Transcript 0:00 But no shitty software, please. Development right now is switching from development by hand to development by prompt. For pro developers, like, vibe coding is kind of the fastest way to ship crap to production. 0:10 Holy shit, I can just, like, ask the computer [chuckles] to start making my change. It's gonna be wild. [gentle music] Hi, Zach, and welcome to The Influence, uh, Show by Turing Post. I'm very glad to have you here. 0:28 Yeah, I'm excited to be here. Thanks for having me as a guest. I've been actually playing with Warp, and it's mind-blowing how it, it builds applications just doing everything by itself. 0:38 And I wanted to ask you, you call it agentic development environment. If you can tell what it is in relation to terminal, to CLI, to IDE, w-what is agentic development environment? 0:52 Yeah, so the, the basic idea is that development right now is switching from development by hand to development by prompt. And so what I mean by that is, like, you know, in the... 1:02 Most of my career, if I wanted to build something, I'd open up IDE, and I'd start typing some code, or I'd then open the terminal and type some commands. 1:09 And what I do today is I open Warp, and I type in English or natural language, or I literally just talk to Warp and ask the computer to start doing something for me. 1:19 And the idea is that as the developer workflow shifts to be more agent first, that what you really want is a workbench that is built to support that workflow. Mm-hmm. And so, like, what should that look like? 1:33 That's, like, what we're, you know, figuring out at Warp. I think it looks a lot more like a terminal than it does like an IDE because what you want is, you know, you want, like, this direct input. 1:44 And, you know, the terminal, the way that it's always been set up is, like, you type commands, you get answers. Today what you do is you speak English, and it launches agents. 1:52 And so Warp began as a, as a terminal, but now the primary way of interfacing with it is through natural language. It no longer makes sense to just, like, call it a terminal. It does much more than a terminal. 2:04 It, like, will write your code. It will give you the ability to, like, pair and partner with agents to improve the code by looking at the diffs. Mm-hmm. It gives you some file editing capabilities. 2:14 So some things that are typically in the IDE are now in Warp. And so we just thought, like, "Hey, none of the existing labels for these tools actually fit." 2:23 And, like, when you sort of, like, stick to the existing categories, you run the risk of not building the thing that's actually the best for the new workflow. So that's, that's how we're thinking about it. 2:33 Did you come up with this term? I did, [chuckles] yeah. I did come up with it. We, we talked about a lot of different ways of naming it, but we were like, "You know what? It's similar to the IDE." 2:40 We liked how it kind of sounded like the IDE, but switch the I with an A and call it an agentic development environment. And I think it's catching on. 2:50 Like, I've heard other people now say that that's what the world needs, and that's how developments can be done, which is pretty cool. So it'd be fun- Yeah... to have actually, like, coined a thing here. We'll see. 2:59 I was actually w-going to ask 'cause, uh, other companies already call themself IDE. I think that we're the only, we're the only ones I've seen where it's, like, the agent. Like, we, we don't even call ourselves an ADE. 3:10 Like, right now if you go to our homepage, we call it the agentic development environment. So it's a little bit of- Hmm... like, a kind of proper noun at this point. I don't know. I was just watching a, um... 3:19 It was another, like, podcast live stream with a couple other engineers who I respect, and at the end of it, one of them, I think it was the Zed founder, Nathan, was like, "You know, what the world needs is, like, an agentic development environment." 3:31 So I do think it's- [laughs]... I do think it's catching. [chuckles] They're just logical after IDE. If you think about the, the current environment with coding agents, how Warp stands out? 3:42 Like, what is actually defines you and differs you from, let's say, Cursor, which is super popular, and, uh, I don't know, Cloud Code, which everyone uses? 3:52 So compared to, like, um, IDEs like Cursor, Cursor is the, the most popular. I guess there's two things. 3:58 So one is, like, because we're more at the terminal level, it's, like, a wider range of tasks that you would want to do with Warp. Yeah. 4:06 So, you know, Warp is really good for, like, the full software development life cycle, so from, like, setting up projects to running a coding agent to then deploying the code that's been built by the agent to then, like, debugging production issues. 4:19 You wouldn't really wanna do that from the IDE in general- Mm-hmm... 'cause then you, you know, it kind of looks like Microsoft Word. It's, like, really a file editing first interface. 4:27 And those are all tasks, like, that you would wanna do from something that looks a little bit more like a terminal. 