Bản chép lời đầy đủ của Episode 587: The Singularity Didn't Happen on the First

of the show. So Coté (02:20.74) What? Brandon (02:37.403) The setup was they gave him three clips, right? To pick, was it a bot? Was it not? Right. And they were all relatively new episodes. So of course I'm listening to this and I'm I, you know, I'm just this is like how much easier could this be? Right. I mean, you just pick the podcast, right? And Matt, of all the people on this show, who's definitely knows AI, can identify it, knows when he sees slop, knows when he sees it. We know Matt Ray is the number one person to do this. And guess how he did? Guess how he did? He got three clips and he got w two right and one wrong. And how do you think he got the one wrong? Did Coti, this is now a question for you. Did he identify w something as AI that wasn't? Or did he identify AI as something that he thought was real? Which way do think Matt would would miss? Coté (03:24.818) I think identifying AI is real. Brandon (03:29.093) Incorrect. He basically he's it was and funnily enough, it was a clip I was on because I remember it very vividly. It was something along where you said something. We were of course talking about the Pope, a long topic that we that we talk about all the time on the show. Right. And I think I said I said something along the lines of like, you know, the like the Pope would wouldn't cheat on their homework or something. I think this is relation to the and Matt Matt confidently said is like, yeah, I think that's AI generated. We don't talk about the Pope very much. He said something along those lines. Coté (03:30.246) Matt Ray (03:38.574) See I was probably laughing. yeah. Coté (03:39.208) my god! Coté (03:49.638) Mm, mm-hmm, mm-hmm. Coté (03:57.298) Ha ha Brandon (03:57.884) And I was like, Matt does not listen to the show. He was actually on the show with me when I said that. So Matt, I'm gonna now give you a chance to defend yourself. Were you under pressure? Did you just you felt like a little nervous? Like what happened there? Matt Ray (04:00.492) I forgot I I'm sorry. Matt Ray (04:10.481) I mean I mean I I I could I could try to to to blame it on the audio clip. you know, I mean an iPhone held up to a mic, it's hard. But also but also like I don't know, it just sounded so bizarre out of context I guess. I just didn't remember it just yeah. Maybe I was maybe I was laughing because we were, you know, onto something and a and I you know, it was a a one liner that just kinda slipped through the cracks. I I apologize, Brandon. I will pay attention to all your pithy wisdom. Brandon (04:42.747) No needs apologize. I think it was it was an incredible s segment that I almost feel like we should try to steal. But you know, it's their it's their show. I will say the clip he got that was AI generated, that was in your voice, Kote. Can't remember exactly what they said, but it did. It had like a weird, for the longtime listeners, it had like a it was a similar voice to you, but it had like a stutter. It had like a thing in it that you don't you don't do. Yeah, it was like something you clearly was like, this this isn't the baritones I've known to. Coté (04:54.296) yeah? I'm looking forward to it. Matt Ray (05:04.757) It's it's the the pacing, the the the weird pauses. Yeah. Coté (05:05.101) Brandon (05:11.719) calm and love when Kote is speaking. And then I think the first clip was the first clip of yourself. I think maybe the first clip of it was so fun. It was a fun thing to watch. I thought it was hilarious to watch it. And of course Matt gives because he's he's underplaying it but he came honestly I would call it the Elder Statesman. They sort of asked him about like what it means to podcast and why it's a good idea. And Matt gave how to keep momentum and how do you create greatness like we've done here. And he gave he gave away all the secrets. So Matt Ray (05:16.821) Yeah, yeah, and I was like, yeah, I remember saying that Matt Ray (05:33.741) How do you keep momentum? Yeah. Matt Ray (05:40.809) Sorry about that. Yeah. Brandon (05:40.891) If you want to know how to how to do it, we won't do it on this episode. We won't give away the secrets of how we create this incredible content. But if you go listen to AMG cyberspace, I don't have the number here. but if you go listen to that episode, I'm sure you can figure out exactly what Matt said and why it was so great. Matt Ray (05:54.049) Five. Coté (05:59.931) It so it sounds like maybe you are like, you know, you go to whether it's a sales kickoff or a a a a sort of conference, a big conference, and they have like they've got like the guest expert that comes in. You were you were like the guest expert. And you know, it's always I I think I'm sure you were much better prepared, but whenever I see the guest expert come into a big conference or a sales thing, it looks like maybe they haven't done anything to prepare. They they just kind of They just kind of ease in. Like I remember one I saw that that sali was the worst of this, or the highest degree of this, I should say. I was at a Dell World and they had Michael Dell interviewing Elon Musk. And and it it was this is before before he bought Twitter and like a f several years before, like full on full on nuttiness. But it was sort of like it was it was it was almost as good slash bad as like seeing Harrison Ford on Letterman. Matt Ray (06:46.56) Yeah. Coté (06:57.34) Just like like just very like, you know, Michael Dell's up there just trying to make it he's always trying to make it happen, right? Just do do the stuff. And then Elon Musk is just kind of there being weird. Matt Ray (07:05.153) Yeah. Matt Ray (07:11.103) Yeah, yeah. Well, I remember you you got me into to Dell World a couple of years before that, I bet, with Bill Clinton. And and and the dude was slotted for like thirty minutes, forty five minutes, and an hour and a half they had to like walk him off the stage. It was just and it was it was interesting and entertaining for a while, but Brandon (07:32.105) But I

think that's better as a as anyone that like does interviews, you always want someone to like have more to say. You know, it's so much easier, although an hour and a half that's quite a long time. But it's much easier to like cut somebody off than it is to be like, no, we've got we got 30 more minutes and like we are we are done. We are out of I so that's a you're look you're staring at like a a void or an abyss ahead. So so maybe that's what you do from now on. You just have a game in the back pocket, like, you know what we're gonna play, but or not. Let's do it. You know, let's let's have some fun. Coté (07:37.732) Mm, yeah. Matt Ray (07:48.407) Ha ha ha. Coté (07:59.443) Yeah, it'd be better than hearing about this year's storage arrays. I think I think it'd be much, much nicer. Well, you know, pretty soon, I think if if if we listen to the people at Stripe, it's gonna be bot yes, a hundred percent all the time. Cause it sounds like the the Stripe people, they know that the singularity is coming. And my understanding, Matt Ray, you seem like the kind of person who would know what a singularity is, right? And Matt Ray (07:59.755) Need some prep though. Coté (08:28.048) The singularity I think is is isn't that when the AI or AIs I don't know if it becomes a collective, I don't know if this is settled in speculative fiction. If if you have AIs ruling, do they all become one AI or are they separate AIs? We don't know. Anything could happen. Matt Ray (08:44.289) Yeah. I mean th there they're there are kind of like different definitions. I I think I think what I think this version of of the singularity implies that AI has gotten to the point where it's making itself faster better at a you know there's it's like the uptick in the curve of of improvement where you know there's no turning back and from here on out it's just you know more and faster and better. and what that implies is you know some of the news we saw this week like AI making its own chips, AI making you know better AIs that make better And so they you know they keep feeding on themselves to you know to to run off into the future and leave humanity behind. Coté (09:32.666) Is it is this like is this like a meme in the making of like the pimp my ride thing where it's like I heard you liked AI, so I put some AI on your AI? Brandon (09:34.706) Mm-hmm. Matt Ray (09:38.573) Yo dog. yeah, it's I mean i yeah that's that's one way of of thinking about it. Yeah. It it is kind of like well Yeah, AI's pipping its own ride. Brandon (09:42.759) Yeah. Coté (09:49.508) It's when AI pimps its own ride. Brandon (09:53.769) But I just want take a moment. I want to read from the letter. I just want read one sentence from the letter. This is 100% exactly what was written. So it says the singularity. And then it goes on to say, it's a funny, fuzzy, and perhaps already overworked term. But we decided that January first marked the beginning of the singularity. And we've been operating on that basis. And I thought to myself, like, you know, you know how like in science there's always this idea about like false precision, you know, like Coté (10:16.774) Ha ha Brandon (10:20.466) When you multiply two numbers and you know, you can't, you know, eventually right. You have to like round up a little bit. I was like, like, what does it take to like when you're sitting there writing a letter, you're like, I'm gonna pick an absolute specific. I'm surprised I didn't besp pick a specific time. You're like, you don't just say in January, we b it's like January first. Like, and it's just like one of these things. It's it of course, but it's like Coté (10:38.136) One forty seven PM Matt Ray (10:38.558) Well well the I y be th they didn't give the time. They didn't give the time 'cause because they did that in Terminator. They told you when Skynet became self aware. They didn't want to they didn't want to be too on the nose. Brandon (10:47.632) Right. I mean that was I mean that that's actually a good point 'cause they told you the exact time. And it's like, but it's just I don't know. I mean, of course this is from the Collisons who are incredibly smart, you know, they've that's been written up a million different times. So so I I don't know. I just was like, I I don't know if I could ever write a letter like this to like really sit down and be like on this specific date, this 'cause I mean the question is like, well what was like what was going on the thirty first? Like did we didn't not have it? Like like what does it like like Matt Ray (11:12.396) Business as usual. Brandon (11:16.998) You know, I mean, like it's almost like an Independence Day proclamation or something. It's like, what do you mean it happened on the first? Like Yeah, this is twenty twenty six. So this is just w you know, however many months ago that is. So Coté (11:23.15) And and and this was twenty twenty five that it happened? Or no, twenty twenty six. Coté (11:32.634) No time zone specified. We can assume California, I guess. Brandon (11:36.169) probably just assume, yeah, New York New Year's San Francisco time. So it's probably true. So it's just I just found that like I feel this is a very weird sentence. I I I mean, it's even because even say in the first part, it's like fuzzy and perhaps overwork term, right? And then they Matt, I think your definition is a good working definition. It's no better than anything I can come up with. And it's like, but then like it happened on January first, and it just it feels like this feels like you know, it's one of these things where we always talk about the bubble. You know, like are we in a bubble? I'm like, when I start reading stuff like this, that sentence jumped out at me as like, I don't think that's gonna age well. That's just like I just feel like that is I feel like I feel well, I want to say definitively the singularity didn't happen on the first. It either happened before or after. I wanna take I wanna take everything other than the

