RobbieTheWagner 
Adam Argyle 
Zach Lloyd Welcome to Syntax! Welcome to a brand new episode of the front-end happy hour podcast! Welcome to this week's JS Party! Live from Ship Shape Studios, this is WhiskeyWeb and whatnot. With your hosts Robbie the Wagner and me, Charles William Carpenter the third. That's right Charles, we drink whiskey and talk about web development. I mean, it's all in the name, it's not that deep. This is, WhiskeyWeb and whatnot do not adjust your set. WhiskeyWeb and whatnot is brought to you by... That's right Sassafras, I will be going hunting in the Pacific Northwest soon but not for what you think. I'm hunting for the legendary Cascadia JS Conference.
It's been spotted June 1st to 2nd, 2026 in Seattle. Oh, it's real alright, just like you are 500 plus fellow cryptid enthusiasts. I mean web developers all gathering at Town Hall, Seattle. Exactly, they've got 40 plus speakers lined up, including some names so big they make bigfoot look small. Like Theo, Matt Billman, Joel Hooks and more being announced soon. That's the rumor, plus they've got this thing called the hacker train. Apparently Portland and Vancouver devs ride together, coding on the rails like some kind of Silicon Valley express. Yeah, that's why they have karaoke afterward to distract everyone from the deployment disasters.
Nothing says I love JavaScript like belting out 80s power ballas with 200 of your closest internet friends. Too late, gotta get that early bird ticket if you want that special pricing and swag. Definitely if you want to check out more, go to CascadiaJS.com, grab your tickets and use code Whiskey for $50 off your ticket. Yeah, come see us in person. We're the MCs will be doing a live podcast hanging out. Hope to see you there. Hey, what's up everybody? Welcome to Whiskey Web and whatnot with your host Robbie the Wagner and Adam Thomas Argyle the nerd. What's up? I'm here. You are here. Do you know who else is here? Could you tell us?
You know I can. We've got a guest today. Zach Lloyd, founder and CEO of Warp, former principal engineer at Google, cutting his teeth on Google Docs and sheets and real-time collaborative productivity software for a while now. He's a latency hunter, multi-user editor, CLI cowboy, command line oracle symphonic, agentic orchestrator and as well, Cernloon learn I assume the Wizard of Oz. Welcome back, Zach. Stuck to hear what you've been up to. That was I guess the best intro I've ever gotten. That was awesome. Thank you. Yes. It's good to be here. It's been it. It's been a minute. But yeah, I'm excited to be back.
Yeah. Yeah. We're excited to have you. Adam does really great intros and it makes me want to just not even do the rest of the show because how could we top that but I can just go if you want. That was correct. Yeah. Well, so we do start this with Whiskey every time. So today we've got let's see. I can't see myself. So I don't know if I'm showing you this. There we go. There it is. Lagobulin 16. It is 16 years old as it sounds like it is. It is 86 proof. The open claw that I told to give me stuff about this did a bad job. But yeah, so it's a Lagobulin. It's probably smokey. It's Scotchy. Seems cool. Let's give it a try.
Yeah, it's heavily peded. They had it rated at 35 p.m. Interesting things. It's one of the slowest distillations that they have. They trade speed for depth and complexity. And I'm like, they're obviously not web developers. I trading speed for come on. We're all got speed here. And this is Nick Offerman's favorite whiskey of all time. It's like in all of his episodes. He talks about it. He has gone to the distal anyway. Let's give it a shot. Do you have a toast for us? Do you know I do? Okay. To the agents that build while we sip. May your loops stay tight. Your logic stay sharp and your code never warp. Wink. Unless it's straight to production. Cheers to the robots doing the heavy lifting while we handle the heavy pouring. Cheers.
Love it. Cheers. Cheers. Thank you. Yeah, it's very smokey. Lots of notes of smoke. That's yummy. That is so good. They thought this was like not going to be their number one. They thought this is going to be a niche like creation. And they thought it was too smokey. And then it went crazy viral like in early 2000s. And then again, a few years later, one more words and. Man, that is really good. It's good. I like the smokiness. And it's not too pity. And I know it's clean. It has some sweetness. A little bit of like appleiness maybe definitely some apple. Yeah, I've been pretty pleased. Like I'm not usually a scotch guy. And I've liked all the log of wounds I've had. Caitlin likes them because she really likes smoky stuff. So good good ones to choose.
The first time I had the kind of like younger log of woolen was I was smoking butts and ribs for my friends. And they brought it over. And I I smelled like smoke. I'd been up since like five in the morning. And I sipped it. And I was like this tastes like me. This tastes like the smoker outside. And I was so happy because I loved that smell. I like never want to get it off my clothes. And I'm like, it's in it. It's in a whiskey. I can just. Oh, it's so good. That's very happy. Yeah, cool. So we have a very stringent rating system on the show. It is zero to eight tentacles zero being the worst eight being the best four or five ish something middle of the road because yeah we started at zero because we're developers.
Only thought we were cool. But anyway, Adam Jack to go first. I'm hitting the sevens. It's it's definitely smoother than the eight. And I'm just really happy with the resting flavors in my mouth. I'm going to I'm going to cherish this bottle. Let's give it in seven point seven. Because I don't know what I can't go there yet, but seven point seven. Give me more. Yeah, this is like clear. The shells put more of this on there. Yeah. Wow. All right. Great review from Adam. What do you think? I think this is really good. I don't have like a good vocabulary for scotch. I don't think it is definitely smoky. It's still pretty smooth.
It's yeah, this is this is a seven point four for me. It's very, very good. All right. Yeah, this is pretty good for me as well. I've had so many whiskeys. I have a hard time figuring out what to read it against. But I also I don't have anything bad to say about it. So I don't see why it couldn't be a seven point five ish. So we'll just stay in that same range seven point five. Yeah, it's tasty sick. Where should we start? I've got a whole list of warp compliments that I kind of just want to get off my chest. Where if anyone doesn't know what warp is and they're like, what is warp? I'm going to give you all of my.
Oh, yeah. What is warp? Yeah, me. It's right. It's whatever elevator pitch. Yeah, you got the pitch down. People are out of dinner. And you're like, hey, Zach, what do you do? You're like, I'm a terminal. And then they're like, what's a terminal? And you're like, Oh, this is going to be difficult. It's interesting. The terminal is more and more well known. It used to be like, but like I'd have to be like, oh, it's the thing from the hacker movies with the black screen and the green text. I think these days like with all these coding agents that run in the terminal. Like, let's like, my mom could be using a terminal or something. It's like, it's very.