4:32 Warps, you know, I would say also, like, Cursor, the, like, real sweet spot for Cursor, the thing that they really nailed, which is very, very valuable, is, like, auto-completion, so, like, AI-based- Mm-hmm... 4:43 auto-completion. So you type. You see the ghost text. I think Warp's agent, like, the actual, like, agent experience in Warp is superior to the agent experience in an IDE right now. 4:52 Like, the agent experience in the IDE tends to be, like, via a chat panel, and because it's, like, you know, a UI that's off to the side, there's, like, a weird... 5:01 It just is not as good of an experience, and the actual quality of the agent in Warp is better. Like, I don't think Cursor publishes the benchmarks of their agents, but Warp is- Mm-hmm... 5:08 like, the number one agent on Terminal Bench. It's a top five agent on SuiteBench. So I know for a fact that the quality of Warp's agent is, like, validated against, you know, the top benchmarks. 5:18 For Cloud Code, it's more interesting. So Cloud Code works quite similarly to Warp in the sense that it's based- Mm-hmm... 5:24 in the terminal, and the primary thing that you do in Cloud Code is, like, you prompt it, and then the agent goes and does something for you. I think that there's a couple advantages that Warp has compared to it. 5:35 So one is, like, Warp is multi-model. So, like, if you use Cloud Code or you use Gemini CLI- Right... 5:41 or Codex CLI, it's like you're locked into, like, using one provider's models, whereas with Warp, sometimes we'll use a mixture of models. We'll always give you, like, the most state-of-the-art model. 5:49 The other thing is, like- These CLI based tools have no ability for, like, file editing. And so it's not easy to, like, view the diff or, like, really tightly iterate as they create code. 6:02 And so they tend to be, like, awesome vibe coding tools is, like, what I would say. Mm-hmm. 6:06 So, like, you know, you can run Claude, and people who are using Claude very heavily might be, like, running, like, five Claudes, building different, like, parts of their, like, you know, marketing site or whatever. 6:16 But if you're, like, a pro developer, the way to have more success using these agentic tools is to, like, carefully, like, pair with the agent and see at every step what the agent's doing and have the tools to guide the agent by, like, literally, like, doing a code review of the agent's code and redirecting it and, like, seeing its diffs, making sure it's sticking to the plan. 6:37 And Warp is just a much better interface for that because- Mm-hmm... 6:41 we have the full GUI experience, whereas if you're, like, you know, Claude code, you're limited to just text-based output of a terminal app, of a CLI app. And so Warp is really trying to build out this, like... 6:51 It's not that we're against vibe coding. It's just that for pro developers, like, vibe coding is kind of the fastest way to ship crap to production. Right. You vibe code something, and then you forget about it. Yeah. 7:01 Or if you, like, look at the Stack Overflow data that, that, from the most recent survey, it's like the number one and two complaints of developers using these agentic tools is, like, the tools, like, generate a lot of code that have subtle bugs. 7:13 The code is hard to comprehend. It takes sometimes more time to, like, understand what the agent is doing than if you had written it yourself. 7:21 And so, you know, Warp, our view is, like, build something that is very, very agentic but also, like, can be used for real production code bases. 7:30 And to make that work well, it's all about, like, can you comprehend what the agent's doing? Can you guide the agent carefully? Can you correct the agent when it does the wrong thing? Mm-hmm. 7:39 Can you make the agent sort of, like, work according your team's, like, coding conventions? 7:44 And so I think we just have this intrinsic advantage when it comes to, like, tight pairing with the agent that these other CLI based- Mm-hmm... tools don't have. 7:51 When I was playing with Warp, I thought about it as a meta tool. We recently did this, um, report with the Focus AI about we just tested different coding agents. Yeah. 8:01 And, uh, Warp came out so strong that my husband, who was running the testing, he asked me if you can hire him as a, an evangelist [laughs] so. That's awesome. 8:10 Yeah, but his main resume was the Cursor plus and Warp as, like, the best combination together. Yeah. So I think at this moment it's more like a combination of tools. 8:20 For me, it felt a lot like meta tool because I'm more of a amateur builder and developer and not a professional developer. So for me, Warp helps do all the steps that I might not know. 8:32 Um, and my question is about the vision. How do you see Warp in, like, development? Is it along the other tools, or is it's, it will become an actual, like, meta tool for professional developers as well? 8:44 How do you see this development? Yeah. So I think today we see people using it in all sorts of different ways. Like, it's very common to use Warp plus an IDE. 8:54 And so you could have Warp do all the things I mentioned before, like the project set up, the, like, you're having dependency issues. You wanna do all your Git stuff through natural language. 