first. And then I want someone to prove that wrong. Matt Ray (12:13.388) Well, Coté (12:14.415) Yeah. Yeah. Matt Ray (12:26.334) Yeah. Coté (12:26.342) I like this position. Your position is not that it didn't happen, it's that it didn't happen on the first. Brandon (12:31.01) No, I did not happen on that date. I don't believe it did. Matt Ray (12:32.16) No, no, Br Brandon Brandon is definitely playing the the Terminator two quote to the hilt. It says, the system goes online August fourth, nineteen ninety-seven. Human decisions are removed from strategic defense. Skynet begins to learn at a geometric rate. It becomes self-aware at two fourteen AM Eastern Time, August 29th, in a panic. They try to pull the plug. I need a time and let's not follow the Terminator two timeline. Brandon (12:55.324) Okay. Brandon (12:59.922) Well, no one's panicked since January first. I mean, I think we've we've done a lot of podcasting. I mean, there's been panic, but there's nothing like there's been panic for other reasons, but there's I have not seen any like robot flying planes attacking or ro you know, robots from the future show up. I've mostly been like mostly using a lot of the same bad enterprise software I've been using for like most of my life. Right. And I was like, I thought the singularity would deliver at least, you know, a better supplier portal than this piece of crap I had to interact with today. Matt Ray (13:03.32) there's been panic. There's been panic. Matt Ray (13:16.214) They're over in Ukraine and Russia. Coté (13:24.196) No I I is is Stripe is public, right? because I was looking forward to some financial analysts in the Q and A being like, so great, great, great earnings as always. Thank you. Thank you. Good good to hear that. I just just one question here. just for our modeling purposes, when you when you say singularity, is that a is that a V lookup or or is is that more of a a control Y repeat last operation? Brandon (13:31.674) No, private. Still private. Brandon (13:39.7) Yeah. Brandon (13:46.106) Yeah. Coté (13:56.843) could you just I was just want some some clarity on that. But I guess that won't happen. I don't know. I feel like yeah, that seem seems ridiculous. That's the kind of thing where you would say, you know, you take out your big red pen and be like, if this is fuzzy and unclear, perhaps use a different word and be be more descriptive. And th and then also, I mean Stripe is the company that like, you know, the they're they're just the digital version of that big chonking machine you used to use with credit cards, right? To they just they just do like Brandon (14:14.098) Well the other thing Matt Ray (14:22.828) Chunky machine. Brandon (14:23.56) Yep. But the carbon paper, right? And just yeah, I know exactly what you're talking about. There's like five people listening to this and know what we're talking about. But yes, yes, that's exactly right. So but I guess the thing that they go on to say is they one of the reasons they say it is they are seeing a lot of like new company formations because then Stripe has a whole like product line that lets you, I think it's called Atlas that you can, you know, start a new LLC and C Corps and stuff. And so Coté (14:26.684) Payments. Yeah. Yeah. Matt Ray (14:33.696) I know what we're talking about. Brandon (14:51.194) Again, good service. Like I've I've seen it. I've I've I think I've even maybe used it impartially once. And it's like, but I was just wondering like, why is that? Like, would you think more companies being instantiated as a sign of the singularity? If I just put this like on an SAT question, like which of the following are maybe this is the game we're gonna play. Is this a sign of the singularity? And I just make up three things. And I guess one the true thing is like company formations is the right answer. But like, wouldn't you think like curing disease and like solving like incredibly complicated mathematical equations that have been out there for a long I I would think it would be these other things. I think it'd be like these like huge like I don't think I'd have company formations as like the reason Coté (15:27.494) Well yeah, yeah. I mean I mean that's Coté (15:32.999) That that is that is the you know, that's the way a a chonk machine company would think. Is like what what is the shortest path towards more chonking? And and I think you no and and like they're also buying or the they they want to buy open router. And and so I mean I guess I don't use open router, but I guess it just lets you choose a bunch of models that you would use. And then I guess, you know. Brandon (15:47.889) Mm-hmm. Matt Ray (15:48.383) Yeah. Brandon (15:56.019) Well it chooses for you. That's the magic. It it does the routing for you, my friend. Coté (15:59.877) And h however, there's a lot of billing going on. So I think a lot of chunk opportunity in in an open route of thing. And so I I I guess I would assume that the the theory, the rationale of that acquisition is like, well, this is just more shit we can collect a percent on, right? Like in w whether whether it matters if it's AI or not. So maybe maybe by singularity, they just mean there's just a lot of AI out there. And and I guess I I guess the other thing is is so long. Brandon (16:04.593) Right. Coté (16:29.982) As you know, Skynet doesn't make it so that you can't do the chonking. like, sure, you want as much AI out there as possible, pe people using it right and left. Matt Ray (16:38.643) As as long as you're the the tool of choice. So that's that's probably where Stripe's coming from also is like right now when A you know, w when AI is like we need a payment processor, like ninety nine times out of a hundred, it's Stripe. But to them, their judgment day could be when when somebody, you know, replaces them, when when there's like the the T eight hundred or the T one thousand comes for Stripe. Coté (16:50.607) And it it Coté (17:00.078) It it do it does like, you