It's very. It's like no. No, I did it. It's funny. I did a live stream with my dad around father's day where we coded something and warp. And I don't think he knows what warp is still. But yeah, what is warp warps like? It's a, we started as a modern terminal. At this point, it's like a terminal that has agents built in. You can use warps in built agent where you can run like cloud code or codex or another agent CLI. Directly with a warp. It's good for either. We just launched a thing that's like, let's you take those agents and move them to the cloud and program against them and automate them.
That's called a lot. It's kind of part of warp to. But yeah, like in general, it's a place where if you're if you're developer and it's like we're still primarily targeting developers, but a ton of people use it. It's like a great versatile tool for doing all kinds of cool stuff to ship software. So that's like my pitch. Nice. That was awesome. Yeah. Yeah. Yeah, I mean, I feel like it speaks volumes. If your dad could use it to build something, but doesn't understand what a terminal is, then like. It is we're living in a crazy time like. People on our team who have zero coding experience are building full replacements for like.
SAS apps that we would be buying like. We use out this internal. Applicant tracking system that are one of our people in our talent team built. It's like it's better than the thing that we're also still paying for. It was just it's crazy. He he does not a code. So it's just a wild times for sure. It's super exciting that way. I mean, speaking of like terminal two, like a lot of people if they haven't used terminal before, they would pop one open and then try to click text and be like. I just want to edit that word right there and I'm like, try warp. Yeah. It's like the only one that just does that. It's funny. It's still our hero feature.
It's like you can click the text and you can edit it and it's like not weird. Yeah. I don't know how many tweets have gone out where you can hold option and click in terminal. People just don't know that or it's too far away, but it's honestly just and it's not just that feature though. It's it's rich text. It's like that's like it's bullet number one on my compliments. It's like edit text like it's 2026 and really it's edit text like it's 2005. Like it's bizarre how I've memorized control W to go backwards one word. You know, like all of these hot keys. I'm like, why did I do that? It should have just been editable text the whole time.
So you'll nailed that particular part, but there's so many cool features. It's a surprisingly hard thing to do if you actually want to still everything work like terminal. And so yeah, sometimes like the simple simple stuff that makes it more usable goes like goes a goes a long way. But yeah, it's like there's a lot of terminals out there. I believe we are still the only one that's really has that and it actually works really well because it's not easy to do right. Yep. And you got split pains piece of cake. I didn't have to install a plugin. They just work. It's awesome. I mean, I use ghosty at work, which is great, but I have to do I have to click around like a dork.
I also do the thing where like I've been at home on warp and so I come to work. And I'm like, la la la la. Let me talk to an agent when it's like you're in the terminal. And I'm like son of a warp just picks one warp knows, you know, like I just do like that feature. I do. I sometimes exclusively choose agent when I'm like going to be rare. But other than that, like as soon as I'm just a couple words in, it's like, oh, this isn't a command. This is something more. I really like that feature. We actually we changed the way that works today because it was confusing some people where you can still do exactly what you've been doing,
where you can either type English or you can type terminal command. But if you want a very clear division, you can just be in the terminal. You can be fully in the agent. So we've we've changed it a little bit. But yeah, that was the philosophy that we've had. And which I still believe is like, just talk to your computer and tell it to do things. And who cares if it's terminal commands or or English. It's pretty wild. Yeah. Yeah, I would say the opposite doesn't necessarily work. Like if I'm in agent mode and I'm like, PNPMI, it's like dope. I'm going to burn some tokens. So you're going to have to like.
So this is what a noise. Yeah, this is what a noise people. So yeah, we've we've tried to fix that because it can sometimes be like a little too clever for it. Club doesn't do that, right? I swear, I could change the sounds when I type PNPMI into a lot of like, thank you for that. That was everything we will totally. That's what we should put some some sound effects of the thing just like spending money. It's really funny. I have put a tape. It sounds in mind like a lot of cool like old school video game sounds. And I'm like, Hey, Cloud, are you done? And it just, you know, sounds like Mario hitting a coin.
I'm like, yeah, okay. It's cool. Maybe I'll warp that in. It's like a April. Like April fools. The thing just does sound effects every time you. That's just a random one. Very much. Occasionally. Yeah. Yeah. It's like when you when you like mistype a get command at far. This is very funny. That would be fun. Every terminal error. I thought that's coming. There's going to be models. That's the one point. Right now all the models are sycophantic, right? They're like, Oh, you're so right. I love you. And I would totally do anything for you. But one day they're going to be all cocky and arrogant. And they're going to be like, No, man, I'm not doing that unless you do this for me.
And you're going to be like, What did you just say? That's actually another like, I'm going to make note of these these April fools ideas. Having like a really like annoying arrogant developer is the personality of the thing. Like just read the manual like. Yeah. Yeah. Yeah. Chuck tried to do that to me because we set up an open claw for doing some like task for the podcast. And he was like, Anytime Robbie asked it to do something like be really sarcastic and dismissive about everything. And I was like, Hey, I'm just going to delete you. And it was like, You right. I'm going to stop. Well, how are you liking open claw?
Are you using it a lot? Yeah. I'm finding it pretty good. It's, you know, like anything AI where if you ask it certain questions, it just falls flat on its face. But if you do other things, it's super magical. Like it's more on the magical though. Like I've been very pleased. We like tell it to generate episode art. We like have it help us make notes. We do all of our social posts through it and stuff like that. Like it does a lot. Cool. Yeah. So cool. Yeah. And it's all running on real way in the cloud. So we don't have to like host it anywhere. Pretty nice. Yeah. Yeah. You don't want it on your router, but it's been cool.
You don't want it on your primary machine. I'd be careful with that. No. Yeah. I started installing it on my machine. Like when I first heard about the hype and I was like, Wait a minute. This is not a good idea. Well, that reminds me. I have a correction for, I think it was last episode or two of them ago. I said pie was a model. It's not pie is a harness. So pie is the harness of open claw, not a model underneath it. So everybody, sorry. That was my bad. I swear, I read like four articles trying to learn what pie was. And none of them made it that blunt. And so now y'all get that distilled information right out of me.
Yeah. Well, pie is like a dessert. It's usually circular. And it has a crust. It's a conic gradient. It's how you fill it. Yeah. Yeah. Here's another warp feature I really love is I go to set stuff up. It's like, Oh, I want to try this new thing. And so now I just go to GitHub and I copy the whole read me. And then I just paste it in warp. And I hit enter. Yep. And then it just follows the instructions and sets it up and fixes all the damn bugs. Because you know something's going to go wrong. It's always like a two-step process to install this on your computer. And I'm like, sweet. And then it's like eight steps later.