9:04 You wanna do all your Docker work or Terraform, like, all these things that are just unpleasant to do in an IDE, you could just do in Warp, and then if you want- Mm-hmm... 9:12 you can go into the IDE and, like, do the coding more through that. That's totally fine. People use Warp that way. Increasingly, I don't-- that's not the vision. 9:20 The vision is more, like, you can just, like, use Warp as your single workbench. 9:24 You know, the reason I, I think that that's, like, kinda, like, where things are gonna go is because you're gonna have less and less need over time to, like, do a ton of, like, auto-completion and hand file editing. 9:35 Mm-hmm. 9:35 And really, the type of file editing that you're gonna need to do is, like, it's all about, like, adjusting the path of the agent and giving feedback to an agent as the agent's working, and w- that's, like, what we're building for in Warp. 9:45 But it's totally very common for people to use, like, Warp plus Cursor these days, and so that's a, a cool workflow. 9:50 We also have the other workflow of, like, people use Claude code inside Warp, or use, like, a CL- and that works too. 9:56 You know, developers, you could say one thing about them, it's like they like picking their tools- Right... and they're best at optimizing their own workflow. 10:05 And so what some people will do with Claude code is very similar. 10:07 You know, they'll let Warp do all of, like, stuff that they think of as, like, terminal-ish stuff, and then they'll, like, run a Claude code session to do the coding. There's no need to do that. 10:15 Like, if you wanna pay for two subscriptions or whatever, like you wanna pay for Claude and pay for Warp, you can do that. 10:21 But, you know, our vision is that there's a lot of value in having a single integrated agent experience. 10:26 For instance, if you do the coding directly through Warp rather than doing it, like, by running a CLI app, Warp can, like, share that context across all of your terminal panes and, like, you can, like, when you're running a server in one pane, use the output of that server as input for an agent that's running in another pane. 10:43 And so there's all this awesome value that you get from, like, an integrated agent experience that we think people, you know, will eventually... 10:51 A lot of people are doing it this way, but, like, for people who prefer using multiple tools, we think they'll eventually be like, "Hey, it is actually easier to not do the context switching." Mm-hmm. 10:58 But it's totally up to developers, and developers are smart and, like, wanna pick the- It's like they know what's gonna make them most productive. It's a lot of rethinking of workflows for developers- Yes... right now. 11:08 And what do you think would disappear from today's workflow once that type of meta tools exist? I think that the existing tools will just become less used. 11:18 Like, I think you're always, or like for a long time, you're gonna wanna be able to use like a file editor because there will just be cases where it's either faster for you to do it yourself, or you started with an agent, it's not doing it right, and you want, or you just like wanna get in there and like change it, or it's like you're working in a language that the agent isn't good at. 11:35 So there's all sorts of reasons that you'll wanna do that. For the same reason that, you know, sometimes as a, very rarely, sometimes a developer, you like need to look at like the machine code or something. 11:45 So it's like it will be there. There will be an abstraction that's more convenient to use than the existing tools. Mm-hmm. And you'll use that for like 95 or 98% of the time. 11:55 But when you have to, like, get closer to the metal, so to speak, you'll wanna go into the code editor, and like I think that tool will exist. 12:01 So I think it's more like the predominant way of working is going to switch to be a- agent first. But like when you'll have these fallbacks for when you wanna go in, like, you know, hand tweak stuff. That's my guess. 12:15 I would say now Warp can be a junior developer already. Yes. You're saying that it will be agent first, so does it mean that this tool can be like, um, senior developer or completely independent developer? 12:27 How do you see it? Junior developer analogy is pretty good actually, in that if you give Warp a, like, spec of things it should do, or you iterate with it on a plan- Mm-hmm... 12:37 and you're like, "Go build this," it can build it at the same quality as a junior developer, sometimes even better. It's like it's pretty crazy. 12:45 And not only can it build things, it can, like, generally solve problems, which I think is like an important distinction. So it's not just a coding tool. 12:52 It's a, like, solve my dependencies tool or, like, figure out my server's crashing tool. 12:56 It can do a lot of the different tasks that a developer can do, but it needs to be guided right now by a more experienced developer. 