know, I I don't know, maybe this is just me being older and crotchety, but like I am kinda tired of tech people being all fancy. Like like with with like singularity stuff and like I sort of want to be like, look, man, just do the computers. You don't need to like worry about like civilization and and like culture and stuff. Like generally there's a very bad track record of like Of tech people getting involved in culture and like civilization stuff. There's a great track record of them doing computers. Excellent work there. But let's just sort of like stay in your lane of things. And like, and then so when it comes to like a concept like a singularity, like that's that is a weird category there. But you could also just say, like, we think there's gonna be X trillion dollars in transactions involving AI, and we're gonna collect Matt Ray (17:38.027) Yeah. Brandon (17:38.092) Ha ha. Matt Ray (17:40.939) Stay in your lane. Yeah. Coté (17:59.623) five to ten percent of of those transactions. And that's what we think is happening. Chart goes up and then you don't have to get into what was it vague and hard fuzzy things. Brandon (18:10.738) Fuzzy, fuzzy. Well, that's what's so funny about this letter. In the first page, there's the singularity. Then on the third page, they do exactly what you're talking about. They talk about open router and then they talk about they give a huge projection or just kind of not even project. They they say the economy could reach like these incredibly, you know, huge GDP levels, right? And then basically say that along the way they need agents will start to buy stuff and that they need an op Open router will basically facilitate this, all these transactions. And it's like that part to me, even if you disagree, it's just very it's just good business, right? Like they're talking their book, makes total sense why they would do this. You can have an opinion about whether the router will be the the most important thing or the model will be the most important thing, but very simple business conversation. But I just think it's weird. Like if you say the f singularities on page one, then like kind of the rest of the pages like kind of don't matter. Right. It's like, yeah, like why would the singularity require an open router? I mean, that's the first question. I'd be like, well, what wouldn't it just get rid of all of this? Like it'll just like like and also too, like what would the sing like wouldn't it be smart enough to just like create new models itself and then just use the model? Like why would it be going through a router? Matt Ray (19:07.569) Yeah. Matt Ray (19:14.997) Yeah. Coté (19:24.363) I I think I I I like where you're going. Like I don't I don't think I don't think a post-singularity world needs payments. Like it's just like Brandon (19:31.728) Or certainly doesn't need an open router for you. You know what I mean? I mean it seems like it's gonna solve a different way. Like it I I don't know, that's just the weird it's like the weird read of this whole letter. Matt Ray (19:40.652) I mean I there there's there are some some elements to what they're saying and I don't you know they didn't call them out but like some of the things we've seen recently like with the open AI hacking hugging face. one of the you know, kind of side tangent threads that po popped out is that there have been other AI models that have escaped containment, you know, during you know, hacking exercises that have left notes for other AIs. They've left like, hey, if you get here, here's how here's where to go next and where to get rations and and you know, how to make it through the next leg of your journey. So it's like, well maybe there is a little bit of a robot uprising happening and it's it's just been, you know, hidden in plain sight. Yeah. And then I sh I shared an article in in in in the the backlog yesterday about the tremendous amount of fraud that's happening via Open Router and other routing tools where they're just like hook up to a payment relay and you can get all the tokens you want for, you know, three cents on the dollar. And so you know, maybe the AI just learns to take advantage of of that, leaves itself some notes along the way and rewrites the next generation and and leaves us all behind. Coté (20:57.919) I I like where you might be leading us here, Matt Ray. Is that well hold hold on, is that us humans, we need to start paying attention to what the robots are doing and copy them. So the next time you start leaving notes where the rations are and and what to do, just under the desk, maybe you tape it there. Just little little a little trail of things that only humans can can can find and read. Or at least yeah, just just you know, you got three envelopes in the desk just for humans. Matt Ray (21:00.619) You like this? Matt Ray (21:14.773) Yeah. Matt Ray (21:19.765) Drawer full of snacks. Coté (21:25.242) Not not for the machines to to look at. That that might be now now I wanna let's go over to the investment division. Like if I if if I if I look the the the the you know over there in the hot tubs and the smoking L's and everything. Now if I look at they they they wanna buy it or they did buy it. Wait, so they have how for is it is it what was the price? One point three billion or is that no, eight billion. Right. Matt Ray (21:25.547) Mm. Brandon (21:46.798) Mm-hmm. They've bought open router, yep. Brandon (21:54.0) Eight billion. Mm-hmm. Coté (21:54.896) Now Brandon, is that enough? Brandon (21:58.245) I I think you if you believe that what the enterprises ultimately want, so putting aside the singularity, let's throw that out and just be like, let's just go back to traditional enterprise software. It's like people want access to LLMs. LLMs are providing productivity to people. So you can debate that as much as you want. But we the one thing we know the enterprise would like to do it for less money. So if you believe that, you say, Yeah, you you go ahead, you put

open router in front of all your developers and everyone that's using it. And you just route to the cheapest model and your CFO gets a report at the end of the month and it says, like, you could have paid X with Claude or OpenAI, and you paid a lot less with Open Router, and everybody's just as happy. So, I mean, that part makes sense. I I mean, I get it. And then I don't know, you know, I guess you just have to believe on the agent side, you know, if you believe that there will be lots of agents interacting and ultimately buying things, then I guess. Open router, that's a I think that's the synergy that they quote unquote talk about. It's like you would want open router not only routing to the best model, but to route it to places that it can get the best terms on doing payments, right? That's sort of the next level of this, right? And that would be a nice synergy between Stripe and Open Router. So So we're gonna get a nice edit point here Kote's on mute. So there we go. Hold on, hold on. Let's three, two, one. Matt Ray (23:16.746) Cutter, you're muted. Coté (23:18.566) I think I'm trying Coté (23:22.716) So I've tried to piece together a a a theory here. Now I think I think in the credit card business, they call it they call it the rails, like what you run the, you know, your MasterCards and your visas on. And so if I'm Stripe, because cause I gotta think, unless they have some enterprise business I don't know about, it's not like Stripe is all of a sudden gonna go sell to like the IT department and people who want to like manage access for all their people to to AI to optimize it, right? They're probably like, what? Like this that's not even their but if we were to say, okay, a huge percentage of all commerce is going to be happening post singularity in the the AI world. And in the same way that all the major major credit card companies have these rails around, we're establishing the network that payments go through. And so therefore we win. And and that's that that Brandon (24:13.926) Right. And the token. But and really think of it. And also we're optimizing the token cost for you. Right. That's you know, I mean it's a I think it's those two in conjunction, right? Coté (24:20.154) Yeah, yeah, yeah. It it it just it just seems like the optimizing the token cost for you is like a whole other top t so like yes, but it's kind of like rem remember well, my my my my current company back in the twenty tens, they were like building Microsoft Office and they stopped doing that. It seems like it seems like a venture into like something you don't quite do. and and it does Brandon (24:39.368) Mm-hmm. Yeah. Matt Ray (24:44.986) I I I I I see their angle though. I mean you know, if if you looked at if you look at the the FinOps Foundation, they're already pivoting towards token economics. You know, so so I I the the paths of of finance and tokens are very, very tightly coupled. I just I just think that there's like this massive chasm of pricing that's disrupting it from the top and bottom. Coté (25:04.102) Yeah, yeah. I I Coté (25:12.658) I guess I guess I guess what you two are saying that I should consider is if we start if we start thinking about tokens as cash or a currency, then we've got something here. And then also we don't have to think about blockchain stuff anymore. That's over there. Bit Bitcoin don't have to think about it anymore. Matt Ray (25:20.744) Yes. Yep. Brandon (25:20.776) Mm-hmm. Brandon (25:27.464) Well, I w I do I didn't w I don't want to burst your bubble. They do mention stable coins in the in the I don't that's a whole another part of the letter. I wasn't gonna get into that, but like that's in there as well. They're seeing the rise of stable coins as well. Matt Ray (25:27.914) Did I I'm sure Stripe made some acquisitions there too, right? Coté (25:34.882) Ugh no. Coté (25:39.437) May Okay, well, maybe there's some synergies there that I don't know about. Get these two things together, sounds good. Love it. Well, meanwhile, it it looks it looks like Hugging Face also kind of a so similar sort of thing, but maybe they are are gonna be acquired for much more. It's like like what do you think, Brandon? Are they hugging face is worth more than eight billion? Is it more valuable? What's what's the what's what's the situation there? Brandon (25:43.943) Okay. Brandon (26:03.836) Well, rumors it's it's thirteen billion is what's what's on the table. So I mean they last raised at four point five billion. So they've had quite a few months, you know, to basically tripling their value. And I don't I don't know the thing that comes to mind from here is it's kind of like GitHub, I guess, in my in my sense of like like I don't know how valuable the it is to just have the marketplace of like here's where all the models sit. But but GitHub is pr I don't know. I that's another qu I mean it's an open question itself. It's like Matt Ray (26:25.46) Right. Brandon (26:32.296) I mean, it's been bought by Microsoft. It feels like it's done well. I don't know. I guess we don't really know what the revenue of GitHub is, but it feels like it's it feels like Microsoft that feels like a good thing that Microsoft bought it. I don't know. Maybe not. Matt, do you would you have an opinion on Matt Ray (26:38.922) I mean Matt Ray (26:45.556) For for for Microsoft or for the world? Brandon (26:48.934) Well, I guess for Microsoft first and then the world second. So what what what like what what would you say? Matt Ray (26:51.626) I mean I I I yeah, I Huggy Face is definitely positioning themselves as we want to be, you know, the GitHub of AI models. I don't know their business plan. I mean, is is it is it that they provide like you know, access and c access control for enterprises and that's how they're like, you know, you can only take these certain models or or there are some models are commercially listed and, you know Brandon (27:21.574) Well, that's kinda why I was using the whole GitHub analogy.