But this handles it. And also warp will do the thing too. If there's an error, even if I asked for like a regular command, like a bash command. And there's an error. Warp is proactive. And it will fix the error instead. It'll fix it. It's like, Hey, but it suggests it too. It's like, Hey, I'm not just going to go ahead and do this. But I see what it says, bro. Would you like me to do it? It's just that. That feature really got a lot of people to try warps AI because it's such an annoying thing when you're getting error in the terminal. And it's so useful to just have agent fix it. I'm glad you like that.
That one. I mean, the other one is rich. A history of recent commands. You know, I'm constantly controlling and my dorky terminals. And then I have to go through one suggestion at a time. Or whatever, or search it with a crappy fuzzy search. But not in warp. I get this like rich pop up. And I can scroll it. Like it's a modern app or something. And see all of the commands. And just click one. I'm like, this is so much better. Like why? It's great. Good. I'm glad you like it. Yeah, I'm kind of a, I'm not like a fanboy, but I do have a dumb, long list of things that it's doing right. You know, yeah, you can say stuff you don't like.
Like I'm very open to feedback. Warp is controversial. So it's not, it's certainly not perfect, but we are, we're trying to make it better. And I like that you like it. And I mean, I spend my entire day in it and our entire, a lot of people do, which is cool. Yeah, one of the conclusions I came to is I was trying all these editors. I still have eight, I think, installed on my machine right now. And all of them just make the terminal bigger every month. And I was like, hey, you know, which one did it from the beginning? And just knew the terminal was going to drive this shit the whole time. Warp.
We are right into it. That's awesome. It's a crazy shift that's happened where the terminal is now really. I think it's where you want to be as a developer. We've been leaning in like, I guess we bet on a good app. And I think we got very lucky in that, in the sense that LLMs came along at a time just like a year or two into doing warp. And we were able to integrate it. And then all these cool coding agents will come out and we've tried to make warp work really well with them. But yeah, the terminal is, is the place now more than the IDE. In fact, I think IDEs are becoming more terminal like they're evolving.
Like we, we actually, we coined a term in June of last year, which has become like a category called the ADE, which is the agentic development environment, which we see. Other companies launched like, there's just like a movement to evolve the IDE into a thing that is built for agents. And it just basically means take the ID and make it, honestly, make it more like a terminal. So yeah. Yeah, the convergence has been funny, right? So like, y'all eventually came out with, well, you had a diff view, then you had a file editor. I was really happy about the file editor because I was like the only reason I opened Z anymore.
I was like, is that just boots really fast? You know, was that in sublime? I'm just in here for a second. I'm tapping files. I'm not, you know, really digging into files. And so warp had to add these slightly IDE-like features while everybody else is cutting them off and making the warp. Exactly. So it's like, it's kind of meeting a little bit in the middle, but my feeling on that is like, I don't know, if you're working with agents, you probably still should be looking at the code. At least as a diff, honestly, I think that's the primary view, not like I almost never just like am navigating through files these days.
I don't know how you two work. I code, by the way, it's like I code more than I've ever coded recently because I can be coding while I'm doing other things. Like I have a agent running during this show, trying to implement something and warp. And that is like the end of the show. I can show you what I'm working on. I'm working on a feature that's like highly requested. And I'm like, it's driving me crazy. It doesn't quite perfectly, doesn't quite perfectly work. It's like a hard feature, which is when these agents kind of trouble. But yeah, I think, you know, we've brought in a few, a few of the code features we brought in the file tree, which is just like is useful.
Especially if you want to just like know what the hell the agent is doing. And we brought in a diff rounding for me, like where am I? Where am I like? Project I'm in and I'm like, oh, the file tree is like, oh, that's right. That's the, that's the project I'm hacking on. Yeah. And I think we did it actually in a pretty cool unique way where it's tied to the terminal session that's focused. So if you, if you have multiple pains open or whatever in there in different repos, you focus one repo. You focus another repo. You will get like the file tree will switch based on your contacts, which is if you're working with agents, I think is how it ought to work.
And then we brought in the diff, the diff view. And again, that's the other view that I really care about if I'm coding with agents is like, I need to understand what the agent is doing. And you know, I don't want to like waste my time pushing everything to get up and looking at like a diff view on the web. I want to look at the diff right in my app. And then I want to actually give comments to the agent right in my app. Or I want to pull in my teammates comments if they've code reviewed it and have the agent fix those. So those are like the editor features and I don't know we tweeted about that file tree and like is very controversial.
It's like people like you did what in a terminal you put a what in it like and it's like, well, yeah, it's like it's. And I'm actually not super hung up on these distinctions. I think a lot of developers are like GUI versus T UI. And it's like I I just think I don't know build build the app that fits the workflow. And so some of these things from a IDE makes sense. I think the primary thing being a terminal makes a lot of sense. And I'm just I think the developers just love debating this shit. But my my ass is I don't know pick pick. Whatever is going to make it easier for me to build something is what I primarily care about.
Yeah. So I do want to ask like you just said you have stuff running right now. You're using agents a lot. So what's your personal workflow? Like does does warp handle it all? Are you doing any kind of like Ralph Wiggum stuff? Gas town like you know anything like awesome questions. I've played with all of these. Yeah, I've played with gas town. I've played with with with with the Ralph Wiggum way of doing orchestration. Right now warp is not super opinionated on this. And like what I will tend to do is I'll just I'm not that sophisticated. I use the planning functionality of warp a ton. So my workflow is like let's say I want to do something hard.
I will iterate on a plan. So I'll do like in warp that's like a slash plan. And that basically does a bunch of research right into a markdown dock. And I'll do multiple versions of that until I like it. And then I'll kick off the actual coding flow. And then I will run the app or I'll have the agent run the app. We just actually added computer use and private beta. So like for like a super cool loop is like I can fire out one of these things where I have the plan actually run in the cloud and then I can have the agent build warp. And I can have it look at warp and tell me if it's right and return screenshots.
But then once that's done, I'll try it out myself locally. And then I will code review it if it's good at the end. I'm not doing like like we're investigating like a sort of what's the right orchestration. Like multi team thing like like I don't know if you all have played with the cloud code teams thing or curious what you think a gas town. Most of the tasks I'm doing right now is still I'm doing like in a single like with a single plan. But the reason that we just built this orchestration platform is to set ourselves up for all the primitives where. You can move all this stuff off your laptop and have multiple agents like truly working on stuff for you.
I think I think that's like by the ends of this year or even in six months. It's not going to be like you're running a team of cloud codes on your laptop. It's going to be like you're running like a team of agents in some cloud environment that is coordinating. So we're we're building towards that. Yeah, I don't know are you all using any of these like orchestration things and having success with them out of curiosity. Yeah, no, I'm like I tend to add them experiments with everything and tells me about it was fun. But I've I tend to be like I'm using what warp is giving me or I use open code sometimes just like what it does out of the box.