13:04 The way I see it evolving is, like, the actual types of tasks that it will be able to do will become more and more complex over time. Yeah. 13:11 And it will be able to do things that, like, good senior developers can do, and this will be driven by improvement at the model layer, but also improvement at the app layer, improvement in terms of, like, the context that's being provided, improvement to how, like, the tool is, like, learning your team's knowledge and your organization's knowledge- Mm-hmm... 13:29 so that it's not having to, like, relearn everything from scratch every time it starts. 13:33 But I still think over the next few years at least, you will have senior engineers who fully embrace this way of working, who find that, like, they can do, like, five to 10X the amount of stuff. Right. 13:44 And a lot of, like, the job becomes how do I effectively communicate what needs to be done and how it should be built- Mm-hmm... 13:52 so that an agent can do most of, like, sort of brunt work of doing it, and then I will review it and make sure it's right and, like, isn't- Mm-hmm... 14:01 gonna introduce bugs or security holes, and it fits into our overall architecture and the product roadmap and all that stuff. So I think that the job will evolve so that it's more like that. 14:09 But a junior developer right now, I think the thing you need to do if you wanna stay in software engineering is become a senior developer as fast as you can and learn how to leverage these tools. Yeah. 14:19 How can these tools help you with that? Yeah, so I think that that's actually, like, an overlooked thing. You know, there's a lot of, like, fear around the tools replacing. 14:28 But the tools can also be, like, a huge accelerant to learning. They can be tools for that. They're incredible learning tools if you use them that way. 14:37 And so something like Warp is awesome, not just doing stuff, but explaining. It can be a very, very powerful shortcut to, like, learning a new code base, for instance. Or it can- Mm-hmm... 14:48 walk you through code line by line, explain, like, what's it doing, why is it working this way. 14:53 Or it can suggest, you know, a cool way to use it is to have it suggest multiple ways of potentially building something and, like, give you the pros and cons of going with each approach. 15:02 It's very similar to, like, ChatGPT in the sense of you can use ChatGPT to, like, cheat on all your tests and write all your essays if you want- Mm-hmm... and, like, that's a choice that you're making. 15:11 Or you can use ChatGPT as a way to, like, accelerate your learning for some topic. It's a little bit up to the intent of the person who's using it, what they wanna get out of it. 15:22 And so it's, like, a phenomenal learning tool if that's, like, your intent going in. I have two questions about languages. 15:28 You chose Rust for Warp's foundation, and so what does Rust enable here that wouldn't have been possible otherwise? We chose Rust because we wanted to build an app that was like, felt fast, honestly, first of all. 15:41 Like, developers, I think, are a uniquely sensitive group of people to the performance of their apps. Yeah. 15:49 They're also sensitive and interested in how apps are built, which is just not a thing for any other consumer segment. Mm-hmm. I don't think anyone who's using Salesforce is like, "Is this built using React?" 15:58 or whatever. But, like, developers actually care about this. They, uh, they're curious about it. So we chose Rust. The primary thing was performance. 16:05 Secondary thing was, or probably of equal importance, was safety and just, like, not having a lot of crashes. I really strongly prefer, you know, strongly typed languages. 16:14 Rust makes it a little bit harder to write a slow app when it comes to, like, just, like, the way the memory management works. 16:20 And then the other great thing about Rust is, like, if you wanna write a cross-platform desktop app, which is what Warp is, it's a nat- fully native app on Mac- Mm... 16:29 Windows, Linux, and also wanna support the web, I'd say you really have two choices today. 16:34 You can use the web platform and do it in Electron app, which is what most people do, or you can use Rust and, like, compile it natively for all these other platforms. 16:44 Rust is definitely the harder path, and I don't know that I would rec- Like, I would only do it if you have an app that has, like, these types of performance characteristics where it matters, and, like, it does matter for Warp. 16:54 Warp is streaming through gigabytes of data, of textual data, and, like, it's very annoying if it all of a sudden starts to, like, stutter or, like, there's all these problems with it. 17:03 Rust gives you really, really nice, like, primitives for doing asynchronous work. 17:07 We just ended up thinking, like, for Warp, it would be better to go one level lower and use something that's closer to the metal- Mm-hmm... than using Electron. 