The GitHub problem (and other predictions) (Friends)

31:32And I'm cool with charging for products, by the way. Totally cool with it. But it seemed like this was a tax on those who want to do runners externally from GitHub. And it seemed like it was like, who wants to make an announcement a week from Christmas? Like, just don't do that. Even Docker made an announcement, which we're going to talk to them soon about like their thing as well. But like, don't announce things mid-December on. Like, wait to the new year. Be a stand-up company and just release when people are paying attention. Don't be sleuthy. Don't be sleuthy. Yeah. So there is that post, Jared, you sent me a post, and someone's complaining that GitHub has this monopoly.

32:21Right. Because what happens when GitHub goes down, you know, it does affect everything. It really does. The thing is, it kind of was the best choice, I think, for a long time. and it was just the easiest to use. So the user experience was just kind of there, and then obviously it got loads of integrations to it. So it sort of earned its place, but then it still does kind of have a monopoly. A lot of Go packages and things are on GitHub, and you wouldn't be able to pull them down, of course, if GitHub's down. We have seen a trickle of people starting to move other places, the most noteworthy of which I think is the Zig programming language, which moved to Codeberg recently.

33:11Codeberg? Codeberg. Like an iceberg? Yes. Code? Right. And it's going to bring down GitHub, the Titanic. Is that the metaphor? Is it? Yeah. Nice. That's good, isn't it? Yeah, and you only see that it's at the tip and everything else is under the surface, you know. Right. Right. Yeah. Which you should know about, Max, it's over there in the, well, they might be in the EU, which you're no longer in. It was on the way, wasn't it? They left the UK and they were going to the US. They were, they were coming here. A lot of controversy around that. Not the Titanic, Codeberg, the platform. It's a European Union platform.

33:50It is true. That's true, too. Yeah, so I guess it worked in both ways. That's not quite a double entendre. It's kind of like a... That's like a triple. Yeah. It's four if they mean the lettuce as well. Triple stamp, a double stamp. You know, there's a type of lettuce called iceberg lettuce. Do you have that in the US? Oh, of course. I feel bad for the listener. I'm sorry, listener. Let's go to iceberg lettuce. Go ahead. I'm sorry, too. No, I'm just saying it could be that. It could be that's where they incentivize their name. Like their name came from. They just love a good salad. Because they want to be associated with lettuce.

34:21They love the smell of salad. If you love the smell of salad, you're going to call your project something salad-related, maybe. That's what I'm thinking. I would. But the main thesis, I think, from... Let me double-check the name so I can be accurate here. Lionel Drycott. Now, that's my Americanized Texas slang, if that's even a thing. Right, that's another Dan Tan, is it? No, it's not Dan Tan. Dan Tan. You got to go Dan Tan. Who's Dan Tan? Come on now. Bring it back. You'll never know. We will never know. But Lionel wrote this post, and the hypothesis, or at least the thesis was this, was that GitHub's near total dominance over open source hosting has become a dangerous monoculture that makes alternatives invisible, not just less popular.

35:13Kind of an interesting phrase there, invisible. Codeberg. Okay, they didn't see it coming. Matt didn't even know it existed. That's true. Yeah. I didn't go on ships. I didn't go on many ships. Yeah. I haven't been on a ship for ages. And I think he's a teacher or the teacher, right? He's a teacher. He's teaching students. Students couldn't, while being told to do things in open source, they only use GitHub or like a large majority use GitHub while also being taught that there's not just GitHub. Like 99%, not even a large majority, like pretty much 100%. Pretty much 100%. So I agree with you, Matt.

35:51I think it earned it. I think it earned its monopoly. Mm-hmm. It dominated for many years. I think it stagnated as well. I think both those things are true in the hands of Microsoft. I think that they don't care about it. They care about pushing Copilot into every orifice of their corporate body and into our wallets. And it's gotten worse and worse and worse and worse. and it will continue to do so as people slowly move away from it until they all move away real quickly one day. Do the competitive products start out using GitHub? Do you think? How so? When you start, what are you going to do? Make a GitHub repo.

36:35Go from there. You've got to start somewhere, haven't you? I mean, even Go was written in C until they could self-host. Yeah. Is that true? I think it's true. That's true. Yeah, that's true. Go was all written in C until Go was good enough that it could then be written in Go. Exactly. Which is amazing. It is cool. I've always loved self-hosted languages. But yeah, I'm sure they have to. But the nice thing about DVCS, for those of us who like acronyms, are distributed. Is it decentralized or distributed? Version control system. I think it's decentralized. It's both. I think it stands for distributed, though.

37:08Is that you could just have multiple origins, or not origins, multiple remotes. you know got your locals like there is no necessary yes your main remote is your origin is github right now but it's so easy to just change that to something else that switching off if you're just using git and those flows it's actually super straightforward however if you're using actions and sponsors and prs whatever whatever else projects if anybody uses that still then it becomes much harder yeah it's the gravity that's why they call it github universe it's a there's a gravity to it honestly i mean there really is and so what do you do do you fight the system you know do you if you're launching something just complain and it's in open source sure you can have multiple remotes of course that is totally possible and you could it's your prerogative but if the users aren't there what's the point well it may it may take a tooling change to actually be significant.

38:13And I think Git is entrenched at this point because of agentic coding, especially it's entrenched. However, the further you get away from that tool, the easier it is to have your agents just go use something else you don't care anyways. And so maybe that's not always the case. However, the reason why GitHub became what it was, was because it put all of these collaborative features around a new tool-ish that was already getting popular and so they kind of popularized each other and so maybe as jj is now jujitsu the cool new tool of the bleeding edge folks i haven't tried it yet i don't like to bleed as much as others but people are loving it maybe a platform not a centered around git a centered around jj which could also support all the git things which is currently the way jj roles right like it's like a superset i think of gits abilities uh has a chance to de-seat github once and for all what do you think matt yeah i think maybe that maybe the whole thing the whole paradigm will change yeah we just don't care maybe you don't care yeah or it's just kind of different maybe it's just a list of prompts all the way down but i don't know it's unlikely isn't it there's some fundamentals we're probably going to stick with we have this thing because we're building, of course, the Grafana Assistant project, which is an AI tool.

39:37And, you know, it's built into the, so it basically can write all the queries for you. You just ask it telemetry questions in natural language. Love it. It's so good. The team that built it. Because I've used Grafana and I love the outputs, but I hate trying to query the thing and like, is it Loki? Is it something else? I don't even know. I always have to go to Gerhard and say, how do I write this query, Gerhard? Write it for me. I just want to tell the thing, hey, show me like the 99th percentile of requests over the last hour. Not only that, though, investigate why this is spiking. Because it's agentic.

