idea. Well, you know, one of the things I've been thinking, I have a much simpler thought, and I've used this example on another podcast, but I was just like, I feel like to help people get up there, it's like you need to have, we need to have like a little lab. And I was calling it like the party planning lab. And like you start with something like this. It's like, you start with something like, you have a bunch of email addresses in like random text in the format, right? And you'd be like, and your goal is to like, I want to have a party. I want to invite all these people. And like what you do is you first show the people, but okay, go to your Claude and be like, ask it to write a parser to get all the email addresses out. And you get that working, right? And then you kind of get everyone's like, oh, okay, that's cool. That's great. Oh, I got all the email addresses. And then the next time you just say, now go back to Claude and be like, don't even ask it to write a parser. Just ask it to see if it can get all the emails. Just ask it directly, get all the emails. And people are like, oh, wow, that was great. I didn't have to do any, I didn't have to write any, I don't know what happened, but it looks like it wrote some code. You're like, don't even worry about it. Who knows what happened? You got all the email addresses. Does it happen to that? And then maybe say like, hey, double check, maybe write something to make sure like you got them all and they're all correct. Okay, and then you do that. And then you go to the next one and it's like.

Brandon:
[53:22] And then you say, and then you, and then you have them like type out an email, and then, and then, and then that, and it's like, okay, now have the thing proofread the email. And you go, well, you know what we could do next? It's like, we can have it just write the email, right? We can have it write the email and then you can say, okay, well now you have the email and you have the email address, you can manually send it, right? So you know what we could do? We can have Claude send the email and the thing to the party. And then you, and you just keep walking up to the end. You restart the whole exercise and you just say like, hey, you say, Claude, like, uh, Help me plan a party for this many people. And here's all, I think all the people I want to invite are in this thing. And just do it like, you know, it's like, you have to almost like reverse the whole thing. Like, let's just go back to the beginning. What if we just ask.

Brandon:
[54:03] The base level question at the beginning rather than like, cause I think that's like the world. Back to the demand question. This is what I want all the people like, the more you walk up this tree, the more you're like, wow, that was impressive. But like people like get lost on the tree. So that's my pitch, right? It's like everyone needs to do something small like that to like feel the power.

Coté:
[54:26] I think if you look at the great DevRel people, like one of our favorites, Josh Long, they do exactly what you said. They're like, I've got this thing I want to do. I'm going to start building it like slice by slice by slice. And each time I do some manual work, I'm going to show you why the framework makes it so you don't have to do that work. And it makes it easy to like add the slice until at the end. I mean, that's a core DevRel like tool right there. And like, I'm remembering, I think it might've been you or to invoke him again, maybe Brian, who this is, this is why, the CMO of AI needs to hire us because of these ideas is like for 2027, I think the other card they're going to give out is like, listen, I know you might feel weird about this, but if you don't, you can use this special activation code and it'll just file your 2027 taxes. And like, I think that's an experience. A lot of people would be weirded out and they would just be like, but you know, this is part of the hard talk, the serious talk that the management consulting people bring is like, they're not your buyers. Like if they're freaked out about that, don't worry about them. Here's a copy of Crossing the Chasm. You can read that when you go to bed tonight. Like, don't worry about the laggards.

Brandon:
[55:44] Yeah, no, taxes is maybe the better. Because it's like, don't just organize. Don't get all the information and organize it so that I can, like, do my taxes. It's like, no, do my taxes, right? It's like, so, I don't know. I hope we get there. And because when that happens, we have the demand. And if we have the demand, that means that all the debt is well spent. And that means stock will go up and all of us will be happy. So that's the end of our fairytale story there. And then along the way, if people, like, have some personal failings, it doesn't really affect us. We're just like, most people are doing weird things and they can be rich, but they don't have to be involved in doing all the work.

Coté:
[56:17] Yeah, yeah. I mean, I think that's the other, that's the last 20 minutes of the meeting that we have. We continue to have in the hallway, you know, they cancel their meeting afterwards. It's like, oh, and also, listen, we got to take care of the weirdo problem, right? Like, these things run apparently. It's almost like, you know how you're taking on a lot of debt to build capacity in the future? This is the human part of it is like, I know the weirdos seem good now, but they're going to be bad two years down the line. So we got to take care of the weirdos. It's going to take you a few years. You got to take care of the weirdo problem. And like, you know, I think probably what you'd end up with, a lot of people, they don't really like it. But one thing that you can do to make sure you don't have weirdos affecting your company is some good bureaucracy. Like it's one of the core functions of bureaucracy is to de-weirdo. Your organization and you know what do we have any bureaucracy this episode.