I think it's sort of like you have all the people. I mean, which is in itself important. You have like this is where everyone goes to get the models. They really haven't, you know, I don't think they figured out how they're gonna make money, right? They haven't done their GitHub professional like if I just gave it you today, gave it to you today and said, Okay, well, you just you can just harvest the profits, well, you'd be like, There are no profits yet. Like we have to like do something. So you have I mean, you could say, and there's lots of things you could do, just like with GitHub, right? Just all these different Matt Ray (27:29.202) Yeah, yeah. Brandon (27:50.545) you know, secondary services you could offer and and things like that. So I could see where they're coming from, but at the same time, I it's one of those things that's like, I d I don't, you know, I don't know, maybe I undervalue like the idea of being the marketplace, you know, or actually being the central location. I don't know, Coach, what what do you think? Is that like some maybe Ben Thompson would be talking about aggregation theory here? It's like, no, it's incredibly valuable to be the place where everybody goes. Like even if people complain about GitHub now, like we're all still using it. Like you know, I know that was a new thing from Matt Ray (28:02.324) Yeah. Coté (28:12.302) Mm. Yeah. Matt Ray (28:16.019) Mm-hmm. Brandon (28:17.958) What the cursor team, they they had something else. I'm like, I'm I was like, I'm good. I'll just complain about GitHub and continue to use it. So what what do you think? Matt Ray (28:24.057) Yeah. Coté (28:24.144) Well well, it's it i it I mean with with GitHub, you at least like you host your source code there and run your projects out of there and there's like the whole process of doing things. Whereas if you just have a directory, it's it it it's it seems like with AI stuff, like sure, you need to host like four to twenty gigabyte files and then have like ratings and stuff. But if you're not like running it, who cares? Right? Like It's this y just having a a a a a directory you look things up doesn't seem like a whole lot. Now, I'm sure I'm missing a lot of of what's going on there. But you would have to provide something in addition to that, because I mean iron not to your question, you could just host that on GitHub, basically. Right. And so but with GitHub, like it's a different thing to hold ho have your source control, all your issues, your pages, like the it the you know, a better source forge, basically. Brandon (29:10.76) Mm. Brandon (29:20.87) Yeah, and to be fair, you really couldn't just throw like a model on Git. I'm I don't think that would go well. But that probably would would end and they and I also want to correct myself, they do have a pro plan for nine dollars a month and a team plan for twenty dollars a month. So I think they are maybe they're like GitHub circa like six, eight years ago. Matt Ray (29:36.259) they

they also provide they they will run models for you and and so that it's almost like paid GitHub runners where you know you can have you know I like this you know I like this Quen model over here can you run and host it for me and give me access so they're they're you know sticking their toes into the the neo cloud space or whatever they call that Coté (29:45.744) Mm. All right, all right. Coté (29:55.45) Coté (29:59.965) Sure, yeah. That's good. Brandon (30:00.211) So looking at it here, I was just double checking. It looks like they're doing like, I don't know, something around a hundred million, two hundred and fifty million dollars and growing quickly. So I mean, you know, I I guess it's like I guess it's maybe it's just a protection bet, right? Like we talk about this all the time. It's like no one's gonna I mean, if you have infinite money, if you're one of these companies with infinite money, it's like just buy it as insurance. Like fine. If it's you know, if it becomes huge, you bought it. If not, then then just roll with it. Matt Ray (30:21.693) Well but but also you know if you're one of the the frontier model providers, if you buy this it becomes like the you know, the poison fruit, right? So you know, it it's like when when Microsoft bought GitHub Everyone was like, it's only gonna be Azure only, right? Or they somehow they're gonna try to make it like tightly coupled to to Azure. And that was like the big worry. They haven't done that, you know, but there are there are a few little tendrils. But if you are if you are open AI and you buy this, you know, does that just you're you're owning the market that competes with you? Brandon (30:59.142) Well, maybe this is maybe back to the Stripe then. Is Stripe the place to buy it? Is Stripe the one that like they buy it and they throw Hugging Face and Open Router together? And then you're sort of like, We're the I don't know, we're the model mall. Like we got all your models here and like we can route you around and and then that that wouldn't make the most sense. Cause I mean, the only way you would say open II or anthropic would just be sort of like protecting their massive trillion dollar valuations, like let's just buy it and just confuse everyone. Like, who cares? If four billion dollars Matt Ray (31:01.136) We need to know who's buying it. Yeah. Matt Ray (31:08.935) Or Microsoft. Yeah. And the pricing? Matt Ray (31:27.335) Ha ha. Brandon (31:28.53) And lets me IPO for two trillion, like who cares? It's like I mean, that's pocket money, right? Like so. So mm hmm. Matt Ray (31:31.855) Yeah. It doesn't matter to them. It's not real money. Coté (31:35.045) Well, open AI did buy that that Silicon Valley news show that talked about other companies. So maybe this would be an extension of that strategy where we're we we talk and and we work in in other areas with great hair and and not just promoting our own. So they've got that that down there. But yeah, maybe so. Yeah, I th I think you're right. The Stripe should just aggregate all this stuff and become an AI company. Brandon (31:40.199) Yeah. Brandon (32:02.384) It seems like the most logical thing. Like if you're gonna do it and the singularity is here, right? Why not? Go all in, like, you know, have have all the models, have the router, put them all together and just become the model marketplace and you know, offer it up at a v a very attractive fee. And then maybe that's to your point, Kote. It's like that's where you send out your enterprise salespeople to just be like, Hey, everybody needs this. Everybody needs to be. You don't go directly to the models, you come to to open router and then you Would go convince Anthropic and OpenAI. They have to put their models in open open router as well. So so maybe that's your world domination singularity message. Coté (32:38.438) That's that's what the singularity is. Well, meanwhile, I think you know, Thanksgiving's coming up in several months. And it sounds like all of us tech people are gonna get a an earful of AI is terrible and ruining the world. I d I don't know I don't know about you two. Matt Ray (32:51.849) Are you having Thanksgiving with Ed Zitron? Coté (32:55.602) But yeah, I I I've I've seen a I I don't know the exact number, but I've seen it the the the surveys that they come out every I don't know, every month or so, the ones that I see, it seems like there's always a percent or two growth. There are I haven't looked at global surveys like I sit around looking at surveys, but it's they're probably all US surveys. But it looks like well over, you know, we're kind of fifty to sixty percent of people surveyed just hate AI. Brandon (32:55.911) Mm. Coté (33:23.74) Just straight up. They don't like it. So I I I I don't know what to do about that. Am I gonna have to not talk about AI at all? Like even here in my house I mentioned using AI and my kids kinda look at me funny and and they're they're not into it. So I think Matt Ray (33:32.585) No Matt Ray (33:38.375) Ooh. No no politics, no religion, no AI. That's the Thanksgiving rules. Coté (33:43.758) Yeah. Yeah. Yeah. What what's what's what what what are you guys seeing in the in the surveys, the people you talk with? You go eat your your deli meat at the grocery store. What's the reply? Brandon (33:52.177) The only thing I I mean the best I can tell is just and cause there's you know, we probably won't get into it. There's a bunch of Facebook I guess lawsuits settled today around like your Instagram and a lot of I mean, and I think some of this is a stand in for that. There's just a lot of frustration around, you know, what quote unquote social media's done. And I think maybe it's just this is just the simplest way to it's just a shortcut, right? It's just like I don't like data centers, I don't like AI. It's just a a nice shortcut to be like I'm frustrated with like what tech is doing. because I mean I you know it's funny when I read all these things, I'm always like, man,