40:09It can go off and sleuths around. It's like the changelog detectives, Adam Logg and Gerard Change. Change. I'm not French. Change. I will crack this case in no problem at all. I love it. That's your Gerard. Gerard Change. Keep going. Don't stop, Matt. Go further. You're offending me. I need more. I'm not offending. It's not offensive to be saying that you're French. I'm Italian, though, so it is. Oh, okay. I'm Italian. So please change it. Italian-American. Yes, please. Go ahead. We call that American. Give me the Italian version. Yeah, but the problem with Italian is it's very easy to do this stereotype.

40:57He did it. It is. We all end up sounding like a Mario. I'm hiding from the camera. My face is too red right now. I can't take it. No, but it's not. I do it with love. I love different accents. I've said this before. I say every time to keep out of trouble. I love accents. You have to say. You should get that tattooed on your back. I have. I have got it tattooed on my back. Wow. How did I know that? Well, yeah, you know. You know. No. It's a family show. I've never even met Jared in real life. Or you, Adam. Well, we tried one time, but it didn't work out. We've never met in real life. No. It's only on this, through this telly.

41:33Do you refuse to come to the US? I can't remember. Oh, one time I didn't want to go. You refused. Well, I didn't want to go, yeah. I didn't want to go as a refusal, right? I suppose so. No. He said no. I feel like I want to write the headlines, though. Not that. But, yeah, this year. Maybe this is the year. Maybe 2026 is the year. didn't you all uh we just i think we just shipped an episode of big tent and you guys were talking about this weren't you like in the most recent episode of big tent yeah yeah the big tent is grafana's podcast all about the people community tools and tech around observability yeah i was reading the transcript and i was like and i was watching it too because that chapter and stuff like that but it was it was a good part i like that what were you talking about i'm lost you're talking about coming to the states no the ai system they're building we're back on top of Okay, thank you.

42:24Both. Well, both, because we're a fully remote company. And didn't you just have a GrafanaCon or something? We do also have these conferences. There was GrafanaCon in London. If you can get to a GrafanaCon, do come, because they're so much fun. And I host them, and it's very fun. We all have a good time. People usually end up trolling me on the Slido. That sounds nice. Which is fine. No, it's fine. They're funny. So I read them out. That's where the material comes from for the show. I really appreciated the clip. I believe it was Ivana.

42:58She said she tried to cover the fire with wood. And Tom's response was just typical. Tom, I love it. It was awesome. He's like, did you learn anything? She's like, of course. She had a home lab where she was measuring fire and measuring, I don't know what I'm liking a lot here for some reason. I think it was a candle so that she could put the candle out. But the thing she was putting the candle out with was made of wood. Yes. Don't use wood to put out a fire. Well, this is the culture we have at Grafana Labs. It's okay to make mistakes. Not only is it okay, it's kind of expected. Because if you're not making mistakes, you're probably being too careful.

43:38So we have a journey. This is a real culture thing that we talk about at Grafana Labs where we have an error budget for our services. And if you're like at 100 % all the time, it means you're probably not innovating enough. You're not doing enough. It's a sign of something maybe. Or they're just amazing engineers, which could be because there are some phenomenal engineers. But Ivana Hutchkova, she works on the assistant itself on the front end of the assistant. So yeah, she's part of the team that built that. And we've heard from leading AI companies that I won't name. They said this is the best implementation or the best use of their models that they've seen so far.

44:17Which they love because they're waiting for people to use them well. you know yeah yeah they are they're very excited they're very excited and honestly like this is this is a good use case for llms for sure totally um but if github goes down it doesn't work we can't we can't we can't deploy stuff i don't know how we deploy stuff without can you be more specific am i did i miss the what exactly you've done can you give us a 30 second version of it that's not marketing speak or on the assistant yeah the assistant like what exactly i know you You can query Grafana. Grafana is a visual tool primarily.

44:51There's not a lot of querying, to my knowledge. You write the queries to build the dashboards. Right. Once you've built your dashboards, but that's okay if you've already thought of something. But it's like something new is happening. And also if you just don't know where, some people have loads of dashboards. This is basically an LLM integration, so it's a chat experience. And then it has all the tools that you have in Grafana. And it knows how to navigate around and things. so you can ask it questions and it can not just take you to the right app to look at it but applies all the filters in the url parameters and deep links you right to that view so it's awesome it's yeah it changes and it makes it makes it so that anyone can now use that get value from the telemetry data whereas before you would usually have to go to like the resident sre or the little team that know it all and ask them now you can self-serve a lot and free up that And they're know-it-alls, so that's the worst, you know?

45:44Yeah, because they have to, they do, it's hard. I never learned the Loki's log QL language. I just kind of go to the docs and make it up, and then edit and things. Same, same. But whereas now you just don't need to at all. That's beautiful. Yeah, and then because it can query data, you can ask it, like, why is the shopping cart slow? And it'll look across all your signals as well. So it'll look in logs, it'll look in metrics. Even if you've got traces and profiles, a lot of engineers don't know how to really get value or use profiling or tracing properly. So you don't need the assistant. We're taught the assistant how to do it all.

46:24Grafana doesn't work if GitHub's down? Well, we can't deploy changes if GitHub's down, I think. I think we probably can manually, but all the pipeline and everything flows through GitHub. Have you ever considered making that some sort of redundant thing so that you're not so dependent? It probably is, to be honest. I'm just saying like the normal flow, yeah, you merge to main and then that gets, we deploy to a lot of instances immediately. What we need is an actions mirror. That's, you still, I guess actually that would suck. Like you still deploy to GitHub. I was thinking like a webhook once you've deployed to main or whichever your branch is, and then something's watching your repository, but that assumes again that GitHub is up.

47:04So your deploy, your change can't get there to get webhooked out to an external actions runner even, or a watcher that mimics what actions does. Yeah. Yeah, that's the need for Codeberg or this decentralized, this double D VCS thing. This monopoly, this centralization. Come on now. Stop centralizing. We just need some competition. That's all we need, some legit competition. All right, JJ. Bam. Time to get it. time to get it what you know i don't know a lot about jj but what i've understand do you all know much about jj all i really know is that they're like similar to get and they can like run alongside of it until it's time to take over that's like my synopsis of how jj operates today yeah i don't know i'm i'm excited to learn more about jj uh i don't know i don't know who they are what about you jared you know anything you got you got a different opinion or a deeper opinion no i do know that the command line tool can do all the things git can do but then it also has its own superset of it and so it's easy to try out a lot like typescript was easy for javascript developers to get tricked into uh jj you know can trick you right into it and it's nice that way it's very nice yeah anywho but i've never used it because i don't have any complaints about git i've just used it for multiple decades and so i can't i've been through the pain i know i i have complaints about it from the beginner's standpoint but i also now when i get stuck i'm just like yo claude can you just rebase this for me and then he does it so here's an idea let me let me wordsmith or think through this guys uh with you guys here in the What if we had a pre-GitHub load balancer where you committed to a thing that was meant to be a load balancer?