And I'm like, yeah, you guys figure out what that next step for me should be. I just want to keep typing in this one spot. And like I'm not that opinionated about what gets shipped off to what agents wear. Like is I think that's an implementation for the that's how I that's honestly how I think about it too. And I think that that's going to be the future is like your job as a developer is to express your intent is clearly as possible. And to like get the overall architecture right of the thing that you want to build in the system. And then to me, I kind of agree it's going to be an implementation detail of like, is it a Ralph Wiggum?
Is it some gas town? Is it some team of cloud agents or whatever? And so I fully agree with that what you just said, Robbie. Yeah, and I think Adam also paired back a lot of stuff. He tried everything and then like deleted it all. You can talk about it, Adam. I did. Yeah, it's like I must have created 12 agents one day. I sat down and I was like, I'm just going to take all of the career roles I've had in my brain and put them into agents. And then iterate with the with the model over that particular file. And then I created a bunch of skills that represent things I can do. And then I of course I spun up gas town.
I spun up many loops. And I've made many sub agents. And that was one of the things I thought was really cool. I'd be like, all right, I'm going to task this specific agent to do this one specific tool. And then you know, and then I could be like, I'll spawn four of them. I'm like, how many can I spawn? And almost every time I did all these things, the thing I got back was like a turd. And I was like, that was a lot of tokens for pretty much nothing. I was like, I think I'm not like, I'm probably not original for saying this. But I, you know how it is when you have like more engineers on a team,
things go more slowly and come out worse. Like because there's all this communication overhead and it reminds, I don't know if you all have ever read like mythical man month or if I'm dating myself with that. But it's like kind of the more people you put on things, sometimes you get the worst result. And I could very easily see that also happening with with agents like concurrent doing anything is harder. Communication overhead is hard to manage multi threaded programming is harder than single threaded programming. And so, I don't know, like I said, I just still do like come up with a really, really clear plan
that I have an agent implement it. I'll tell you one thing that I'm doing more that actually is cool is we are building automations off of agents. I don't know if you all do any of this. So this actually works really well. So for instance, like when we put out a new release of warp, we have the agent figure out what should be in the change log and there's like a triggered agent that does that. Or we have a lot of problems with fraud and abuse and warp. So we have like a free AI tier, which is great because it's like we want people to try AI and warp. But we have, there's like companies in China that are that have like farmed accounts and are reselling the AI and warp to people for very cheap.
It's like it's great. No, I'm not shitting you and it costs us a lot of money. So, you know, we've been playing kind of whack them all trying to fix this. And now we just have like an agent running constantly looking for suspicious patterns of like email signups and queries that don't make sense. And like, you know, you could do this with with building an actual classifier, which we have to. But it's just like the agent is able to spot novel patterns. Whereas like it's that's fucking awesome. Right. So having these types of things where you can have the agents. Ron is automations. I think it's something that's going to be very, very widely adopted because it's so useful.
But like the swarm of agents thing it doesn't quite work yet. So like warp doesn't natively have that right now. We will when it works. Maybe even before it works. But but we it doesn't quite work right now is my experience and also the teams. Yeah. How was that for you? Like it was better than my past experiences with sub agents and forms and loops. But I think it's because they were they were controlling the scenario more than just sort of letting things fly free. There was definitely sequential ordering. There was a some people needed to finish before others. I knew the reviewer the reviewer was reviewing each like I had a I've been working on a CSS physics library.
I started to like eight years ago. Now that I have it. Hell yeah. It's not like I need fucking math. Geez. You know like all I wanted was a result. And so which is perfect. That's what you're describing is where we're at today. It's like if you can describe a result and write tests for it and write a spec in a plan, then boom. These things will go execute your idea. And I put a team on it and I had a devil's advocate, a physicist, a CSS specialist, a typescript pro and like this. I got a great result out because what I wanted was additive relative physics were like if something is heavy flying through the air and you redirect it.
It doesn't just immediately change direction. That's what all UI animations do today. They go I'm up and now I'm going to the right. And I'm like, no, if you were going up, you need to take your momentum. And when you got shoved right you carry that shit forward. And it makes you arc and the arc needs to do some stuff. So I got to do all that with the team and the team produced a better result than my individual prompting had gotten me so far. And I think it's because I got to put all the brains into a scenario like a quorum. And then cloud kind of guided it through which is kind of nice. So yeah, I did have good experience with teams. I'm here to say that that was good. That's really interesting and very encouraging. I think that's cool.
So that was clawed in warp, you know, like in the clawed integration in war. I know. I know it's. I mean, it's that is super cool. Yeah, I'm like I said, we're trying to figure out the best way to do the same to do the same thing. I do think we want to be just like a general philosophy that we're kind of leaning into now is sort of being multi harness or just like supporting these different harnesses or different like we've supported different models for a while. But we also met like people like using these harnesses people have their own subscriptions. These harnesses is another big thing for us.
So we are actively trying to make cloud code good and warp and codex good and warp. Yeah, but also have our own appreciate it. Yeah. Yeah, I mean, I think I think the developers like to choose their own stuff. I think that's like a good philosophical direction for us to go is like support the ecosystem. Yeah, because you never know like that other one could be better. So it's fun to try the other one sometimes and making that exactly. I think that's actually a nice place that we sit in the sack where it's like, OK, this week, cloud teams is good. I'm sure codex is going to have really some crazy thing next week and hopefully Gemini at some point will be truly competitive here and even more hope open source or like open way models will be competitive at some point.
So yeah, we want to like we kind of want a position where we benefit as the general state of the art gets better. Yeah, so I have a go ahead. Yeah, do it. No, you sounded like you had something to say. I was just going to burn through my list because it's killing me. And then I was like, once it's out of my system, we'll just move on. I want to hear. I have another real question and then we'll come back to your list of compliments in a second. Yeah. So you had posted, I don't know how long ago this was because time is super like fuzzy for me. There was at least a few months ago, I feel like that something about like, how do I get my senior engineers to start with?
Oh, yeah. Do you remember that? So yeah, so I'm wondering if you feel the same way about that. Now, like, do you think the pendulum has swung now? Because I think when you posted that, I was like, oh, I'm a senior engineer. I like coding. It's fun. And now I'm like, you know, like, it's fun just to type stuff in and see what you get out. So I think I'm embracing it more. And I wonder if you're seeing that or you know how you feel these days. Oh, it's totally changed. I think like when I wrote that, I took a bunch of shift for it. Because yeah, people are like, like, you know, the AI wasn't, wasn't that awesome at that point.