Brandon:
[57:15] Well, as always, if you would like a sticker, just send your postal address to stickers at softwaredefinedtalk.com, and I'd be happy to send you a sticker anywhere in the world. And also from the community. So last week, and this is going to lead into my other recommendation here in a second. Last week, I got to spend the day out in San Jose at a tech field day, and this is the AI infrastructure six. So I was going to throw it out. Like, I was a delegate with actually some of the people that you know, John Willis, some of the previous guests, and several other people that you know, Kotein. And so what these things are is they basically recruit people that are not traditional analysts, that have some level of expertise, that are interested in typically not working directly for a vendor. Maybe you're a consultant. Maybe you're doing your own thing. Maybe you're doing something small. And they basically bring you out. They don't pay you to do that, but they cover all your travel. And you get to get exposure to both a bunch of other smart delegates as well as get briefed on, in this case, they got briefed on six...

Brandon:
[58:12] Different vendors. Maybe we'll talk about those going forward. It was very interesting. So I thought I'd throw it out, though. They're always looking for tech delegates. And I figured, you know, we have a big community. There are probably people out here that would be interested in doing that. Like I said, it's not a paid gig, but they do cover your travel. So people fly in from far away. So most of them are in California at these events. So if you are someone that feels like you would like to be a delegate, I'll put a link here in the show notes where you can just apply online or if you wish you can find me in the slack or, via email and like just tell me who you are and i'll be happy to probably pass on a recommendation because they're always looking for smart interesting people and i like it too because it's sort of if you will uh i know you're a traditional analyst cote or what were right but this is sort of if you will i don't know what you call it like the pragmatic analyst practitioner analyst people that have kind of been in the trenches so um you'll meet lots of cool people and it's a fun environment so if you're interested it, apply online or hit me up and I'll give you a recommendation and introduce you to the people that run that.

Coté:
[59:11] Yeah, they have one going on at KubeCon, which is happening. And I think we finally achieved, was it Craig McLuckie, who was like, I hope one day KubeCon will, I mean, Kubernetes will just be boring. And I think the fact that we've gone this long without talking about it and you have the biggest tech conference, probably one of the biggest, and we haven't mentioned it, it's now boring. So congratulations. Victory.

Brandon:
[59:32] You won.

Coté:
[59:33] But they have one going on there. And, you know, that's a good,

Coté:
[59:36] I think there's lots of, lots of listeners who would probably do great. And then you get a free trip to KubeCon and things like that. So... Check that out. Well, there's some conferences coming up. Not as many as there usually are. Next week, I'm going to be in London for VMware, Explore London. I think it's called Explore on the Road. I'm going to be speaking on November 18th, but it also goes during the 19th, and I think there's workshops. And then on the 19th, I'll be in Copenhagen giving a talk at Cloud Native Denmark. And then continuing on the Where Kote is Speaking list, I'll be at Build Stuff in Vilnius, Lithuania. And I don't have anything confirmed yet, but I'm still going to be at Configuration Management Camp February 1st to 3rd. You know, FOSDOM is the weekend before that, if you're interested in FOSDOM. I was thinking that might be fun to take one of my kids to it to show them what conferences are like. I don't know. We'll see. Also, another fun place, another place would be cool to take your kids if you wanted to show them what the computer nerd world is like. There's scale 24x in Pasadena, California, April 1st to 4th. And speaking of things that are cool, what's your recommendation this week?

Brandon:
[1:00:57] Well, picking up on where I was last week. So one, well, you don't get paid to go to become I'm a delegate, sometimes they give out gifts, the companies that they'll give you different things. Some of them are good, some swag is fun, all swag is welcome, some of it's really good, some of it's very interesting, but I had never gotten this, Kote. I was given something called the FN pen. Have you ever heard of the FN pen? Do you even know what I'm talking about?

Coté:
[1:01:19] Now is that like a Steely Dan song?

Brandon:
[1:01:23] Well, I don't know about that. I don't think so. But like in John Wick, he carries a pen and he kills a bunch of people with it, right? If you are familiar with that show. And so this group, Stabby Labs, has created essentially, it's technically a pencil, right? They've created the FN pen. And it's basically the most over engineered pencil that you've ever seen. It basically looks exactly like a number two pencil, except it's, I don't even know what it's made of, like some type of steel. It literally feels like a weapon, right? I mean, to the point of when I was given it, I was like, I don't know if they're going to let me take this on the plane home, right? Because it was like, I was like, this is, because it comes with like not only graphite tips that you can use to actually write, but it basically has like, you know, essentially like an end that is, you know, basically, I don't know how to describe it. Like it's basically something you would like puncture somebody, someone or something with, right?