49:01Your primary host is GitHub, but you still have downtime, of course, but you're not on GitHub's downtime. And it's not monopolized by GitHub. So what if you had a load balancer that you pushed to, but ultimately the transaction landed at GitHub to run your actions and your runners and all those things? but what if it could also webhook somewhere else and say hey runners over here can still do it where that way if github is down you have sort of one layer before that kind of like the low bounce effect to your git pushes i don't see why you couldn't do that get like a proxy that pushes to multiple places it's all in sync it'd be a fight against it at least it'd give us a chance if downtime is your major concern yeah well my major concern is in downtime it's it's crappy time.

49:47I don't know what you call it. It's just like their bloated JavaScript UI, the fact that like copilots showed to my face everywhere, stuff like that is what bothers me. Like actually using the website is slow now. It used to be super fast. It's like way too much React code, I think. Those are the things that bother me more so than GitHub being down. But obviously that is a problem and we can't deploy when GitHub's down either because we don't have this proxy that you're talking about. But I'm certain that it's a good idea and that they probably exist out there. Like a green blue almost in a way.

50:20Let's say a green blue to your Git repositories. That way, if ever you did need to eject from the GitHub world, you could. And it's like, it's almost what they're doing with passwords and quantum mechanics and stuff like that. Like the quantum computers, they're making defense towards, you know, encryption. And there's a lot of effort. I just saw this talk at GopherCon, actually. thank you go for publishing your your talks most recently I can't recall the person who did it I'll put in the show notes but it was a talk on defending against I believe it was FISO 140 or FIPS 140 I think is what it was and it was essentially like defending against the future fact that the encryption we have will eventually be de-encrypted by quantum computers it's almost like that's like a defense against the future this this green blue effect we can apply to git hosting in general, not just GitHub.

51:18I think you could pivot and do that. Just keep it called the changelog. It still works. Yeah. Bam. Good idea. What do you do when GitHub goes down then? What's your favorite go-to? Because you can't work. Walk. Go for a walk. Take a walk. Git E. Git T. Git T. Go and get some T. Take a shower. Well, I mean. Use it as a prompt to shower finally. I'm too careful. Time to shower. Have a shower. Well, you only shower when GitHub goes down. I use it as a prompt, you know. I mean, I shower like I change my underwear, you know, monthly. Okay. Phew. I thought you were going to say something gross. TMI.

51:57I don't know. What do you do? Do you sing a song? I always go reach for the guitar if GitHub goes down. How do you know? Yeah, it's my go-to, really. My favorite keyword in Go. I use all the time. Is that a real legit Go keyword? Because, you know, GoTo is considered harmful. It is a little harmful.

52:39I forgot what a tree was. I thought it was a thing with a light on top. I forgot what tea was, but that's because I'm tea. So tell now, I'll go outside. When GitHub goes down, I'll smell the flowers. When GitHub goes down, I'll look at the birds and tread in turds and meet with nerds. When GitHub goes down, we'll use the DTVCS. If GitHub goes down, but if Google Maps goes down, then I won't go outside. Because I'm not entirely confident I'll be able to find my way back. I use Google Maps way too much. I'll be becoming too reliant on technology.

53:31When GitHub goes down. Woo. that's a double entendre too I mean it's not just saying down time it's like when they go down like as if it's a demise you know predicting a potential demise here that's the end of a titan so many entendres in this pod triple double stamped yeah that entendre yo

53:56well friends this episode is brought to you by Squarespace the all in one platform for building your online presence whether it's a portfolio a consulting business or finally shipping that side project landing page you've just been meaning to do, but never get to. Here's the thing. You mass produce code on the daily. You deploy new services, new infrastructure, new hardware. You're versioning your APIs. You're servering all over the place. But when someone asks you about your own personal website, it's like, ah, I'm still working on it. Does that sound familiar? Squarespace exists so you don't have to treat your personal site like a weekend project that never ships.

54:32pick a template and drag and drop your way to something that actually looks good and move on with your life. No wrestling with CSS. No, I'll just build my own static site generator again. It's just done. If you do consulting or freelance work on the side, Squarespace handles the whole entire workflow. Showcase your services, let clients book time directly on your calendar, send professional invoices and get paid online. It's the boring infrastructure that you don't want to build for yourself. And for those of you out there who are doing courses or gated content or educational stuff, tutorials, workshops, that intro to whatever series you keep talking about.

55:08You can set up a membership area with a paywall and start earning recurring revenue. Set your price, get the content, and you're done. And they've also added Blueprint AI. This generates a custom site based on your industry, your goals, your style preferences. It's not going to replace your design skills by any means, but it'll get you about 80 % of the way there in about five minutes. And here's the call to action. This is what I want you to do. Go to squarespace.com slash changelog for a free trial. And when you're ready to launch, use our offer code changelog and save 10 % off your first purchase of a website or a domain.

55:45Again, squarespace.com slash changelog. Who knows what's going to happen in 2026? Like it's unwritten as they back to the future is my favorite film. genuinely this hat is actually a back to the future hat from back to the future 2 is it yeah it's lovely i'm very jealous of that guess who got me this french friend your wife french friend oh your french friend yeah so he is a friend absolutely yeah if he's bringing you a hat it's an amazing gift isn't it i'll show him this because he definitely doesn't subscribe well i know somebody who thinks they know what's going to happen in 2026 who knows tomas tungus i'm sure that's how it's pronounced i am not sure that's how it's pronounced but if you go to tom tungus.com slash 2026 dash predictions you'll find that tom or tomas as i like to call him has written 12 predictions for 2026 and he wrote these you guys won't believe it in 2025 that makes sense why would it be weird if he wrote it in 2027 it'd be too easy it would be way too easy if he wrote it 2027 in fact i should write my 2025 predictions tomorrow you should actually publish it no one's most people aren't sleuths like you detect your blog and change change gerard change i'll look at the dead time stamp not everyone does that's a great sound to hear slandering me as French that's the offensive thing that's the offensive thing it really is what's this post about Jared this post is Tom who is a venture capitalist so he's a venture capitalist at Theory Tomas Ventures yeah Tomas but the website is Tom so I'm going to call him Tom because I don't know if Tomas is correct it could be Tomas it could be tom ozzi do host yeah or do host i do like i like these names that they've got like a zed in and he's got two z's at the end of german friends out there just saying like what tom yeah i saw i saw a german keyboard the other day and some of the letters are mixed up it's like the bloody enigma machine i had to crack it crack the code before i could use it i made that joke in front of Germans and it was just deadpan.