I've I always feel like it's taken like so many of these tasks that I don't like to do and made it, you know, even putting aside the singularity, I'm just talking about just like simple productivity. I can get my work done faster. I I I don't have this feeling at all. Like I it's been nothing but helpful to me. So I don't know. I guess, but it if people are frustrated, right? And it's just I guess it's an easy way to label something and just an easy way to be mad. what do you see, Matt? Matt Ray (34:51.491) I mean I definitely see a lot of it it's just an extra layer of crap at the bottom end, right? Like everyone was already used to dealing with spam. And now we're yeah, and robocalls. And it's just like you're starting to see this continued grit in in society of just more crap. And and yes, at the top end, there's great stuff. And and you know, for for for folks who deal with, you know, who who work in email or who who work in phones are like, yeah, I can call so many more people now and I can send some I can reach out and I get response rates and I get all this like high high you know high value. But at the other end it's like, Yeah, but like I'm having to deal with the higher volume of just automated annoyances. And I think I think that with the you know, the the societal of like, well, they're also just they want to take all your jobs and they want to poison your air and take and make your electricity bills higher. I kind of see where people are coming from. Brandon (35:54.077) Yeah, I guess I it's just one of the things too. I I don't f I always just feel like have have any of these things really like happened? Like that's the part I guess I don't get. It's like Matt Ray (36:00.339) Well, electrophills have gone up for everybody. Brandon (36:03.218) I mean, but I I guess it's like it doesn't feel like it's gone up significantly more than like last year or whatnot. It goes up every year a little bit. I don't know. So I don't feel like I don't I'm sure somebody has lived in a place where like somewhere they it's come in and it's gone up significantly. But there's been plenty of places where they built data centers and it's gone down, right? They basically just paid people. So it's fine. I'm not like telling listen, if people are mad, I'm not telling them they can't be mad, but I just I kinda come back to like, I don't know, there's just a lot of stuff you can get done faster. And I and I you know the one the other part I do see, which I think is interesting because like The group that in theory that is like maybe more productive but less happy is are just developers. Right. I think you do see fair enough people kind of post online just anecdotally. They're like, you know, I miss writing code. I would, you know, it was better when I was just writing code. Now I'm just like managing an agent. I'm not doing it. I don't feel as productive, even though they're, you know, outputting, you know, by some metric, you know, more things. And so maybe that's like the Derek Thompson thing. It's just sort of like if you developed a craft and you don't that craft is you feel like the craft isn't needed as much or it isn't something you can practice as much that you just there's just like some natural depression. Right. And I guess and maybe that that exists for people as well. Right. If you know I just really like doing this thing and it's just not as valuable. But again, I mean I always think that's offset by so much is like, well now it's like for someone that's sort of like I never really had time to do that and like build stuff. I'm like, wow, this is awesome. I can build stuff all day long. And it's like works and it does stuff. So Matt Ray (37:04.85) Yeah. Matt Ray (37:17.404) Well Matt Ray (37:23.154) Mm-hmm. Matt Ray (37:28.924) Well, I'm yeah, I mean it's the perspective of like all the taxi drivers who got displaced by Uber who are gonna get displaced by Robotaxis, right? And so it's like, the thing that defined my career for, you know, thirty years is now gone. Brandon (37:29.448) So I guess it's perspective. Brandon (37:38.63) Right. Coté (37:46.661) Replaced by a phone. Brandon (37:48.498) Yeah. Matt Ray (37:48.956) Ha ha. Coté (37:50.161) Well, it it is, you know, there is this weird I don't know, it's more than more than three points to it. But on on on the one hand, you got what Brandon is saying, like, v I'm very productive, very empowered, get to do a lot of stuff. And and then you're over here, the another point of the triangle, which is like, we should burn this all to the ground and send our P to data centers. So so you got that crowd. And then there's like the continuous, like Is does lub lubrious mean like you're overly happy? I should go look that up. There's a continuous like overhappiness of like the enterprisey space about how it's great. But then they're always like, I'm not really sure if this is paying off. I can't really judge like if if things are better or not. Like there was there was like a new McKinsey thing that came out. And of course, you know, despite all the happy talk, the way that like higher level executives at companies and quote unquote share shareholders are probably going to judge the value of AI is how many people can you fire, which feeds the the peace enders over there. It kind of validates what they want to do. But yeah, the the I I think I think in in this this survey, it was like in twenty twenty five, people like people thought like thirty two percent of people thought that they were gonna be laying people off was gonna be like a big thing they were gonna do with AI. But then in in recently surveyed, only fourteen percent of them actually did it. So like there's this way overestimating of like I don't know, what the effect of things are gonna be is. And so it is it makes it equal it makes it confusing because it's hard to hold up and say like, this is paying off and doing great. And then there's Bandon over here who's like, this is the most

productive I've ever been. And then you got the people who who are upset. I I I don't know. It's it's a bizarre situation. Brandon (39:35.314) Yeah. But what I said this, you know, what if I just said to you, like what do we because you know the there's a good note on here, the haters guide, haters manifesto from Zift Ed, our good friend Ed. And you know, he kind of says something like, you know, he's it changes tune slightly, still very negative. But he says, you know, LLMs are fine for small contained tasks, but dangerous and useless once you outsource them to for the judgment. So but what if we just sort of like cause I think that's when he says small contained tasks, they've just said like What if just change that to like AI is like a really good tool? This is a really good tool to have in the toolbox, right? Still need human beings to drive it, but like, I mean, for like normal people that are like frustrated when they set up their monitor or they can't get on the Zoom call. Well, there's just a a they can either say it or a a chat windows like set up my computer so I can get on my Zoom call, right? Like just all these frustrations that like I see like normal people interacting with computers with all the time, right? Coté (40:11.964) Yeah. Brandon (40:34.568) It's like, what if it's just a tool that like just made all that stuff easier? Like the literally the human being could just tell the computer or be like, do this for me, right? You know what I mean? Or like convert this, just convert this to a PDF, right? Or make this thing, right? Or like there's like, I don't know, I just feel like there's like a limitless amount of modernization of software that could take place that can make all our lives a lot easier. And it's so, and it's and it's like what Ed would say here is just a small thing. Like, Matt Ray (40:43.784) Mm-hmm. Coté (40:44.604) Just automation. Brandon (41:03.162) What if someone just like sent you your tax bill and it was just like here it's all done, right? What I mean, just all these things that computers could do for you, right? Like, here, let me fill out the form for you that you need to send in for the doctor or the permission. So I just think like all of these, like just simple stuff. I'm like, this feels incredibly valuable to me. Like Coté (41:20.954) Or or like attend this meeting that you didn't want to go to where nothing was gonna happen. Brandon (41:25.37) Or summarize it or like just like or even you know, everyone's about writing. It's like proofread my email, make my email more concise. Like I I mean, everyone's like, writing's going away. It's like, listen, man, we're sending we've we've not had AI and we've been sending bad emails for like decades. You know, and humans have been producing poor text since the beginning of time. Why can't we have something to help us like make it a little bit more succinct? You know what I mean? And it's just like, so I just look at like these when when someone says, Well,

it's just these small things, I'm like, this seems like pretty Matt Ray (41:26.072) Matt Ray (41:32.178) Yeah. Brandon (41:54.227) Pretty incredible, especially when you look at like normal people interacting with with systems and the amount of frustration people have canceling their their sell their cable service. And like I could go on and on. And I'm just like, that seems like a really big win. I don't know if it's worth two trillion dollars or if that's the singularity, but it feels really good to me. Feels like there's a lot of things we could make better for in the coming decade. Matt Ray (42:18.663) And I I I think that's why everyone points to Apple and they're like, if they can unlock this for everybody else Yeah, I I think that's that that that feels why there's a gap is most people haven't put the time into the tooling into the harnesses, into, you know, d eating living this lifestyle and and you know, day in, day out, like, AI do this for me, do this for me, do this for me. And instead we end up with like, it kinda did a crappy job and I don't know the tool that well and so I'll just, you know, it does some things for me. And then I see a lot of other noise out there and I I'm not feeling the benefits. But what it needs to be more accessible. That's that's the l limiting factor, I think. Coté (43:01.786) It's it sounds like the I forget the name of the the the fella over there who's running Apple now. But but that that person needs to call up Derek. Is it Derek over at Stripe? No, Yeah. Tell me about this singularity, brother. What's what's going on here? Maybe maybe if they can get the singularity rolling through Cupertino there, we'll finally use the go to market power of Apple to to deliver it. Brandon (43:11.41) Derek Thompson? Or no? Patrick Collison or one of the Collison brothers? Brandon (43:27.282) Do you think you know, I mean, because sometimes I just think like, what have we just talked about a little differently? It's okay, like the tools and stuff, very realistic. I think lots of things that we can make better and will get better just with the normal development of software. Then I also think you could, if you want to make the singularity cases, just say, like, I feel like you could say, like, this may enable like new scientific research, but then it's sort of like no one needs to get out of their lane. It's like, let's give it to the scientist, right? Like, I have no idea. Like, maybe when they tokenize, like all of the genome and all the molecules and the physics and there's like a world model. Like you could I mean I you could see that coming up with incredible scientific breakthroughs. Maybe that will happen. But like I feel like that kind of stuff is where it's like nobody knows, right? Like and I don't think we have to like it it feels like it'd be okay for us to say like we don't know. It's just an interesting new tool and field of research that scientists are going to use across I would assume every scientific discipline and probably many Matt Ray (44:10.268) Mm-hmm. Brandon (44:25.318) New things will be discovered that are hard for us to understand right now. And it's like, it's okay if we don't have the answer to like, well, what's gonna happen there? Coté (44:28.239) What's Coté (44:33.094) Well let may maybe we can push we can do a little do a little PR consulting here. We can push the tools idea a little further and say, like, all right, we're let's focus on the outcome of of better tools. Cause yes, better tools, we love better tools. Maybe the messaging that the the singularity people need to have is like, what if a 35 hour work week? That's what AI is gonna do for us. And then I bet I bet all those people would just put their P in the toilet. instead of shipping it to a data center. They would be like, sign me up. Love it. So maybe maybe it's like a simple, simple transference. And you could go you could go to to, as you say, our good friend Ed and be like, what, you don't want to work thirty five hours a week? Like like what what are you opposed to here? Matt Ray (45:17.799) Well, I I I think that's I th I think I think that's part of the the the haters manifesto is it's not that you get to work thirty five hours a week, it's that you don't get to work. Coté (45:30.074) Well, they gotta change the message. Matt Ray (45:31.367) I'm just saying, like, I I mean, you know, he's he's taking the the the most negative path and it's like, yeah, like employers they don't want to pay you less. They don't want to pay you. That's his take on it. Brandon (45:32.76) Mm-hmm. Coté (45:34.938) And Coté (45:46.375) Th this this th this is this is true. I like you know, I I th you gotta you gotta wangle it through through the brand and thing is like, hold on, we just got better tools. The tools are better. That's it. And and then because we have better tools, you can go home a little earlier, an hour earlier every day. Everything will be cool. Matt Ray (45:56.581) Yeah. No need to lay everybody off. Matt Ray (46:04.337) That would be great. Brandon (46:04.412) Well, I don't know, I'm sure we have to wrap it in a minute, but I I guess I just kind of come back to like I think sometimes Ed and all the people like just underestimate like inertia. Like let's take funny, let's just take banking for once. They're always like, you know, they're not gonna need all these junior analysts to like do spreadsheets and make pictures books. And it's like, here's the honest truth. Like, we probably don't need that today. Like, no one really reads the pitch books, to be totally honest. But that's the process this investment bank has basically decided this is how we indoctrinate new people. And we keep it going because like investment banks existed all before Excel, right? All of this existed before. And so what will happen I mean, to me, it's like, and they're not no one's gonna be like, I just want, I just want everybody gone from the bank. I want nobody but at the bank but