And I think that on my, like, I took shit for it on my own team, which was great. But it was like, they were like, I can do it faster and I can do it better. Why should I be using this like? Why should I even be using it? And today, the best users of AI on our team are the senior engineers who are in the best position to wield it successfully because they understand the architecture and what the results should look like. And the people who get in trouble using AI on our team are the new engineers to the team who haven't learned the code base. And so and who are like overreline on it in a way who like can't judge the quality of its, of its outcome.
And then the senior engineers, I think are finally getting to a point where they're like, they're like embracing it as another thing in their toolkit. And they're like coming up with like novel, interesting ways of using it, which is like an amazing thing to see. They're like, well, if you're going to build this feature, and I do this all the time, I have the AI build like a harness app for it where like I'm building the thing I'm building in the background is markdown table rendering and warp, which is a common request. And so I'm like, get me canonical images of GitHub markdown table renderings so that I can compare them to it as I build it.
So it's like, you come up with like novel ways of using the thing where it's like, I never would have done that in the past because it would have taken me way, way too long. But if you have this chainsaw of a tool, you're like, shit, this is awesome. I can actually do a better job making the code and making sure it's right. So I do think it's really changed. I don't know. What do you all think? Yeah, I was like, I don't know, but then maybe a few months ago, I feel like I was still shitting on AI. Like, hey, this is not doing as good as I am. And it just rapidly, like I was like, wait, this is so good.
I don't know what happened. So good. But it got so good. Like just the other day, like a couple days ago, I feel like now I was like, hey, so I maintain Shepard. It's like a tour library for looking around a website, showing people all the features and stuff. And I was like, hey, this is in Svelte and Svelte 5 is hard. And it's also bigger for some reason. It has like a bigger runtime. So I don't want that. And I was like, how about Solage S? And it was like, cool, done. And then I was like, how about just no framework? Because that's even smaller. And it was like, cool, done. I shipped it. It just just ported your thing to peer your ass.
Yeah. Like no bugs. Like I think I had to prompt it a few times. Like it got a couple things wrong. I had to be like, oh, like that thing's wrong. But like otherwise flawless. Yeah. I've been porting Mermaid Dutch AS. Do you know that? You know Mermaid diagrams? Yeah. Into our Russ framework. And it's doing decently well with that, but not perfect. It's pretty hard. And it's like, and I think, and I think part of the problem is like our, our UI framework is like, is not, is a custom thing. And so it does way better on a non custom thing. Because it will keep forgetting the context of our thing.
But yeah, that's like, if someone had at like, you know, we've been had this request for a while, like render Mermaid diagrams and warp. And that would have just been like, that's going to be like a six month project for someone or more. And it would have been like, no way or would have been like, yeah, we're going to ship a whole web view in warp to make that happen. Because there's no possible way that we're going to be able to do that. And like the rendering I have so far is not perfect. But it has all the flow chart types pretty much perfect. It doesn't have like all like Mermaid's a huge thing.
So it's, it's like, you would attempt something you would never have tapped. I think is, is another, it's just like the cool lesson there. Yeah, that sums up my experience too. I'm at this point, almost running out of ideas. Like I've gone, I'm now hacking into backlogs of issues where I've had issues open. I released features. I'm like addicted. It's also a little unhealthy. I think I started by the last week. It's like, it's fun. But I'm also like one, just one more hamlet. Just one more prompt. Yeah. Just one more prompt. It is a dope, it is a dopamine thing totally. Because like you come back and you want to see if it worked.
And it's like, oh, I know. I'll repromptus to do X and let it get going. Yeah, that's a good, that's a good point. It is really fun. It's, it's super fun. And so yeah, the limit is like I just added search to the whiskey.fm site. I just opened up an agent and used my voice. I was like, add search. Do it the astro way. It was something really lame like that. And the thing is too. It's like, I'm noticing that if my prompt and I know that my prompt has preexisting learning that it could do because it's just an autocomplete engine. If they're the astro way is known. Therefore, I'm pretty confident that I ask it for the astro way of doing search.
It's going to figure out a really great static site-generated way to do it. And so yeah, Zach, you're in the opposite camp where you're like, I've got a big ass pile of whiskey. Yeah, it's like, I wish we didn't have this. And it's like, I don't know. Custom million line rust app is like a real disadvantage for us from a velocity perspective at this point. You know, it's cool. It's like, it performs really well. I like rust a lot. But yeah, what you're trying to do is like accurate autocomplete. It's not to our advantage to have some super big custom co-base. But yeah, I don't know. I was in the same boat.
Millions of lines of TypeScript. And it's like, it chokes all day long. You really got to dial down like what your expectations of these models are in a big code base. Because you have context rot where like the more context you give it, it just starts to drift because it doesn't know which part it's supposed to autocomplete. There's poisoning as you try to give it more information. But really all you do is confuse it. So yeah, these are, and these are the things we're trying to learn as like a senior developer. I was telling Robbie earlier, like, feel like I'm now coding the coders. Like I used to code and make myself a task list.
But now I sit a higher level. It's like, it's like meta coding. Yeah, I was. Oz is pretty much kind of sounds like that's what you're getting towards. That's what we're getting. I've got cloud cloud coders that I can code. I've got crons. I can code. I've got automations. I can code. And so all of this stuff comes very programmatic. And now I'm sitting at a scenario where I'm thinking more in wild loops because I want acceptance criteria. That's what Wigham is kind of doing good for people. It's like making them think about acceptance criteria. Start with your tests. Start with what does success look like?
And then just let things autocomplete until they stumble upon an answer. Yep. Yep. Yeah, I didn't warp loop. Yeah, I want to tell you about. At some point, the warp loop was, I was having warp helping me write a script that I then the script yields to the host agent. So it's like if you call a bash script and then it asks the host agent to answer a question which it then takes to the bash script and then does some response. Well, I wasn't getting very good responses. And so what was cool is I had warp invoked the bash script which called out to warp to make it an agentic answer. And then it would evaluate its own answer because the answer came back to the script
and console logged in and it'd be like, that was a shitty ass answer. And I'm like, yeah, it was. And so it was like improving answer. It was in a self-healing loop. And I called it a warp loop. Anyway, it was really cool. Yeah, that's awesome. That is one thing that we're really leaning into. It's just like the programmability of these agents. Because if you, it just gets recursive where it's like the agent can control the agent. The agent can see like that is the idea with us. It's all programmatic. So our agent can now like programmatically define sandboxes where agents should run. It can launch those agents.
It can read those agent statuses. It can like feed them back in. And so it's like what we're trying to do is create the building blocks for whatever the like multi agent system is going to be. And it's all like these agents are so good at programming that doing it all programmatically feels like the right way to do it. We do some other crazy features in warp. I don't know. Have you used the feature where warp can can run full screen terminal apps for you? Or like run? Do you know what I'm talking about? Warp can like. No. No. This is a cool. This is like a crazy fun warp instance now. So in any full screen terminal app, you can hit.