58:10It did not go down well. Nine. Nine. So Tom has written 12 predictions for 2026. We thought it'd be fun since it is the new year and Matt has all these resolutions. I'm still rocking the same resolution I had last year, which is 1512 by 982. Yes. It's classic. Classic joke. You could have done that low res eyes before you had your eyes fixed when you first saw a leaf for the first time. and you were like, oh my gosh, there's more pixels here. That's like when we got the retina displays for the first time. We were like, wow. Yeah, it's true. That's what life looks like. The text was really sharp.

58:48Yeah, it really did. And we thought it'd be fun to go through some of these. We don't have to hit all 12, but a few that we think are either right or wrong or we have comments. If you have comments, sound off as I read them. How should we do this? Should I go, I'll read all 12 and then we'll stop, I'll pause if you guys have anything to say. Matt say nothing read all 12 Jared if not okay you see how I preempted that here Matt say nothing well he just said to talk that's why I was gonna talk well no more talking for you for just a just a minute here go ahead see them all you prefer when you sing your songs are great yes but commentary you know for the listeners sake like if we could have a swap if we could have a mid-show swap you know like Matt if you had a tag team yeah maybe Tom Wilkie could be your tag team and like Tom could talk and then you could sing.

59:38That's a good idea. Let's swap out Tom next time. No, the other person could be worse. Just right in the middle. No announcements. Back to the devil you know. All right, so speaking of the devil, what we know and what we think we know, Tom thinks that in 2026, businesses will pay more for AI agents than people for the first time. Pausing for comments. Oh, I thought you were going to read all 12. Oh, I thought I was going to read each one and see if you guys comment. I'll read all 12. Here I go. Yeah, go fast. Number two, 2026 becomes a record year for liquidity. That's very exciting. Number three, vector databases resurge as essential infrastructure in the AI stack.

1:00:18No. Not to be confused with the Atom stack. Number four, AI models execute tasks autonomously for longer than a workday. Dang. They'll be working hard for a living. Number five, AI budgets receive scrutiny for the first time. Number six, Google distances itself from competitors via breadth. Breadth in AI. I'm not sure what that means. Number seven, agent observability becomes the most competitive layer of the inference stack. Okay. Number eight, 30 % of international payments are issued via stablecoin by December. I don't know why you had to put December in there. We know the year. Okay. Number nine, agent data access patterns stress and break existing databases.

1:01:05I can definitely see that one coming through. Number 10, the data center build out reaches 3.5 % of US GDP in 2026. And number 11, the web flips to agent first design. Say it ain't so. Say it ain't so. Number 12. And the final prediction of Tomas Tungus is Cloudflare becomes the gatekeeper for agentic payments. There you have 12 predictions for 2026. Which ones should we talk about? Let me just say, whatever his portfolio is, I want to invest. Okay. Because these are all great predictions. Okay. You like these? Fantastic. Yeah. I'm not in my head to like, I'd say 10 out of them, 10 out of 12 for sure.

1:01:47And I'm not iffy on them. Am I just like no less? Yeah, I can't see one that I like hard disagreed with. I don't know about GDP. You know, obviously that's like a specific number. I'm not sure about the stablecoin one, but I can see it happening. Yeah, I don't like hard disagree. I don't like number 11. The web flips to agent first design. In fact, I'm going to say that's not true. That's the one I'm going to disagree with. Break that down for me and for the listeners. What does that mean? To you at least. Well, when you create a website today, what do you think about? What do you design it for?

1:02:20Well, you design it probably for a mobile device first and then also for a desktop, right? And then maybe you think we also need an API because we need to have programmatic access to our website. But you're human first for sure. And you're probably mobile first. Yeah. Maybe you do it all at once. But he's saying that in 2026, the web will flip to agent first design. so the first thing you're going to think of the number one thing you think of when you start a new website at first is how will an agent use this and then comes everything else i hope that's not true i can see it being like sometimes the case for certain websites for certain uses but like for all the web he calls it the web i'm thinking that's an awful lot of change quickly for the web, which generally moves somewhat slow.

1:03:10Some pushback on that. I, I do concur with this thought, but the lens I would shift just slightly because I feel like every, it's almost like behind every good man is a good woman, that, that whole shtick or phrase or whatever. It's like behind every great human who's doing great work is, or maybe in front of actually, since we're front loading this is in front of every great human doing great work is an agent. So I almost feel like you're helping the human do better by being agent first. So I'm sort of, I'm conflicted there because that's how I'm thinking too, really. I think that if this last year has told us anything, people like agents, developers like agents.

1:03:55I think people like agents and people are going to start using cloud code who don't even code software. They're going to be coding pros or whatever. Um, that's where my lens is at. I'm thinking like you've, if you're not taking the agent in mind, then you're in the past. Like you have to think about, I'm not sure I would say quite agent first, but definitely agent as well. And almost everything, everything, even like simple CLIs are human first, right? You have to, like, if you have an error from a CLI, like you threw a command out there that you misused a flag, have a proper error, not for just the human, but for the agent that gives them context.

1:04:34Hey, you meant to do this, or this is what that does, or here's documentation and you give them context. I think that's where my lens is at too. So I'm not sure if it's for the web and, you know, agent first necessarily, but you definitely have to be thinking about every new interface with agents for sure. Hands down. I agree with that. I don't disagree with agent as well, which is what you said. I think agent first in 2026 is too fast. And also I think it's, it's both too fast to call. And I think it's too fast to do. I think agent as well makes a little sense. Matt, what do you think? I kind of get this for sure.

1:05:11Cause we did, we've been through transitions like this. Mobile, you mentioned that was one of them. It used to be web before you'd be building for desktop. And it is about how the users are interacting with it. That's always been the most important thing about technology. So, yeah, I could imagine like if it's like booking something or if it's like a hotel or any of that stuff, I feel like total that's all gone. Basically, it's going to be presentation of something like here's what I found for you. Here are some images and things. But you want to ultimately you'll just say to an agent like, oh, we're thinking of me and a few friends are thinking of going here.