agents. Like the leaders of the bank don't want that, right? They want there to be a pyramid structure. So so maybe it will be. Maybe the next generation will be using LLM tools to create like 3D videos that replace pitch books, and like no one will watch that anyway, either. And like deals will still get done the way they're done today, like. couple partners get together, they meet with the executives, there's a pitch, there's probably a steak dinner, there's probably some talk. There like a couple of leave behinds, right? That are and like it just feels like that inertia of like how humans do business, like, I don't think that's changing. Like and and Coté (47:18.332) I think I think I think you're right. If if you didn't want to talk to a junior analyst, you're not gonna want to talk to a robot, right? Like you and and and I get now, on the other hand, we got the singularity where it might be robots talking to robots. And so we're just gonna have to pick up what we can. Now, maybe maybe a good book into this is I think this might be the last sentence. I need to go check. the last sentence in that McKinsey PDF. But it's it says something Brandon (47:23.558) Right. And it's like Coté (47:45.159) I don't know. Almost wickedly delightful. This is a whole survey about like how executives are saying AI is helping or not helping them, about their predictions about firing people or reducing headcount, as as they say in the business. And then at the end it says, and I'll read it, it says, this little coda despite increasing expectations for reduced workforces, most respondents do not see AI as a threat to their own careers. Just thirteen percent said, quote AI makes me feel anxious about my career prospects, quote, unquote. Which seems like it's kind of odd that the people surveyed who are figuring out the fate of workers' lives based on AI are like, no, I'm totally fine. What whatever whatever it's doing and whatever I'm doing, not replaceable. Couldn't couldn't couldn't do what I'm doing. Nope. So it it it is it it is a strange thing. Well, I just wanted to highlight one thing Brandon (48:30.268) Not replaceable. None of us are replaceable. Coté (48:40.304) that I think is a good result of of of using AI. And that is Whitney and I talk with him. We got an upcoming interview with with this guy, Victor, and he made these two great little films. And I talked with him last night, of course, he used AI to make them. Lots of effort went into it. But it's basically the the history of infrastructure is code and then the history of Docker. And I I w I won't give it up, but they're fun. They're like little little fun videos to watch. And they're not if if you ever watch Victor's stuff or or heard interviews with them, they're a little more kind and gentle because they're bedtime stories than than what you might be expecting. But good good job, AI. Came came out with a good product there. Brandon (49:23.078) And maybe this fables, I mean the one, the history of

the forty years of inf infrastructure as code is really good. And and maybe that is just a fable that gets told over and over again. It's like every time the new thing is gonna fix all the problems, but the new thing doesn't fix all the problems. So so and that's the and I guess sort of that's where he ends. The last one is AI agents. So so maybe that's a little fable for us all. Coté (49:35.634) Yeah. Coté (49:40.463) It's good stuff. Yeah. And, you know, it's gonna be oof, maybe six weeks in the future if I think about the schedule. But if you wanna hear that interview and hear how he made it, you should go add software defined interviews to your podcast and eventually in five or so weeks you'll hear the interview of Brandon (49:58.653) Well y Cote, you should say there are plenty of interviews for them to listen to to keep them busy. Right. I mean, you've got a lot. You got a full backlog there. Coté (50:02.2) yeah. Th the there'll be several. Yeah, we've got we've got lots of great interviews in there of of people in the tech community. And it's a lot of it's a it's it's a lot less of like what what do you do at your company and more just like what do you do? What's what's up with you and how how that stuff works. So that's enjoyable. Now speaking of how AI can help us, there's one thing I think, you know, if we know anything about humans, it's that they just need something to fucking do. Matt Ray (50:19.088) What's up with you? Coté (50:32.006) during the day. So a lot of some people might call it bullshit work, others might call it living. And they just you gotta do something, you gotta stay off the streets. But there is something that I think we should strive to never replace, and that is bureaucracy. Do we have any Brandon? Brandon (50:49.148) bureaucracy is if you would like a sticker to send your postal address to stickers at software defined talk dot com like Chris did during the show. In fact I I haven't even sent him one yet. So Chris, who's in looks like Pennsylvania, I promise to send you a sticker. And then if you would like a sticker as we get into the the school season, as as I say, you know your water bottle is not complete until you have a software defined talk sticker on it. So in fact my son was asking I was at the fundraiser last night. And you know, one of the things the pitches they were making is like corporate sponsors, right? They're like, why don't you and my son was like, Why don't some software-defined talk become a corporate sponsor? So maybe that maybe that's our next thing as we break into the singularity. We'll maybe we'll start sponsoring schools and give them stickers so everyone has it on their bottles. That's that's a market that's untapped. We'll see. We'll see if we'll see if the business development team likes that idea going forward. Coté (51:41.936) Well, you know what they said in podcasting school, Brandon? children are future listeners. Brandon (51:48.297) That's right. That's right. Maybe that's what I do. I will say in the cafeteries, I'm gonna get everyone in the school together and be like, first of all, I gave you guys a hundred dollars. Now I need every one of you to open up your phones and download software to find talk. Maybe that will be the quid pro quo. So Matt Ray (51:54.374) Ha ha. Matt Ray (52:00.25) Hmm. I I I I didn't I didn't share that that guidance with the the AMG Cyber Chat folks. Coté (52:01.884) Yeah. Brandon (52:09.446) Matt, what are we gonna do for you? Coté (52:11.382) You know, we gotta keep some stuff proprietor proprietary. And other otherwise we don't want to level the playing field. It'd be so too much work. Well, also lots of conferences. I there's many DevOps days, which I'll go through. but we also have We Are Developers, North America coming up pretty shortly here, about a month out, September 23rd to 25th. Now, last week I talked about how if you if you were to look at like Matt Ray (52:13.125) Huh. Coté (52:40.24) The the way European conferences are run, very well attended, very broad and general. And the We Are Developers people, they posted the topics that they have up on their site there. Hopefully you went and and browse there. Now, I think the other part that's always good about a conference is what they call the hallway track, the people who show up there, who you can talk with. And that, I think in this this area, at a conference that's composed like we are developers, is is interesting. Is it's not just focused on one. type of technology or one stack or one part of the stack. You've got topics across the stack. So you can expect a whole bunch of people coming there of different roles, different attitudes, hopefully happy attitudes, perspectives, backgrounds, foundations. But also, you of course have it in San Jose, which means there's a a good concentration of like tech world people. There's plenty of other like non-tech world people there. But Matt Ray (53:24.24) Perspectives. Coté (53:38.331) I suspect if you go there, if you look around, talk with people, be lots of interesting people to talk with. Now, if you're interested in those talks, those topics, and then people who are in your your sort of community peers you could discuss with, you should go to We Are Developers North America in San Jose, September twenty-third to twenty-fifth. Now, if you want to buy a ticket, you should use the code dev pod fifty to get probably fifty percent off, or maybe fifty dollars off. You won't know. Until your check. It's it's a it's a good thing to to look into. Also, there might be anywhere between zero and twenty-five free tickets left in the pool they've given us. So if you want to try to get one of those, you can go to softwaredefined talk.com slash five eight seven and look at that. Now, also before you go there, you might want to go to DevOps Days Gros September 4th to 5th. There's also the Cloud Foundry Summit. It's gonna be right before. We are Developers North America September 24th to 22nd. So