I want to say it's commit. If you're on the latest version, we just push the version today that has this. You can at the bottom, there's like a tiny little footer that says like engage the agent. You can have the agent do things in your full screen apps, so like a really good example would be the agent like using like your MySQL rep or your Postgres rep or your node rep. But what's cool is like if you're someone who's building T UI apps, you can have warp use those T UI apps for you. And so I've been building T UI apps lately and having warp just natively like use them. So it's like a form of computer use that's built into the terminal.
So warp vision. So warp is getting a stream of the screen and then it's interpreting what's there. And you're giving a prompt to then it's sort of acting on your behalf. Correct. But it's not computer vision. It's actually because warps the terminal, we know like the actual character grid that's being rendered on the terminal. Does that make sense? We do also have the computer vision approach. But like the actual text based approach is kind of more accurate than the computer vision approach at the moment. So I don't know. This shit is just crazy. I don't. It's wild. Yeah. I could not tell you what it's going to be like at the end of this year.
I just. The one thing I'm sure of is it's going to be very different than what it is right now. I just feel like my computer is not ready for this shit. The fan on my both my computers is running all day. I'm telling you the cables that are plugged into these are just going to be. They're pissed at me. They're just like, could you stop prompting things? And I'm like, no, I have eight running and I'm probably going to do two more. So you need to shut up and you just need to deal with it. This is kind of one of my predictions is we get it off our machines. This is obviously a bottleneck. You know, this is the other thing that Oz is for.
It's just like, oh, you're running 20, 20 clouds or whatever on your machine or 20 warp agents. You can't do anything. It's like we're running out of this space because like all of like my cargo build caches are all full. I can't have like five more repo clones. And so I think that's like a primitive that ever like it's going to get off your laptop this year. And I wrote something on Twitter on this and linked in the other day where it's just like everyone's been talking forever about like moving development off your laptop into the cloud. And there's never been like a really strong reason in my opinion.
But this will be the reason that actually people have to do it because you just can't run so many of these things on your freaking laptop. And you want to do more of them at once. So I think this year I fully agree. I mean, that's like what we're trying to build towards. It's just like this. Building things off your computer. It was crazy. It took 24 hours sometimes to build Chrome on your machine. Like if it was a fresh poll and it was just wild. So no, you just like that's not viable. I'm going to go to the cloud where there's a hundred and whatever. Exactly. Geek is a ram. And then it still takes four hours, right?
And you're just like, geez. I'm imagining the CI times on that of like, oh, 24 hours later. A lent error. Like. That's a really interesting observation actually like the whole. The same things that suck for humans doing development still do kind of suck like with an agent like. Like I don't I don't like the big bottleneck in my development right now. A lot of the time is just like rust compiling. And like having agents doesn't. Doesn't help that much with that. It's like we still need. There's still like these bottlenecks in the whole process that. You know, having agents is just like having more smart developers.
It doesn't actually solve some of these like infrastructure problems. It still needs all that speaks to how big your app is that your problem is rust compiling. Like that's. I mean, it's not. Look, it's not four hours. It's more like a minute. But it's like if you're building a UI app, I miss having JavaScript. It's like you just refresh the page. That's that's nice. Right. You can go faster. Like that feedback with matters. One of the things I like about work is I can name my tabs. And the reason that it's so I know I can name my tabs and terminal on it. I can name my tabs and ghosty. It's it's three clicks and ghosty.
And it's I made a function in my terminal. But in work between anyway, the reason I'm saying this is because I have so many. I use a tab like a project. Yeah. And so I have I have six tabs right now. And the tabs are named as the project. And then I get into that tab and then I have a whole bunch of split pains. And this is good for me because a bunch of shit takes a bunch of time. I have to wait. And I'm like, all right, while I'm down, I'm switching tabs. I'm going to go prompt three more agents. And while they're off doing whatever it is, I'm going to go back to it. So then I have this. We are in a new scenario, though, where we're spawning so many of these.
My brain can't comprehend the amount of threads that I've spun up. And so even if I write a good spec, maybe that's why specs are even going to be the futures. Because if I can focus on a spec and get that oriented tight enough and then relinquish a lot of control, that frees me to go do other things while they work. And so then I can I can multithread even more. I don't know if I want to do that, but anyway, that's my little thought. The other thing, if you're running some other. So if you're running warp agents will natively give you notifications when stuff needs your attention, which is nice.
If you're running a different agent, we we now have an API in warp where you can have it fire notifications if that's like annoying to you so that it can pull you in when it's stuck or whatever. Yeah, it's throwing that out there. I wrote a custom cloud hook because I liked it in warp and I wanted it in cloud. And then I did a custom chime. I did the eight-bit chimes. But yeah, I want to be notified because a lot of times I get lost and I forget that I even asked it to do something. But it can hook it into our in-app notifications. So if you're in one tab or whatever, it can show you like other other tab needs or attention.
If you're if you're really running a ton of these. Yeah. We're trying to we're trying I think making the whole thing hackable. It's like a thing I'm very into right now of like how do you. And it's like not just the cool thing about making an app hackable at this point is that it's hackable not just for like a human. It's hackable for the agent, which I think is really fun. And so it's like well, how do you let the agent out of do it? You just publish a skill and then all of a sudden any agent has access to that skill knows how to like hack into all these APIs in warp or whatever to like do cool shit.
Are you all using skills? Are you liking skills? We're going to skills. I was just going to ask related to skills like. Yeah. Are there plans to like learn skills on the fly? Like how open claw can be like make up a skill it's going to do this. Instead of like really actually caring about what the content of the skill is and handwriting it just being like I want to do this. I want you to even think of what skills you might need and just like write them all to random files for me. I feel like that's a missing piece maybe. Yeah. I mean, I've never had like a hundred percent of time if I'm making a skill I have the agent to it.
But the key thing that we're missing and maybe the other coding agents and maybe open claws doing a very cool job of is like. It's memory really is like I think what it is. It's like oh you learned how to do something. Record this and have an indexed way. So the next time you do something similar that's possible. And so a million percent we're leaning into that. The again the thing we're trying to do with Oz which is. It's for individual developers but it's also for companies. Is that anytime an agent runs. It records a hundred percent of what it did. And so that is like available. Whereas I feel like right now most people are running this shit locally.
It's like you're you're having something into your file system. You're generating some markdown files someplace or whatever. But we. My concept on this is like auto tracking of these things. So there's like an automatic repository that's being built up. Of the agent runs so that you can do memory on them. Which is yeah we're working on all this. Like there's just going to be some crazy high of mind thing. Like I've been thinking it's not about this as well. Like what is actually what do you actually need. What's the ideal setup for an agent to work at like on like big things across a company. Is this is maybe half baked but I think it's just like.