1:05:45Can you ask them for their availability? One of the Fridays coming up would do for me or check my calendar. it goes off and pings their agents to find out their availability and if they're interested in it you know and then that all happens they just get some kind of notification and the agent just asks them a question like a text or it makes the text look like it's come from you even more worryingly probably um and i think that those kinds of flows i think will happen and they'll they'll have to happen but yeah would does this mean you're then not going to have an experience where you are choosing what to present agents are all at the moment you are prompting them you have to ask a question for something what if you don't know what to ask it what if you don't know what it can do those kinds of things i think is where we'll have some kind of other experience we had the same thing with dashboards because you know someone said does does the assistant project mean you don't need a dashboard now because you can just ask it the question but there is something about um having being able to go and look at something without having to ask for it and just having an easy way to go and find that and i think dashboards still sit there like they're still i don't think dashboards are going to go away for example but it'll be new things alongside it yeah it's the hub and spoke model you always have a you know an api first and you always have a client it's like if you live in that api first design world then you totally get what's happening here because if you've always been API first and a CLI is a client, the website's a client, and an iOS app is a client, that's pretty easy to sort of grok that direction.

1:07:22So I get that. This kind of conflates to some degree with Paul Dix. Jared, he mentioned, trying to figure out the best way to say this, but while I was reading the great engineering divergence, one thing he mentioned was Amdahl's law, which is the principle is when you speed up part of a system, your overall speed up is limited by the parts you didn't speed up. So it almost reminds me of that in the fact that if you don't think about an agent and you don't think about that first mentality like even Matt's describing here, you don't enable yourself or the things you're building to be as fast as it can be because agents are going to be much more faster than we are because they're designed to be.

1:08:05They're a machine. I think if you don't start curbing that idea, you're going to be dwarfed by the folks who do and embrace these new systems that have to move faster. And they retool their entire pipeline towards agentic things that move faster. And if you stick in the old way, let's just say, you're going to be slower by nature and less fast. So slow to market, slow to think, slow to experiment, slow to fail, slow to succeed, all the things. I think it's, I got to say it, I don't like saying it necessarily, but I am thinking, I guess, agent first. Yeah, I guess I'm thinking agent first, honestly.

1:08:49I mean, could you imagine a service that is only an MCP server that you plug in? It doesn't have any other kind of web interface. The only interface is through the agentic kind of chat. I don't think anybody wants to be that. I don't think any business wants to be that. I mean, nobody wants to be a utility company. And that's why they fight to be not just a utility company. They want to provide services and apps and like all that. They want to be your provider, not your dumb pipe. and I think that most profitable and desiring to be profitable companies will will buck against that as they begin as they begin to get commoditized by agents so I think there'll be a fight there but yeah as an end user of course we want the simplest easiest cheapest thing right that works and that's where I'm at with 2026 I don't think agents work yet I think we are living in the golden age of coding and we think that everything's like a coding agent agents are not booking anybody anybody's flights, anybody's hotels, anybody, anything.

1:09:50It's January, 2026. They don't work yet. There's way too many edge cases, too many problems, too many errors. They are writing code and they are summarizing text and maybe they're, and they're helping out with your reflux stuff, but that's code. And that's kind of where it's at right now. Like they're all just demos. And so I don't think it's going to move as fast as I should design my website agent first today. I do think eventually you do. I just think that the timing is wrong. Yeah. Yeah, I think the thing is the speed of changes is increasing as well as more and more people are AI accelerated with their work.

1:10:29So I do think things are going to change much more quickly than they ever have before. And things have changed quickly before. But yeah, look at like Cursor, OpenCode, Zed, ClaudeCode, but all these AI now enabled things, a lot of their UIs are shifted so that their agent first experiences now, where the code is kind of an afterthought that you review. Right. Once you've got it kind of going, it is quite interesting. I do think there are questions about what that's going to mean, but there's no doubt in my mind that it's an enormous acceleration of human productivity. And that's why I'm on the side of like, yes let's use ai and let's use it for good um and we need to solve all those challenges around energy and the climate and um some of the bad things it can do and some of the the way that it's like quite greedy with with taking knowledge from people without kind of giving them credit and things like this there's a lot of problems with it but once you've seen like if we felt it like building building these products we're building and and also because they themselves are the same kind of thing grafana assistant is basically like cursor for grafana the the productivity boost just the amount of people that that's enabled is is significant so it's not it's not just a fad thing you know yeah i also hate to say that i agree with you directionally as well and i think adam does as well my only pushback is on the fact that we are in the perfect we're like the goldilocks case for agents, which is software and digital creation.

1:12:07And the rest of the world doesn't do that. Obviously, it touches everything, but they do all kinds of other stuff that agents have no ability to do yet. And so that will take time. I think it's coming. I think it's coming. It's true. We just look at everything through this cursor and Claude Code rose-colored glasses because it's amazing for us, but it's not amazing for most other industries yet. Maybe law is the other one where it's really making moves because again, that's so formal. And then everything else is just sloppy at this point. Yeah. We are at the forefront usually of change. Technology has been the driver of it.

1:12:43So it makes sense that we would see it first. Yeah. We didn't have to build it for the rest of them. That's our jobs. And the nice thing is, and the accelerating thing, like you said, is the fact that we are the builders. We're moving faster than we were before. And so you can build faster. Okay. Another one. What else caught your guys' eyes or ears? yeah the uh businesses pay more for ai than human labor that's one of his predictions isn't it does that mean that we become the cheap choice again we can get our jobs back i don't think that's how it works yeah i think that will be the case in the case again of software workers i think i've seen a lot of the budgets exploding and i think that at a certain point one engineer plus you know an engineer's equivalent of agentic coding is probably better than two engineers i don't know exactly when that flips again is it this year or next year but i could see it happening adam i was just reading it to get uh more of the context of tom's tomaz's perspective here i'm gonna read it just because it's it gives me some context i haven't formulated a thought yet so this is kind of a delay uh this has already happened with consumers waymo rides cost 30 percent 31 percent more than uber on average yet demand keeps growing riders prefer the safety and reliability of autonomous vehicles for rope business tasks agents will command a similar premium as companies factor in onboarding recruiting training and management costs also that's not time on budgets it's talking about uh per capita yeah it's just like you're gonna pay maybe that you'd pay if you had a choice between the human doing the work and the AI doing the work, you'd pay more for the agent doing it because of the perceived safety and reliability.

1:14:30If that Waymo option translates there. I'm against, I'm personally, I chose the other way when I was in Phoenix. Okay. Because I wanted to try a Waymo because of course, but I'm cheap. And so I looked at Uber and I looked at Waymo and it was like an extra$12 to go to the same place. And I'm like, I'm just taking the Uber, but I can see where There are circumstances. Yeah. For instance, if I'm sending my children with, I would maybe, I would trust a Waymo where I wouldn't trust a random Uber driver with children, for instance. I can see where parents might prefer it. I can see where over time it becomes more safe, especially if you're in an area you're not, you know, used to.