you might want to leave early from Heidelberg on the 22nd, dri take the train or drive up to Frankfurt, direct flight to San Jose. I don't know if that's possible. You might have to call up NetJets, but you can get over there and and attend that conference. I'll be speaking there in Heidelberg. And then there's DevOps Days Rockies, September 22nd and 23rd. Sadly conflicting with Cloud Foundry Day. So hopefully people show up to that one. and then there's also DevOps Days Dallas, September twenty eighth and twenty-ninth. You've got DevOps Days Vilnew, September 30th to October 1st. And then there's DevOps Days Prague, October 5th, where I'll be speaking. I I found out recently. And then I'll also be at DevOps Days Istanbul, October 24th in Istanbul. then there's also the VMware users group in Orlando, October 20th to 22nd. The c and then then I'll also be speaking at Cloud Native Denmark, November 19th. Now I recently, you'll be happy to know Brandon. I received a picture from one of the conference organizers who has a sticker on their laptop of this podcast. So Brandon (55:38.769) love that. Denmark's my you know what? Everyone needs to go to Denmark. That's my favorite then. You know? That those I've always said the Denmark people are my favorite people. I haven't I've said that every time on the show. Where are you going? Copenhagen? I've heard that's a good place to be. That sounds like a good that's a good city, right? So all right. When are we going? November? It could be cold though. Is it not too cold by then? Is it real cold? Is it gonna be kind of cold? All right, we'll bring everyone bring a sweater. Matt Ray (55:43.365) Yeah. Coté (55:43.57) Yeah. Coté (55:48.709) Yeah. Yep, they know how to operate a sticker. Yeah, yeah, Copenhagen. Yeah, there there's a big Hans Christian Andersen statue there. Maybe put a sticker on that. Mm. Probably chilly and wet. Brandon (56:06.704) So yeah, it sounds like that would be a good one. And put a sticker on your jacket or something. You know, that way they know. That way they that way they know who you are. Mm-hmm. Coté (56:11.428) Yeah. Yeah, you can find each other. And then and then I'll be speaking at build stuff also in Vilnew, December second to fourth. And then just as a as as as a I don't know, community service, the configuration management camp just launched their opened their CFP and they have the dates for their conference. February first to third, great conference. I try to go every year in Ghent. You should go. And then there's also Scale 24X in Pasadena, California, April first to fourth. So you can plan what you're doing. Speaking of planning what you're doing, I know that you two always plan what to give as your recommendation. And here's a little secret that maybe we should tell the the AMG cyber people. If you have a recommendation segment at the end of the show and you're the host, always ask the other people first because you probably haven't thought of something. And then as they're answering, you can think of something. But what would you have to recommend this episode, Brandon? Matt Ray (57:02.053) Ha ha ha ha. Brandon (57:09.544) My recommendation, I have two. I have one, I did an interview over on the Enterprise AI show. So I interviewed the co-founder and CEO of a board where they do a lot of modernization, doing a lot of AI stuff. So good conversation. I think Rich has some good takes on like where AI is today and very pragmatic takes, which of course, you know, I welcome. I like that stuff. So go listen to it. Tell me if you like that interview. And then if you w looking for something to watch, I saw this movie a while back. It's it's called Tuner. It's about a piano tuner, which it sounds boring, but it was actually very good. So no aliens, no time travel, no science, no dark post apocalyptic world, just a good story about a piano tuner that is surprisingly good. So if you're looking for something, I'm gonna say mellow. I mean, I thought it was good to watch, but if you're not looking, you just like a drama. You're looking for like a drama, just something you maybe watch with your your significant other that's maybe not into the sci fi stuff, check out tuner. Matt Ray (57:44.537) No superheroes? Matt Ray (57:55.469) It's the drama, thriller. Coté (58:06.78) Do you do you think there's a category of movies when they're pitching them where they're like, you know, this is one of those movies that sounds like it would be a really boring movie, but it's not. It's like, just like, remember that movie about the guy in the subway who made sushi? That was a big deal. Remember that? Like there's there's a whole series of movies that's just like, and you know, remember that guy who used to sharpen pencils? Brandon (58:15.816) Mm-hmm. Matt Ray (58:24.854) Absolutely. Yeah. Brandon (58:27.206) they do pitch this because is it like coming of age or is it like family drama or 'cause it doesn't there's not like a like like if you just say the say what it is, it's n doesn't like have the hook. But maybe maybe Hollywood people like they're not looking for like special effects and you know I don't know what what I don't know. Matt Ray (58:42.765) No, undoubtedly Netflix yeah, Netflix is mining their their customer data and they they're like, Okay, Kote watched Jiro J Jiro Dreams of Sushi and he likes these these dramas that are they're kinda quiet, but you know what? We can make these things for two, three million dollars a pop. We keep Kote for another month. The spreadsheet show that this thing pays for itself and in ten months after release, let's order forty of them. And and then they're like backfill in the content. Brandon (59:06.13) Mm-hmm. It is y Matt Ray (59:11.929) Ha ha ha. Brandon (59:12.54) the you know, I want a full Netflix documentary about how Netflix cannot make good movies. Like that is what I want. Because I understand, I get it. They don't have to, but like I think a lot of times, like Tuner was a movie

that doesn't have any doesn't have a I would not expect it to have a huge budget, right? I mentioned Furious, the show on Hulu, phenomenal television show. Everyone should watch it. Again, it's just normal acting. It's not there's no huge budgets. And I'm like Matt Ray (59:19.151) They don't have to. They're they're filling space. That's my point. Brandon (59:41.605) If you're Netflix and you're making all of this stuff and you have money, like why not also make it good? Right? You know, it's like, cause you know when you're watching a Netflix movie, like I could put one on. Like that's it, that's the segment we should do. Be like, we're gonna watch a clip. And you tell if this is a Netflix movie or like something from Apple TV or Hulu, something where somebody actually cared about, like if it was good, right? And you're like, and it's just like cause the Netflix set, the set's always like a little odd, the green screen. Matt Ray (01:00:00.173) Oof. Brandon (01:00:10.652) Background is a little off. The acting's a little woody, but they're still paying good money. And I think they have plenty of people that want to make good things. So that's my pitch to Netflix. We're going to make a documentary about how we can't make good movies. And I believe it would be a number one, at least in the United States, it would be watched by millions of people. And like, why not? And just be like, hey, we're in the content game. It doesn't matter that we don't make good movies. Let's prove to people why we can't make good movies and they'll watch that too. Matt Ray (01:00:37.348) Ha ha ha. Brandon (01:00:38.544) It's like, you know, it's the whole snake eating its tail thing. I love that. Matt Ray (01:00:42.292) did did did you know, we're going off on a tangent, but did you read that article about the Roku's AI channel, like everything's made by AI? it was horrific. yeah. Yeah, well we'll we'll put that in the show notes in case you missed it. Brandon (01:00:51.164) my gosh. No, I haven't seen that. Is it live? we need to get that going. We need to start I need to put that on in the background during the day. See what's to see what happens. Matt Ray (01:00:59.874) Yeah, it it's it's it's the the dark end of that future. Brandon (01:01:04.946) Well, that's the other thing that they always say, like, Netflix always tells everyone, the the directors that they have to like, they assume everyone's watching the second screen and they have to like remind people what's going on on the show. You know, another another note is like you can make something that was fun to watch and not so boring that I'm compelled to watch my phone. Right. That's that's another way to solve that problem, right? Make something that's really good, right? 'Cause I'll because we'll watch it. If it's really good, I'll pay attention. Right. If it's has bad acting and a bad thing and we just Matt Ray (01:01:14.54) Yep, exactly. Coté (01:01:21.474) He's Brandon (01:01:32.776) I need something on the

background. Yeah, I yeah, I am gonna watch the TikToks of my phone. Might as well. Coté (01:01:36.262) This is this this is like this is like if if dad advice worked in Hollywood and ever every meeting would be like, Yeah, this good good stuff, but my recommendation is try harder. is Matt Ray (01:01:47.908) Ha ha. Brandon (01:01:49.98) convinced the people at Netflix know. I mean the people at Netflix when they go home and they put it on, the people that make the stuff, they're like, it's not that good. I I I think they know. Coté (01:01:57.138) Yeah, yeah. Matt Ray (01:01:57.294) Well th I I'm sure there are people who are like, you know, th three million dollars, that's your budget. And then some people come into that and like, Yes, finally, three million dollars. I can make I can make tuner. This little, you know, movie, it's gonna be perfect. That's what we needed. And then there are other people like three million dollars? I'm gonna cut some corners. But yeah, okay, you need seven of them this year, I can do that. Brandon (01:02:17.606) I don't know. Even listen, they this little Matt we're gonna end on a question here for you. We'll a little test question for you. Matt, what Ben Affleck and Matt what's it on what Damon or no, who's who's his partner? I'm sure Matt Damon, yeah. They made a met those two made a heist film on Netflix. Almost like you can't make a bad one with these too, right? They have numerous credits to of successful films. I want you to tell me the name of that Netflix film, because I'm sure you've seen it. What? Matt Ray (01:02:30.179) Matt Damon? Coté (01:02:30.47) Yeah, yeah, that's right. Matt Ray (01:02:38.965) Yeah. Matt Ray (01:02:46.178) I'm didn't did not see it. Brandon (01:02:47.322) Exactly, Matt. You're making my point. My point is like even when they get the right actors, they spend tons of money, they still mess it up. What? It's called the rip. And it's not that good. It's the rip is the name of the film. And it's right there on Netflix. You can go watch it. They I think they spent like forty to a hundred million. And it's it's and it's not and you're like you're like, these are the same people that made Academy Award winning movies, right? And they've been in tons of good stuff. And they go there and it's like somehow they're like, let's spend all this money and w and we still can't do it. So there you go. Matt Ray (01:02:53.292) So so why bother? Matt Ray (01:03:00.601) wha what was the one about the the the Matt Ray (01:03:08.845) Yeah, yeah. Coté (01:03:16.274) Hm. Well, you know s you know something that we can do is get a recommendation from Matt Ray. Brandon (01:03:17.04) Anyway, mm-hmm. Matt Ray (01:03:21.924) Well, so so I I I added to the show notes my anti pick about Roku's new slop channel. So, you know, I I the article was all I needed. Yeah, so that brand that's gonna be Brandon's like permanent listening, it's like you know, teletubbies for adults. and then Brandon (01:03:30.53) I'm checking