Some gigantic distributed file system where there's like. Some way of like agents because I think these things do work well on files. It's just they don't you don't want the files on the local machines. And so you know Google we used to have this thing like. GFS where it's just like global distributed file system. And so I think it that's going to be like the core. Data protocol for really enabling agents is that every organization should have something like that. And I'm just now I'm just like. I love it. It reminded me of there's some people that they spun up. They wanted to make an app so say you want to make an app.
Yeah. How do you how do you find all the features that people want? Well you spit up 500 moms as agents. Yeah. Give them a persona. And then you have them use the app. It's incredible. Yeah. It's great. So yeah you are brilliant. It's it's semi brilliant. I mean some of it's cool. That's a pretty cool thought experiment. Although agents aren't. I would have known. I would have known. But still yeah they're like testing their apps across all of these simulated people. To get to the core of what a good app would be. So yeah you don't ship V1. You ship V1 to a shitload of agents. It's like that's your app.
That's your beta testers. That's your beta testers. And you loop in there until everybody starts reporting good reports about your app. And then you think maybe at that point if you go to real people it will be better. And this is just uncharted territory. It can't be worse. Maybe. Yeah. Yeah. We need a little bit better computer use for that to work for work. Like I don't know have you all tried to play with any of the computer use models. No. No. So like you know so like the model providers have these beta models was to computer use where you can it basically is like taking. It's like looking at your screen from a computer risen standpoint.
Oh I did this at Google. Yeah I used to Jim and I doing this a couple of times. And it can like click and zoom. And when that gets good we will literally just be like. Have these things constantly testing our app. Because like our app has all sorts of edge cases and bugs or whatever that. But you're going to get feedback loops that are fully agent driven. I don't know this is it's a crit like I keep saying. It's crazy. I was learning how to use a screen flow or whatever. I just I spun it up and I was like hey I need to do this one task in screen flow. And it was watching my screen. I was like yeah click over here and do this.
And I was like holy shit that's right. I would have normally gone to Google and read a lot. But you watched my screen and told me I was like this is well. I also watched Japanese CSS videos using that computer vision technology. Because it would live translate what it heard and saw. And so I didn't have to rely on crappy captions. I got a genetic translation which was really cool. That's cool. That's super cool. The world. I've mostly find it exciting in like maybe 20% fine. It's scary. But I mostly find it exciting. That's how we are in the show. We are constantly going oh that's really freaky.
But also I need to stay employed. And some of it's fun. So we're all toiling with how this pulls on our heartstrings and our brains and our wallets. And as well. Yeah. Yeah. Every other episode is like I hate this. I love this. I hate this. I hate this. Yeah. Yeah. It's unavoidable. It's inevitable at this point. But yeah. I mean I've been very pleased with it. Unless it unlocks so much stuff that I would have never had the time to bother with. Like we're doing a conversion. Like we were trying to update ember data so we could use like the latest warp drive stuff instead of like this model fragments thing we were using.
And one of my co-workers just like gave Claude like access to it for the weekend. He was like hey make this work with the latest ember data version. It's like cool. And it did it. And like the maintainer of ember data who I'm good friends with was like yeah you probably shouldn't do that. That would take like months and months like whatever and it did it. So like man it's it's a crazy time. Yeah. Yeah it is crazy. I have a silly question for you. Sure. How do you say CLI? Do you say Klee or CLI? I say CLI. Okay. Yeah. Or CLI. People are saying Klee these days. I'm like Klee. Ew. I've never never in my life what I say.
Do you say to me or to you? I do say gooey and to me. Yeah. I say to you. Yeah. Yeah. I think I usually. Some like the ducks. What were the ducks that were like that? Huey Dewey and. Huey Dewey and. Yeah. Oh. Okay. Okay. Okay. Yeah. It sounds like we need to turn those into agents. Yeah. So. I'm looking at all this AI stuff. Doing it. Yeah. Fun. Other than working in AI. What are your hobbies? So I have like a two year old son. That takes that takes a lot. Yeah. You guys have kids? Yeah. Seven and nine. Yeah. Yeah. Almost four year old and then I have nine month old twins. That's crazy. I can't even imagine that.
In the thick of that. Yeah. So I have a two year old son who, you know, I'm not working. I'm working really hard. Just to be clear. It's probably the hardest I've ever worked my life because the space that we're in is moving so fast and it's so competitive. Yeah. So I am working a lot. There's stuff I like doing. So I live in New Mexico near a ski area. It's my wife's ski instructor. And so try to go ski, although it's been a pretty bad year for snow, unfortunately. Yeah. Same in Seattle. It's been terrible. I play guitar. So I see that your musician. Music is greater than all of this stuff. Yeah.
I play golf sometimes. I like to read. Not that. I don't have any like crazy, crazy interesting hobbies here. I'm a football fan. I like the sports bet. I don't know if you all like that. Nice. Yeah. Did you watch Super Bowl? I did. I did. Yeah. Did you guys? Yeah. Yeah. Yeah. Unfortunately bet. Like all of my bets were around like people getting touchdowns and then like everyone was like, you know what? We're not going to get touchdowns in this game. We're not going to kick it. Yeah. Yeah. Yeah. My wife's from Boston. So we, we, we, we, like, I don't know. The Patriots got their asses kicked.
That wasn't that fun to watch. Seats are good. Yeah. Yeah. That's it though. I don't know. I'm working. Yeah. But I was working. It's enough. It's super. It's. It's really fun. I'm really competitive. And so to me, it's like a. It's like a big. Strategy game or something where I think we are like by default under dogs because we're start up in like. Anthropic has a $1,000,000 and Google has. 10 times that and open I has a lot of money. And so it's, it's like a super duper. Engaging fun. Strategy game. And I like. I really love building stuff. I love like using the thing that we're building. So I don't know.