1:15:08Not just driving safe, but like this person could drive off in the place and whatever me. I can see that. So while I did make the choice the opposite one time, I could certainly see where I would make the choice to pay more for the robot because I don't trust the humans. Yeah. See, that's just it right there. Yeah. Yeah. Now, see, if you read headline only, you're thinking salary swap, right? I thought budgets, yeah. Yeah. Yeah, that's where I was torn. I was like, what is the true context here? Again, context. Yeah, I'm torn on that one. I think, you know, I think I would, in the case of a Waymo, let's just use this as an example.

1:15:46I'm in the vehicle, right? My life is literally on the line. Do I pay 12 bucks more for the assumed? And I suppose if data shows the reliability and safety is higher over a trend of time, then I would pay more every single time because my life is literally priceless. And if I can't be here, then I can't even care about spending more money. So in that case, I'd probably spend more every single time on a Waymo if the data and it wasn't smeared or tainted in any sort of way. If the data was true and over time Uber was less safe while Waymo was more safe, every single time I'd pay, I'd probably pay double, honestly, if I had to, if I knew that I was going to be more safe and point A to point B would truly happen.

1:16:32And I have determinism in my, my trip every time. Yeah, for sure. I'd pay double if it flew. If it flew. Yeah. I don't want to pay double Waymo, but I'm just saying like, if, if knowing the date and the choice, then I'm going to pay more for the thing that gives me more safety and security in a time where my life is literally on the line, you know, in the case of a real business task, maybe not so much. You take it to like surgery, you know, over time, robot surgeons will operate more precisely and correctly than humans will because they don't have the margin of error. They didn't have a bad night.

1:17:06They aren't tired, et cetera, et cetera. we've all seen prometheus not the prometheus that you all uh wrangle over there at grafana but the prometheus the movie what was her name shaw i believe her name last name or her name was something shaw 512 the last name on it maybe so she had to hop in this thing at the end when she was giving birth to this alien and then had to patch her up she happily got in there and was just like pushing all the buttons every single button go back to the it's all about the button episode from the beginning of 2025. That was an amazing show. Yeah, I got stuck in that moment there for a second, but she happily pushed every single button possible to have the machine help her deliver this alien baby and patch herself back up.

1:17:55And then she went and conquered the mission, right? Like if that's a science fiction is kind of predictive in a way, if that's a version of our truth and our future, I mean, she's kind of already trusted the system, right? I mean, that's something I personally have said for almost 15 years now, trust the system, but verify, right? But verify. I think in the future, if the data shows a AI-assisted or an AI agent surgeon is better, I mean, I don't want to choose a machine over a human, but in those cases, if the data shows it, then it just makes sense. It just makes sense. It does, and I think that is where we're going.

1:18:33We are increasingly going to just be doing that more and more. for sure so we do have to figure out how we deal with that change because that is an enormous change but i trusted the waymo when i got in in san francisco i trusted it immediately and that's probably because i've been brought up on sci-fi films johnny cab from total recall where you get in the guy spins around and yeah he's you know he's a little robot boy takes you on a little journey um and the key thing about waymo is you've got to you don't have to talk to anybody you can just not talk to anyone it's awesome and that's worth 12 at least true actually uber actually do give you an option in the uk i don't know if you have this but there's a uber comfort option and this basically allows you to choose the music and decide whether they talk to you or not so you do pay extra for them to not talk to you yeah i saw that option in the uber app i didn't know that it costs extra maybe i just paid extra and didn't realize it i definitely said like i I think it's nice because sometimes you feel like talking to somebody and you're like, yeah, I'll have a conversation with a stranger.

1:19:37And other times you're like, no, I just got off an airplane. I want to sit in quiet and get to my place. And so it's cool that you could just get a pick like, yeah, don't talk to me. Well, the other factors you're not thinking about and maybe it's just not mentioned is like the smell. Right. Humans have habits and they also have odors. And, you know, you can be a smoker. You could be a not smoker. You could prefer a certain sense in your car. Right. S-C-E-N-T-S sense. uh not just see it you know you know how to spell sentence the other sentence forget it y 'all you're smart people out there listening to this podcast i'm done trying to spill on a podcast but yeah i mean how often do you get into that or you get the music right you've got all these human nature things that you're like you know what i just kind of rather avoid a human in this moment yeah that's a scary there's comfort in just saying that and i'm smiling very big if you're listening on the on the actual audio pod.

1:20:35I'm kind of blushing in a way. Cause I can't believe I'm even thinking like that. Like I'd rather have a ride without a human. If I had the choice just because, but there's sometimes too, Matt, like you said, I'm like, I'd love to, you know, I'm down with the humans. Right. But I'm also down with the non-humans cause humans smell and have just things. Yeah. All the things, you know, opinions, habits, smells, music choices. Yeah. But see, the other thing about the self-driving taxis is you can have the night rider experience. I call it, it's my new startup, night rider. You go to bed, you get, you know, it's like a little hotel room that's on wheels.

1:21:14You get in, you get in, yeah, you sleep, you wake up in a different city. You then spend the day in that city, you sleep, the next day you're in another city. You're traveling while you sleep. It's the closest teleportation we'll get, probably, because the EU keep ignoring all my letters. But I know how to do it. This reminds me of Matt World. Isn't this how Matt World works? Yeah, I think that was Matt World. One of your inventions was basically the Knight Rider car. But they're building it. It would work. For sure. It'd feel bad to get a human to just drive for 12 hours to take you to Edinburgh.

1:21:47Yeah, you have to have empathy, right? By nature. That's why I like Claude Code better than Junior Dev, honestly, because it's like I just don't have to even give you any empathy or anything. Like I don't have to afford, there's no affordances. Well, there's some true psychology in that too. It's not just personal preference. So don't feel bad for yourself. So if you're listening to this, here's your escape hatch, right? It's mirror neurons. So you see this a lot in married couples. As they age, they tend to dress similar, look similar. They don't literally look similar down to the wire, you know, so to speak.

1:22:25But there's mannerisms that sort of merge and it's mirror neurons. Or when you're around somebody, the reason you have that empathy factor or you begin to cry because somebody else is crying is because your brain is literally wired to mirror neurons. It's called mirror neurons. It's a psychological fact. It is brilliant. People do end up looking similar because they'll pull the same facial expressions. Yeah. I got this frown I keep doing. I hate it. Gosh. Yeah. And then so you're working. I'm always doing this face because my wife's always doing it. She looks better than that. Yeah. But then my face changes.