that out. That's that's the place I'm going after this show. Go on. Brandon (01:03:37.948) I'm watching it. Can't be worse than what's on Netflix. Matt Ray (01:03:46.092) And then obviously the the AMG Cyber Chat that was fun. Yeah, so that's that's gotta be one of my picks. enjoyed doing that. But my quality pick is a BBC series called Line of Duty. started 2012. it it's you know like all BBC series, they they they're on again, they're off again. I believe they've had six seasons and another one's coming. So I've watched the first two seasons and it's you know it's Apparently the show was universally praised. Rotten Tomatoes gives it a ninety six. So I enjoyed it quite a quite a bit. Coté (01:04:22.926) Mm, right. So first Roku slop and Matt Ray (01:04:24.687) police po police procedural, you know, it's it's an easy easy subject to do right. Coté (01:04:30.94) So first Brandon (01:04:31.058) Trying to figure out where do I watch it? I definitely'm in. Is it like on Brit Box? It looks it looks like a okay. okay. All right. Well, well, that's another that's another you what, Netflix? Why don't you just buy that shelf? That sounds like a good show. Just buy that. They don't have you know, they don't have to do anything. All right, I'll wait for it. That that looks like a good one. Matt Ray (01:04:33.924) well you get yourself a VPN and a BBC subscription. Brit box. Matt Ray (01:04:45.09) Yeah. I'm sure it's it probably comes in rotation periodically to to keep you on the hook. Coté (01:04:50.032) Yeah, it's it'll be on like Netflix Saturday night masterpiece theater showcase. Is is what what where where you could watch that one. Well, my my recommendation, I've I've I got this some time ago and I've had a chance to use it several times. So I'm gonna recommend the Patagonia Black Hole Mini MLC Pack 30 liter. Now they make a micro one, and there's of course the classic MLC, which is much bigger. But this is like right in the middle there, and I think it's Brandon (01:05:11.191) my gosh. Coté (01:05:19.726) It's of course you can have an over-the-shoulder thing. You can have convert it to be a backpack. It it works out pretty well. It's I went into the store, I don't know, four months ago and I was like, what is this? And they said this it's new. Can't we came out with this? And, you know, of course they set me up to play with it. The guy was folding shirts and he's like, let me put this up. You can use this table to to look at it. It's very nice. But it's it's good. It's got, I think it has As always, there could always be more pockets, especially more pockets higher in the bag. But the layout's pretty good. I think it works well. If you're if you're looking to be one of those like one bag people on a on a two to three day trip, I think it'll do it for you. And the backpack part is pretty good. So check that out. Bags. the micro. Matt Ray (01:06:01.624) Yeah. Yeah. I've I've got the twenty five liter. Matt Ray (01:06:11.02) No, no, the micro's the twenty. The I've got the twenty actually I have the micro too. Coté (01:06:17.4) There you go. But it's it's a nice bag and survives being wet. So that's good. Well, you know what else survives being wet? Another episode of Software Defined Talk. Always comes out dry, sparkly, and a little more cheery, I think, than when you started the show. You know how you can continue having fun with the show is you could go join our Slack channel. If you go to software defined talk.com, you can find the word Slack. I think up there, click on it. Come discuss links, other stuff. We got all sorts of channels in there. All all sorts of things. I think we even have a channel where you can share a picture of your dog, which I haven't looked at recently. I should go check on that. Also, if you want to get links to the things we mentioned, like those anywhere from zero to twenty-five free tickets to go see to go to We Are Developers, you can go to software define talk.com slash five eight seven and check all that out there. Like Brandon said, what you should do. Tomorrow is go sponsor some event at your children's school. And just say as part of the corporate sponsorship, all we want to do in my little five minute pitch, everyone get out their phone, open up Apple Podcast or whatever it is Andrew people Android people do, Android Playcast Pod. And search for so search for software defined talk and subscribe for six to twelve weeks. Brandon (01:07:21.009) Yeah. Brandon (01:07:38.617) Mm-hmm. Matt Ray (01:07:39.916) You're struggling. Brandon (01:07:46.694) Yeah. Coté (01:07:46.84) And that'll be a good sponsorship. And with that, we'll see everyone in less than six to twelve weeks. Bye. Matt Ray (01:07:54.318) Hi. Brandon (01:07:57.747) So my question, Kote, is this your daily driver backpack? Or is this your is this like your going away backpack? Or is that one and the same? Coté (01:08:04.21) you know, it hasn't really come up, but no, i if I if I if I I don't need a d I I'm not avoiding I'm trying to avoid your question, I guess it sounds like. I have another Patagonia bag, the one that you can fold up into into your hand, and I put a like a padded laptop thing if I just need to go about town with with a laptop, but Yeah, yeah, yeah. But yeah. Brandon (01:08:17.5) Mm-hmm. Brandon (01:08:23.248) Okay, so this is more like a travel backpack than a Matt, what do you what do you got, man? How you got the micro and you got the what w how many do you have? Matt Ray (01:08:31.379) so I've got the the twenty six liter black hole which Brandon (01:08:36.829) wow. They only have what they don't have they a twenty two liter pack, thirty liter pack. So you must have you're in like a different part. Matt Ray (01:08:40.227) let me find it. I've got the black hole or twenty-five. I've got the black hole twenty-five liter. which I bought when I had like a broken I dropped it in the the chat. I bought that like I ha my backpack broke a zipper or something on a trip and I shipped shipped it off for the lifetime warranty. and then I bought one of these and it's been it's been really good. Brandon (01:08:45.052) Twenty five. Matt Ray (01:09:10.337) But it yeah, it's all waterproof. And so it's it's big, so I use it for travel and then Google sent me a new Patagonia and I'm trying to figure out what model it is, but it's a small one, so it's gonna be like my go into the office one. Yeah. Brandon (01:09:25.286) Okay, so it's got on your daily driver and then you got nice, nice. Coté (01:09:25.296) yeah. Coté (01:09:28.88) Yeah, they have this little twenty two liter one. That that looks nice. But yeah, this is like the perfect size for like if you gotta travel somewhere. It's not too big, so it's not a goofist thing to walk around with it just while you're out during the day and it'll fit your clothes. Not too heavy. Brandon (01:09:43.1) Although I've done that a couple of times recently where I've just gone just the backpack, but then I'm like, I don't know. It's like not to carry the backpack. It was kinda nice to have the wheels. You know, it's always one of the things like I don't you know, it's like kind of lazy. So I learned that about myself. Matt Ray (01:09:43.107) Mm. Coté (01:09:50.828) It's true. It's true. I I go I go back and forth. J just like, you know, j just like I wanna use Apple notes and then I wanna use Obsidian and then I wanna use plain text files. Just it's it's just in the rotation. The only question is how many years between each? Matt Ray (01:09:54.861) Ha ha. Matt Ray (01:10:04.812) I I ni I need the twenty five liter 'cause I'm still hauling around two laptops. But the end is near and according to to the the QR code on my laptop, I mean on my my new Google Wiz backpack is it's a discontinued model. So Brandon (01:10:11.944) Mm-hmm. Brandon (01:10:26.312) There you have it. Well, on that note, thanks to everyone that wa watched us along in the live stream. Saw Michael, saw several people chatting away. It's always fun to have you along for the ride. And of course, thanks to everyone that watches us later on YouTube. We appreciate that as well. And with that, we will talk to you next time. Coté (01:10:45.542) Bye. Matt Ray (01:10:46.573) Bye. All right, I gotta run.