17:15 And we did prototype Warp in Electron first, and we started to run into some performance issues pretty quickly, and we're just like, "You know what? 17:21 Let's, let's just pick the, even if it's a little bit harder to build, let's pick the, the thing that we think is gonna result in a better product." 17:27 Now when you work so close with AI, and AI is famously not built on, but machine learning is famously u- using Python all the time, how do you feel about it? 17:37 And do you think that AD era shift parts of the stack forward Rust or, like, newer contenders like Mojo, other languages, or it will be still, like, Python? So great question. 17:49 So most of our agent, like, logic Is on our server, which is not built in Rust. The server is built in Go. Mm. Again- Wow... sort of the same question applies, like why not- Yeah... Python? 18:01 If we had to do it again, actually, and we talk about this internally, we would probably use Rust on the server as well, 'cause there's some code that we wanna be able to share between client and server. Mm. Go is okay. 18:12 I'm not nearly as big of a fan of Go as I am of, of Rust, if I'm being honest. Go is kind of like a no man's land. I don't think we're getting that much from it compared to either using Rust or using Python. 18:21 Like Python would be nicer, I think, for doing sort of like, you know, the more AI-oriented work. But the truth is, you can build anything in any language. 18:30 We're able to build and iterate fast enough on the server side with Go that it hasn't been an issue. Why do you think they stuck with Python machine learning? 18:37 I think it's the main language 'cause it's very, very easy to use, and it's like very ergonomic. 18:42 Also, some of like, for whatever reason, the core data libraries, like the stuff that's done at the data science level, like the, the Python libraries have caught on for that. 18:52 And also just like it's a very interactive, like dynamic language so that you can- Mm... you know, I think a lot of AI work starts in a notebook. It starts in like a, you know, in Jupyter. 19:03 And so you want something that's, that's not compiled, that's like super interactive, that is very high level also. So it's just like it's not hiding any of the concepts. 19:14 It's just is like making the concepts of doing the machine learning front and center. 19:18 So I think for those reasons it's probably, it's a, I think it makes sense to me that it's like the primary place where all this ML and LLM work starts. 19:26 Now when you not just a software in software development anymore, you're in agentic development world. Yep. What excites you the most, and what are your concerns about the future that you helped build? 19:39 The thing that excites me the most, and I've been a developer for a really, really long time, I don't even wanna go into how long, but the thing that excites me the most is just like how much easier it is to build awesome software- Mm... 19:53 and how much more is possible to be built by like fewer people. And also how it expands the sort of like number of people who can build software. 20:03 So it makes it, it gets rid of a lot of like the gatekeeping around like who gets to build software and apps. 20:10 I think that historically that's been a, you know, a real problem in that there's like relatively few people in our society who have like, who know how to build software. Mm-hmm. 20:18 And software is such an important thing in the world. So it definitely democratizes that. It lowers the barrier. 20:24 You know, we have people on our team who are not software developers who are making apps to help themselves do their own jobs better. 20:29 Our head engineering recruiter has built an MCP server that connects to our applicant tracking system, and he's not a developer, and it's so cool that he now has this ability to like make his own job so much easier. 20:43 So that's like what excites me. What concerns me is that there's potential to be a lot of like really crappy software that's built and- Mm... brought into the world. 20:53 [laughs] Not just a builder of software, but a, a user of software obviously. I like hate using shitty apps. [laughs] It's like a real... I hate it. I hate when I go to like, you know, my pharmacy's website- Right... 21:05 and like I just wanna refill a prescription and like it doesn't work. I can't log in. It doesn't find my stuff. [laughs] The buttons don't work. 21:12 And so I'm just afraid as a user that this is going to lead to an explosion of like really shitty software- Yeah... that I have to endure using. 21:22 And it's one of the things I fundamentally don't get actually about like vibe coding is like it's awesome that everyone can build software, but like what matters is actually not just like the pure volume of software that goes out in the world, it's like does this software solve problems for people? 