1:23:00Does she have the goatee as well? She does, yeah. That's a mirror neuron for you? It's a woman. That's when your neurons are mirroring too much. Yeah, too far. You need to get that checked. It's too far, yeah. Well, if you're around somebody, even, and think about this too. The next time you're around somebody and you're just standing there having a conversation, if they cross their arms and moments later you cross your arms, guess what? Mirror neurons, okay? That's how it works. you start mirroring somebody else just because that's just what we do as humans i don't know how to describe it that's why i keep saying dan tan i mean i never used to say dan tan all right let's go dan tan to my next place here i'm going to change the subject uh let's see there was two that stood out to me what was the database access patterns breaking things and then vectors these two stand out to me you know why does it stand out to you matt since you're concurring uh the date we've already seen this we are hammering our databases now the agents can do queries a lot faster than humans can and it can do more complex queries and stuff now we're lucky this is just the example at grafana labs um we're lucky because the teams yeah that's where i work we are hiring there are teams uh the teams are very good the loki team like these are like i've never worked with engineers with that particular kind of speciality because you know they care about storage uh obviously all the data formats are bespoke that they have to invent you know all the indexing all the kind of complexity that they build to make these systems work really quickly so they they're up for the challenge but we hammered we are hammering them uh you know we had the drill down apps last year that we did where the ui is basically ddossing and we've made that even worse now with assistant and they so they have to adapt and change and they're up for the challenge but is everyone um are some old data techniques or data techniques even going to change are we going to start changing how we store data so that it's more it's ready for agents uh indexed differently vector databases i think as well plays into that big time this is where you create a vector from some kind of content a chunk of content you put it into some multi-dimensional space and then you can query that very quickly and it's as long as it semantically means, like things that mean the same end up in the same area in this multidimensional space.

1:25:20So you know that this is roughly what you mean, the nearest thing to it. Yeah, unbelievable. And that powers, that makes, Cursor does this very well. The others have the same kind of thing. Indexing your code base like that, Cursor makes it basically very quick. If you ask a question about the code base, it can answer it extremely quickly by just consulting its index that it's built. It used to grep all the time, and it would just take longer to go and grep everything to learn. And then it would end up filling its context window too much. Now it will use the vector database, gets the answers right there very quickly.

1:25:55And you really feel that difference. And I think vector databases are going to be a massive new concern for 2026. Yeah, vector databases are interesting. I'm just barely scratching the surface thinking about some of them. But what do you do whenever, like you said, the vector space that you operate in, like those embeddings are created by, let's just say an algorithm, maybe even an LLM or a model or something like that. And that model gets superseded, you know, becomes part of your architecture to kind of keep your original embeddings maybe, or maybe the, the, the, the original data sets you can re-embed quickly to get maybe even better embeddings as that vector database gets used.

1:26:38And there's performance enhancements, there's new technologies. What do you think about that? Yeah, I think, um, so the vector embeddings are different. Um, yeah. Yeah, so we use like, there's some open source ones that you can use. There are some other models that do it. But they, yeah, so I could imagine there being new vector technology, which means you then want to re-index things potentially. If it gets better, sure. But the ultimate, the in and out of it to the LLM is just, here's a sort of search query, and it returns just results that match semantically. So that interface is probably quite safe to keep.

1:27:15But who knows? Different innovations could happen. And one of the problems is like clustering to caching, right? Is it like caching? Yeah. It's like a cache. It's an index. Yeah. It's a, it's an index that, but it's an index that you can, you to, to look it up. You're just doing simple geometry. Like they're quite simple functions to actually find the answers. Cause you do all the work at the time you generate in the embeddings and things that takes a lot of time and process to get, but, but reading it is very quick. Yeah. But if you have lots of content, the more content you put in, then this vector space can get crowded.

1:27:53That's where you end up with problems where it's just, it picks things that aren't relevant. You wouldn't consider it as relevant as other things, but it's all too tightly. Which is when you would do a re-index or re-embed, right? You would re-vectorize. I'm not sure what the terminology is for these because I'm just touching it a little bit here. But when the space gets crowded, you need to figure out a way to give distance and give more meaning because the whole point is to create meaning and create similarity in the vector space but not have to stay there forever yeah so it might be like you would keep a vector of recent stuff so you keep an index of the recent stuff that's what you're going to search because that's in this in this domain that makes most sense yeah but maybe you also have an older one that contains past things and you do you should do multiple requests into this what about the database what do we do there just just better indexes uh more vcpus you know what do you what do you do more ram dedicated machines i think we're going to end up storing data differently so that we store it so it's in a format that the llm needs which is going to be just natural language in a lot of cases there'll be some cases where it's that for what we do with the assistant just as another example because it is on my mind a lot is we have a smaller model Crafon Assistant, so hot right now.

1:29:18So hot. So hot. Well, according to a big AI company, it's one of the best implementations they've ever seen. Which model are you guys using? We use Claude Sonnet 4.5 at the moment, but we're excited about others when they get more affordable. But what we use, we use a smaller model to look at the data. So we use a, so, you know, it'll make a query, but we can easily fill up the context window too much. So we take the data and give it to a different model and say, we taught it how to describe a graph, basically. It's got a spike here and then it dips, or it's flat generally, and it describes it in natural language.

1:29:59And it's that that then gets fed back to the main agent. So the main agent, it's a bit like saying, look at that graph for me and tell me what you see. And it might, and the agent, there's one will say. It describes it, yeah. Describe the color green for me. Very spiky. Yeah. It'll just, yeah. it's all right it's trees that is interesting how natural language has become the language of choice across these things like even if you like even context is simple just like it describes it in words yeah i like it because i can read it thank you yeah yeah until i can't read it right that's what i mean so like the the apis actually like being being just like text that you can read is quite nice for humans as well we don't really do that we have tends to have json api or something you know but it's open to interpretation so right how we store data is going to change and did you finish your thought or did i interrupt you and you didn't finish no no i think that's it yeah i think vector databases are going to shine this year and we'll have to see innovation there.