Brandon:
[1:02:16] And then on top of that, it has like a little screw cap at the end where you take off the eraser and you can put on actually a little pen tip as well. And then they gave us like these pen inserts that you can like write underwater with and stuff like this. So to be clear here, this pencil is completely unnecessary, but it's awesome. It is absolutely awesome. I mean, it is just, I don't know what, I don't even use pencils that much. I don't even like you're the one that was always like taking notes and like writing things that has this great carrying case. And so, I mean, again, like from the from the area of like, I don't know why something should like this exist. And just to correct myself, it was that it has the penetrator tip. Right. Which is, I mean, basically, it's just something that you would stab somebody with, I think.

Brandon:
[1:03:03] But again, so I don't know why anyone needs this. I don't know really why it should exist. I'm just here to tell you it's awesome. It's sitting on my desk right now. I love it. When I, in the rare times I'm actually writing, I'm now writing with it. So if you're looking for something that's just fun, fun in a John Wick kind of way, check out the FN, I think this is the 2.0, the FN Pen 2.0. It's definitely a good gift for maybe the person that has everything. And something that if you're of the right type of person, it's one of these things like no one would ever ask for this, but everyone needs one. There you go. That's the FN. And I want to thank, no free ads, but we'll give one a shout out. DDN, they were the ones that gave me this. So I appreciate them. And they gave a great briefing. Maybe we'll talk about what they do on another show. But more importantly, they gave me that, and I will always remember them for that.

Coté:
[1:03:58] I think maybe this is the only pencil you will find that has not one, but two of the product pictures with guns next to them. You've got your daily carry with your, I don't even know what it is, a pistol. And then, of course, you know, as as you do, you've got your sniper rifle over there with your pencil ready to, I'm assuming, hunt some deer or turkey or something.

Brandon:
[1:04:22] It's probably or if you need to just take take notes on what the presenter is talking about. You know, who knows? Maybe you never know what all three of those things are going to happen at the same time.

Coté:
[1:04:31] When you've been flown somewhere to get briefings from six enterprise software vendors, who knows what will happen? Yeah.

Brandon:
[1:04:38] I tell you what, I'm bringing this 10 from now on. I'm prepared to take notes and defend myself.

Coté:
[1:04:44] I think that was the intention. Well, we'll go to a little D&D here. So my recommendation, everyone knows I play a lot of DMs, solo role-playing with AI. And I finally, I thought I would try to crack one of these Wizards of the Coast official adventures that's been vexing me for a while. And that is the wild beyond the witch light. Now, the premise of this, Brandon, I know you're interested, is that there is there's this traveling carnival, that only comes to the same place every eight years. It's there for two days from the Feywild, the land of the fairies and other things like that. And it's landed outside the players area. And of course, there's some intrigue. You go in there, but it's a very, as they say, sandbox open game environment where there's a lot of just like, here's the environment with some plot lines and let the players go in.

Coté:
[1:05:42] That with AI, this kind of adventure is terrible. The AI needs very strict railroads, as they call it, just narrative things.

Coté:
[1:05:49] And so it's kind of difficult to wrap your arms around it. However, I think I've gotten the AI to work well enough. And I think it went through a critical step where it kind of like planned things out earlier. And it's been going very well. I'm in the second stage of playing it with it. It's near perfect, the way that it's working. And I think contrary to what I was just saying, I think if you tune your AIDM correctly, something like the setup of the wild beyond the witch light or witch light, as the insiders say, is it actually is perfect for an AI because what you have set up is basically just like here's an environment with extensive details and the mechanics of the NPCs and a little bit of plot with clocks. They call it like every every 15 minutes of play, the clock advances. And this is the kind of structure that I think a well-tuned AI can deal with. And I don't know. And also, it's kind of a fun adventure because it's also it's from I forget what I forget the the exact year that it's from. He says trying to look it up on wikipedia it's from 2021, so it's from this era where we we we kind of were like where people especially who work on this kind of this book were like what if we were kind, and so you can you can go through this whole adventure without combat there's lots of opportunities but it's easy to go through and just do pure role-playing which is uh kind of interesting.

Coté:
[1:07:15] I guess maybe this is a seven-year-old recommendation, but check it out. The Wild Beyond the Witch Light. It's a fun little adventure. Kind of goofy, but it's all right. Well, speaking of things that are kind of goofy, you've kindly listened to another episode of Software Defined Talk. This is episode 598. In a couple of weeks, we'll have episode 600. That'll be fun. Maybe just in time for Thanksgiving, if I know how the calendar works. But anyhow, if you want to get the show notes for this episode, all sorts of things we didn't talk about and things we did, you can go to softwaredefinedtalk.com slash 598. You can also join our Slack channel where we've got conversation going on all the time, link sharing, little snarky things. If you're into memes, lots of meme posting, things like that. But if you go to softwaredefinedtalk.com slash 598, you can find all of that. And with that, we'll see everyone next time. Bye-bye. I like to say around here, I'll never do this, but maybe I should spend a week naming my agents.