21:36 Yeah. Does it like solve them in like an kinda like useful and like fun to use way? 21:42 And my worry is that there's gonna be just a lot of bad, shitty software shipped, and that even when you look at the more pro use cases for the agentic stuff, you know, my worry is that it's like going to lead to like security holes or like unmaintainable code or there's good reasons that the practice of software engineering has evolved the way that it has. 22:03 That's why I think that in order for this stuff to really work in production, you need to go way beyond vibe coding to building- Mm... 22:11 a tool set that like lets you really, really carefully specify what you build, wanna build. 22:16 But not just that, that lets you really, really have a very, very short tight feedback loop and make sure that you understand what's being built and like are guiding the agents to do it the right way. 22:26 Like, I think that's where we're at today, and so my fear would be there's a, just a, gonna be a lot of crappy software. [laughs] You need to change system prompt in the world, like don't build shitty software. 22:36 Just don't do it. Yeah, like that, like if we could just tell it that, maybe that would just solve it. No shitty software, please. It's like a kinda like- Yeah. [laughs]... savant. 22:44 It's like it, it will do things where, you know, if you just give it a goal, you gotta be really careful. It'll just start like deleting tests and be like, "Oh, the tests pass now." It's like, "Well, no, not really." 22:55 And so it doesn't understand- Right... the user problem that it's solving, which is to me like absolute key thing if you wanna even be a good engineer. 23:03 It's like a person, it's like you always have to be like, "Why am I building what I'm building? How is it useful to someone?" Mm-hmm. 23:08 And until the agent can understand the user problem, I think it's like bound to produce stuff that's like is not like that awesome, and you need a human to be providing that context. Right. Human in the loop. 23:19 Do you think at all about AGI and super intelligence? I do. I'm not like, uh, qualified to like pr- prognosticate on like when it's coming. I don't know. 23:27 All I'll say is like this stuff is changing super-duper quickly, and like where we're at now compared to six months ago implies a sort of- Right... exponential curve. 23:37 I don't have like enough sort of knowledge to say like how long will that exponential continue? When will it become more of an S curve? 23:44 I do think the typical pattern is like you'll have some period of exponential and it will become an S curve. It could be that for AI though, the, it becomes sort of recursively self-improving. 23:54 I think this is like what the people who are on like the ASI side- Right... think is gonna happen. I don't know. It's not there yet, but the rate of progress is obviously super fast. 24:03 What I heard from you is that human still is needed, is needed in this loop. The bet that we're making, I guess, in relation to this is like we're betting that the models will improve and that they will improve rapidly. 24:14 I do believe that there will, these agents will become more and more autonomous and be able to do harder and harder tasks, and that you need to build, you know, we're building more- For that future. 24:22 However, if you look at, like, where we're at in terms of the adoption of the agentic workflow and the current state of the models, the first big AI breakthrough was autocomplete with, like, Cursor. 24:32 I think we're starting to see the widespread adoption of agents, but, like, most developers aren't using agents yet. Like, that's just, like, a fact. It's a foreign workflow that developers need to adopt. 24:42 And then I think that the third thing that will happen is that the agents will become more and more autonomous, meaning, like- Mm-hmm... you might have, like, system triggers. "Hey, there was a crash." 24:52 An agent then goes and investigates. And so I think that's coming. 24:55 What's needed today is, like, this tight iteration loop, and even in the, like, more autonomous future, what's needed is, like, maybe it's not an iteration loop, but it's, like, an orchestration loop. 25:05 It's some way of- Mm-hmm... expressing intent and managing what the agents are doing even if, as you have more of these agents at scale. 25:11 So in terms of timeframe for that, again, I think this stuff's happening very quick. Like, six to 12 months, I think it's gonna look pretty different. 25:18 But today, as people are just, like, getting used to building with agents, I do think you need that really, really tight loop in order to ship production quality code. 25:26 It's very hard to make predictions, but if we talk about adoption and some timeframe for that for software developers- Yep... what do you think will happen in the next, let's say, 12 months? 25:36 Where I think we're at right now is, like, the mainstream developer has adopted an autocomplete tool like Cursor. 25:43 In terms of agentic tooling adoption, what you see is, like, AI enthusiasts plus non-technical people, like PMs, designers, and an increasing number, but still relatively small number of, like, kind of, like, bread and butter pro developers being like, "I'm gonna start every task with a prompt." 