1:31:08Is there any particular vector database or package or module that you're using in like maybe the Go world that you want to give a shout out to? What's got your fancy? Well, we tend to, so we're basically building our own. Postgres does have the ability to store embeddings. Yeah, but you still need to decide how to generate the vectors. And that's a separate piece that you need to figure out. and that's domain very domain specific right the better you do that the better your search results will be yeah so that's what that's i don't know yeah we tend to i think when we did machine box we did the machine box startup we have the same thing because it would do it with face detection so it would look at the face and it would have it had a big model that was trained on loads of faces we trimmed off the last layer this is kind of a spoiler alert of how we did it so instead of it giving you the answer of a person of the face it gives you basically the vector and then we we have the the spatial index where we go and look up who the person is which allowed us to do one shot learning and also you could delete and forget things because it's just editing in an index so we tended to um do that stuff ourselves it's not it's not that complicated but like doing it at scale and doing it at you know nicely redundantly and horizontal scale all that stuff you know that's where we want services um but yeah i think we'll see them more of them coming out i don't know loads of them um but yeah pretty good interesting so build your own is where you're at at the moment but yeah but also don't i mean yeah we we are doing we tend to we tended to do that in the past um but that's because we we weren't sure exactly the use cases and we wanted flexibility to be able to innovate so it's kind of worth us having our own thing um but i think once there's a once like there should be a service that open source thing there probably is yeah you're playing with uh parquet parquet how do you say that how do you personally say that i say parquet parquet okay cool do you play with parquet at all like do you maybe do like you said running around on the streets and that maybe over jumping over bins and that yeah run up the stairs backwards hardcore parquet that's right as you would say jerry right that's right they do use that in grafano labs i haven't used it myself but that that uh is used and i don't know i don't think that's a spoiler or anything i think that's known yeah but no i haven't no no real experience with that well i have a prediction for 2026 oh yeah uh it's short term short term prediction i predict that matt's gonna get his guitar Oh my gosh.

1:33:49How did you do that? Oh my goodness. Get your parquet on. Okay. Happy with the key? I love this key. What is that? C minor? Where is, what's his name? Knuth? The famous guy who's perfect. Donald. No, it's not Donald Knuth. Singer.

1:34:20singer singer songwriter Toby Keith Charles Charlie Puth Charlie Puth Margaritaville guy Charlie Puth Pitch Perfect Knuth according to Google likely refers to singer Charlie Puth that was an embedding right there. They vectorized that because a lot of people just jacked that one up. They're like, listen, we're going to speed up this search, okay? This is pretty close in vector space. I'm going to parquet that and then vectorize that and then boom, there you go. All right, Matt. Go ahead, Matt. Tell us what you got here. All right.

1:35:13I'm cheaper than an AI now. you're gonna hire me back it turns out chat gpt knows how to negotiate hard and learn from the best lawyers and everyone on reddit with opinions turns out opus 4.58 your lawyers for lunch and it was just a single prompt and the response that they gave to the actual judge contained emojis

1:35:46I'm cheaper than AI now You gotta hire me back Don't give all that money to the robots Humans need money too To give some to me, please

1:36:05Alright Yeah, we've touched the whole range of AI subjects and different musical subjects in a way We didn't really bring it home, man You brought it home. Oh, shit. I don't know what that means. Is that good? That second line was a little if you, but gee, you've rounded it off pretty good. Appreciate the immediate review. You're, uh, yeah. I like to fail fast. Line two. I like it. It's like a linter. It's like error on line two. Adam's the linter. Musical linter. I'll give you three stars, but I'll let you decide if it's between five or ten. Well, I'll give you and your podcast two thumbs up.

1:36:41Oh, thank you. Nice. Out of ten. I do want to give a shout out to Tom Tomaz Tongue Sorry dude TheoryVC.com I really do Appreciate this post I didn't think we would have such a Great I guess somewhat great Mostly great and a song Conversation from Not against venture capitalists But this is good These are good predictions very well thought through so maybe theoryvc.com can be your friend at least for some information follow him on LinkedIn maybe who knows maybe Tomas would like to invest in ChangeLog all day Gerard Change has started a new company doing a distributed VCS put your source code in it or don't, I don't care it's that kind of thing I asked you not to do that I'm sorry I can't help it would you like it but I can't even do your voice really I need to meet you properly and then I can do you you need to meet us properly well I'm gonna meet you properly it's never too late man as long as we're still both breathing which so far yeah so good so if you're not breathing you don't want me to visit it's never too late I might have to hold my breath when you get here but well that's because it's been a long flight you wouldn't have to if you were visited by self-driving Matt you could sleep on the way over you'll be refreshed when you get here that's true oh it should have a shower in it as well she'll press a button and it showers you listen i like your idea of how we'll travel in the future i'm i'm down for some version of that honestly i think i would love to teleport via a trip i'm i don't mind the time because if i can use the time well then i'm cool with the time because i'm gonna use time anyways if i oh yeah that's just i got no choice right i got i got no choice This guy uses time.

1:38:38We don't just pause. He uses it. Yeah. When given a choice, he uses it. But sleeping, sleeping and traveling, I do think is good. Like, have you ever fallen asleep on a long flight? Yeah, because what else are you doing? Exactly. Fall asleep on a long flight. Beautiful. I've done it. I did it last time I went to the United States of America. I'm working, man. I can slay some work over three hours and just be like, what? Three hours? Yeah. You could broadcast from it. Yeah. You could broadcast from your car. One day. Just fake background. Yeah. Don't need to. So if we go agent first, Jared, agent first web leads us to this world.

1:39:11I like agent first travel. You know, let the agent travel for me and then let me know how it went. Yeah. Or you could just email it. Just email the agent. It's an attachment. I just want to stay home. Matt, thanks so much for hanging out with us again. Kicking off 26 the right way with you and your guitar and your accents, even though you kept dissing me. I've still enjoyed it somehow. It's like I'm a masochist. yeah, or I wasn't dissing you. Yeah. But I've had a great time. I've had a fantastic time. Welcome to 2026, everybody. What are we going to do? Let's do something good. There's a lot of kind of trouble going on.

1:39:51And, you know, but there's a lot of us still, we've got time on our hands. Let's do some good stuff. Let's use it. Yeah. Yeah. Yeah. That's how I'm going to do it. Use your time wisely, friends. Use your time. Yeah. Thanks, Matt. It was awesome seeing you. Bye, friends. Bye, friends. And Matt. And Matt. All right, that's changelogging friends for you. And that's Matt Ryer for you. He's truly one of a kind. Oh, and we do have a little changelog++ bonus at the end of this one, in which Matt tries and fails to give us a Grafana Assistant demo. No, 100%. How do I share? I'm clicking the share button and nothing is happening.

1:40:29Thanks again to our partners at Flight.io and to our sponsors of this episode. Namespace.so. Notion.com. slash changelog and tigerdata.com. Thanks also to BMC for the continuous stream of Dope Beats and to you for listening. We appreciate you. That's all for now, but let's get back together and talk again real soon.

1:41:09Can you share more about the Grafana assistant architecture? He would love to. Yeah. How much can you share and what can you share? I'm really curious. I'll tell you anything. Ask him anything. I'll get fired for this show. Change log plus plus. It's better.