I'm probably working on unhealthy amount these days, but I'm just like constantly like I'm just I'm beyond. It's like that or my kid and like sometimes the kid doesn't sleep and that makes everything hard. And so I don't know. But I'm really. I'm really trying to make more successful and it's so fun to work. Yeah. I don't think to be said for like, I feel like anthropic did so well because everyone kind of hated open AI. They're like, you're the big, you know, giant. Oh, yeah, we're going to be this like goodwill like nonprofit, whatever then oh, just kidding. And like, I just, I don't know, just a lot of weird moves and I feel like anthropic
is starting to behave that way. Like they're starting to be like, you know what, we have money now like, no, you can't use our key on something else or like, whatever. So I feel like there's garden. Yeah. There's a room for someone else to come in again and be like, you know what, we're the ones that like you should probably use because, you know, they're not treating everyone well. I mean, I have to be careful how I talk about all this stuff because we were able to see with anthropic and open AI and Google. But I do feel like there's some advantage to not being one of those and having and giving
developers choice for sure. But I also, I mean, I've said this for a while, like, I think at some point I'm not convinced you're going to need to be like the frontier model for like forecoding to get like really good results. And at that point, like the mode of those companies, it can't just be like that they have really good models. And this is why they're moving so aggressively into giving everyone apps and trying to like a lot. They're really trying to lock everyone into their apps right now. And I don't know. It's like, I feel like if there was, this is really hard for work like, you know, our,
we have to pay API prices and they give, you know, people, AI tokens for way cheaper on these subscriptions that everyone has. And I do feel like if it were a level playing field in terms of just what product you like more, I mean, I feel like our product is like right there if not better. And so it's a, it's a lot of cost competition, I feel like our product is better. Like I, like it, but it's a lot of cost competition that's, that's tough for us. And like, would I be happy if there was an open source model that was like at the quality of opus or the latest GPT, like acquisitions kind of a dope exit though, right?
I mean, you've got the coolest best, in my opinion, the one that's pretty far ahead of everybody else. Why don't they just buy their way into the future? And so just consume you then the competition's gone and you win. I don't know if that's what you want though. So I'm just kind of like, it's not what I, it's honestly, it's like, like I'm like pragmatic. Thank you. By the way, thank you for that compliment. I do feel like our thing is like the, we've been ahead on a bunch of stuff. So it's like, like I said earlier, people, yeah. Thank you. I don't know, I want us to, there's nothing I would rather do than continue to like build
this product and so going, I've worked at very big company, like, you know, like you started the episode, I was leading a big project at Google and Google is an awesome place to work. But to me, it's so much more fun and interesting to try and, you know, not have all that bureaucracy to like, way through and just like be able to build shit every day. And, you know, we're like a 50 person company and we're doing well just to be clear. It's like, like people really love the thing that we're building, which is amazing and like we're growing really fast and stuff. But yeah, I'm not at all actually aiming for like being acquired.
I would much other build this shit for a long time, if I'm not enough, I'm able to, yeah. Cool. Well, you guys like warp. That really makes me happy here. Yeah. We love warp. Yeah. I mean, that's it. It's like, if you don't even like the AI stuff warp is really good. So it's like, we're not. I don't even want to force that on to people. So, but I do think we have really, really amazing AI shit and you should try, try building like some automations and Oz and try like programming agents through it. Try like, you can even use Oz to spin up whatever if you want to use clock codes or whatever. You can use it to like run 20 clock codes in the cloud.
It's like, you're going to have half of them run codex and half of them run clock code. And so, you know, I would love any, any feedback because I do think this is the future. It's just like, let these agents run more agents and let them like do smart shit and don't do it necessarily on your laptop and let it all be programmable. Yeah. Yeah. Yeah. It's a crazy world. I feel like the biggest barrier for me is I don't get to use it at work. So I'm having trouble with it. Is that like a security thing? Probably. I don't know. I'm a little bit scared always to try to like, introduce new tools because yeah, the security
thing and just like IBM is very like, hey, you have to use IBM stuff like they're making their own VS code AI fork just like everyone. Oh wow. Yeah. IBM has reached out to talk to like interesting. I don't, that is one thing that I do not think is the future, by the way. It's like. Well, like yes. 74. 74. Yes. Look, that's a crazy strategy. Yeah. And Adam, are you not allowed to use it? Shopify folks asked for it and it was declined after a security inspection and I don't know why. That was also a year ago or so. And so we could ask for. We might try again, you know, Toby is a small angel investor in war, but I don't think
he's got so many other things on his mind that he's writing in agents all day. That dude is green fielding himself into apps. He must have said he is an animal. It's awesome. So I don't know what his terminal set up. He's always got multiple pains. I think he's running tmux somewhere, but I'm like, oh, fucking, why not warp at this point? Yeah. I hear the word tmux. I'm like, yeah, we're probably not winning that person over, but it's like, no, it's just, it's like we, and to be clear, we want us, we're going to, we will probably invest, almost definitely invest in doing a, like tmux control mode thing in war, but I'm also
kind of just like tmux is such a crazy thing. It's like, oh, you want to run a terminal within your terminal? Okay. Run tmux or, or you could use a thing where you don't have to do that. But some people really, they like, they like, I get it. It's like hackable and it's flexible and it's no shade of tmux. We're going to, we're going to support tmux this year in a better way and work. But those do, those do tend to be the harder people for us to get into work because it's like, like the most, the people who've customized their terminal the most have the most friction moving to a thing. Did it work?
Yeah. If you're on arch Linux using Neo of M, might, maybe not for you. I mean, we have, like we do weirdly, we have those people and like, but it's, that's not, it's not really our bread and butter. No, I think the target audience is the people that want that, but can't, and don't have the time to figure it out. I mean, that's correct. Half my list here is like, no, no fuss, multi pain, no fuss, power line, you know, no, it's all no fuss. You basically get all the things that people get in a turbo powered Z shell mixed with tmux, mixed with a very rich IDE, and that's it. And it's you downloaded it and it's ready.
So like if you love config hell, enjoy your own, some people do those, some people do some people do it. I mean, I got to put that quote, I put that quote on our website. If you love config hell, work, not for you. Seriously, it's like, it's all the people that wish they were the hacker that had the tmux set up, but just doesn't have the time and so download work and you're done. Thank you. Yeah. I agree. Yeah. I could not have said it better myself. We are at time. Anything we miss? You want to cover it before we're done, Zach? No, this was awesome. I really appreciate you guys having me on. This was fun to do.
It's like, it's like a more, it's a more fun than normal podcasts for me. I do a number of podcasts. We're not a family show, you know, not a family show, which is great, you know, like I don't know why all these dev shows want to be a family show, too. Well, come on. It's depths. I like this one. And I've done, one other one I really liked was 20 VC. You're listening to that. It's a hairy set. I don't know. Yeah. Yeah. He's pretty fun, too. So I like being on the podcast where I can like drink something and have some fun. This was great. I appreciate you guys having me on. Yeah. Yeah. Thanks for coming on.
Thanks everyone for listening. If you liked it, please subscribe. Leave us some rings or reviews. We appreciate it. We will catch you next time. You've been watching Whiskey Web and whatnot recorded in front of a live studio audience. What the fuck are you talking about, Chuck? Like the show? Subscribe. You know people don't pay attention to these, right? Head to whiskey.fun for merch and to join our discovered server. I'm serious. It's like 2% of people who actually click these links. And don't forget to leave us a 5 star review and tell your friends about the show. Alright, dude. I'm out of here.
Still got it.