25:59 I bet you within, like, six to 12 months, that the agentic workflow is primary workflow, and the autocomplete workflow is no longer the primary workflow. I'd put a lot of money on that. 26:10 That's, like, the bet that Warp is making, and so I think that adoption is gonna happen really quickly as developers, like, fully get, like, "Holy shit, I can just, like, ask the computer [chuckles] to start making my change, provide it the right context, tell it how I want to do it, and it will do it." 26:23 And then I think the more autonomous workflow, which is, like, you know, you might look at a company like Cognition, or I think Google has a thing called Jewels, or OpenAI has a thing called Codex. Like- Mm-hmm... 26:34 that's also happening right now. It's just happening for a very, my understanding is, like, a pretty small slice of tasks that agents can actually do that way. 26:43 And I think a year from now, or six, six to 12 months from now, I think that that will be becoming more and more mainstream as well. 26:50 And so, like, the simpler tasks will be started to be fully automated and the, the real, like, meat of development, which is, like, building features, fixing bugs, fixing stuff in production, will be, like, this agentic workflow. 27:03 But I do think, like, the full automation of, like, simpler tasks is gonna be happening in parallel as well. So it's gonna be wild. And then, like, two years from now, I don't know. 27:12 [laughs] I just straight up don't know. Like, three years, I really don't know. It's crazy. It's crazy how fast it... And my last question is about a book. 27:18 If there is a book that you read this year that you can full-heartedly recommend. 27:23 Yeah, so I'm reading this book called, um, The Beginning of Infinity by David Deutsch, who's, like, this pretty famous physicist who's also, like, a quantum computing person. Mm-hmm. It's a pretty cool book. 27:34 It's taken me a long time [chuckles] to get through it 'cause it's, like, it's a pretty dense book. Like, he's a very- It's dense, yeah... very smart guy. I don't know if... Have you read this book? Started. You started? 27:43 I think it's really fascinating. Like, my background even before I did engineering, I have, like, a philosophy of science degree. And so, you know, it's a lot about, like, what makes human intelligence special. 27:55 He's like- Mm-hmm... 27:55 human intelligence and the ability to, like, produce explanations that are hard to vary is a really, really powerful physical thing in the world and, like, talks about sort of the nature of progress and what that enables. 28:07 He talks a bunch about, like, quantum physics, and- Mm-hmm... he's, like, a subscriber to, like, the multiverse theory of the universe. 28:15 And so, I don't know, it's a really, it's, like, one of these books that's kind of like all-encompassing explanation type book, which I think, like, Guns, Germs, and Steel is one that I read a long time ago that's kind of in that camp. 28:25 Or, like, Gödel, Escher, Bach is another one fr- by Douglas Hofstadter that's another, that's in that camp. Mm-hmm. So it's one of these really cool, like, I'm gonna give you my whole worldview type books. 28:35 I do recommend. I really like- Expand your thinking. It will change the way you think about the world a little bit, yeah. That's cool. Thank you so much. Thank you for your time- Yeah... and insightful conversation. 28:43 Cool. I really appreciate you having me on. Um, this was awesome. 28:47 [upbeat music] So this week we have a major launch that we call Code on Warp that is, uh, should be live by the time this episode comes out, which there's a few key features. 28:58 So one is the ability to actually review diffs with an agent and pair w- on editing those diffs directly within Warp. 29:07 And so the workflow here is, like, you start a coding task, the agent produces a bunch of diffs, and then you basically as a developer, get to code review the agent's diff and have it correct itself. 29:16 Or if you want, you can edit it. 29:18 The second thing is, like, we've invested a bunch in file editing capabilities in Warp, and so you'll find, like, the actual file editor is a much richer, like, experience, kind of at the level of Sublime Text now in Warp, where it's not meant to be a place where you start by typing code, but it is a place where it's like, it's great to iterate and edit an agent's code. 29:38 And then the third thing is that we have, you know, added some, the agent now has the ability to, like, spin up a new project or bring one down from GitHub and create project-based rules and a project-based code base index, and we've made that whole flow a lot simpler for people. 29:53 So including, like, the ability to have different slash commands, which if you're familiar- Mm-hmm... 29:56 with some other agentic tools, you'll find similar in Warp, like /init or /codereview, and so these things that if you want to, like, make Warp your primary coding tool, the experience has gotten way, way, way better this week. 30:08 Awesome. Thank you for this news. [upbeat music]