Warp with Zach Lloyd
===

~Hello. I see you, Kate.~

~Yeah, I see you.~

~I can hear you too.~

~Yep. Where are you? Where are you guys?~

~Boston area.~

~Cool.~

~Yeah. Yeah, it's hot in mass today. I think we're 8 83.~

~Is it humid?~

~Yeah, that's the thing. It, it's like a wet, the wet bulb heat gets reached pretty bad. Pretty, pretty quick.~

~Got it.~

~hey, I've never lived in Florida, so I can't, I can't say anything.~

~Yeah.~

~Test, test, test.~

~Test, test, test.~

~Okay.~

~Hi there and~ [00:00:00] welcome to Pod Rocket. Hi. A podcast brought to you by Log Rocket. Log Rocket helps software teams improve user experience with session replay, error tracking, and product analytics. Try for free@logrocket.com today. My name is Paul and joined with us is Zach Lloyd. He is currently pioneering the development in the company Warp and he has tons of experience, such as, ~you know,~ working on Google's sheets and Google Docs.

~He's, ~he's been around the block and~ so~ we're here to talk about. Warp and some of the newest updates come into warp, such as Warp Drive. Welcome to the podcast, Zach.

Excited to be here, Paul. Thank you for having me.

~So like, like I mentioned,~ you were on about a year ago. ~We had Ben interview you and, and talk about what Warp is and kind of the compelling pieces of reinventing the terminal and be,~ before I spoil too much, for folks who haven't listened, I'll hand it over to you to just tell ~the,~ the newcomers what is warp

~Yep. So,~ so Warp is a, ~uh,~ as a company, we're trying to build software that helps other developers, ~uh,~ ship better software more quickly. As a product, we are a reimagination of the command line terminal. ~Um, so it's kind of~ a [00:01:00] complete rethinking from the ground up, from the perspective of ~like,~ The terminals user experience, how do you improve it?

~Uh,~ how do you make it smarter? How do you make it collaborative? Those are ~sort of ~the three big product, ~uh,~ points we're trying to hit on. And, ~you know,~ the reason that we're focusing on the terminal is that ~it's, uh, you know,~ it's a very widely used tool. Most developers are using the command line every day, whether they like it or not.

~Um,~ it's ~a, uh,~ a place where a lot of important work happens. ~So, you know,~ for a lot of developers, it's where they do their sort of like bill test run. Cycle. For some developers, it's where do they code? If you're doing like production work, any kind of DevOps, s r e or backend work, it's where you're controlling all of your cloud services.

It's probably the most popular platform for writing internal tools. ~And so, you know, we,~ we basically saw an important tool that's used by most developers and. ~You know, and,~ and ~kind of~ my opinion, it's not a great piece of software historically. It's something that hasn't really changed much in the last 40 years.

~And, uh, you know,~ I think that manifests as a being a difficult tool to learn, a difficult tool to configure, ~um, you know,~ basic UX stuff that you [00:02:00] expect, like the mouse working doesn't really work. And ~so, you know,~ what we're trying to do at Warp is build the tool that, ~uh,~ we think. Developers really would want to have to do all these tasks.

~And you know, we're,~ we're a few years into that. ~Uh,~ we're only on Mac right now, but are working to bring it to more platforms and, ~you know,~ I think we have ~a pretty,~ a pretty cool start on it. ~So that's, that's kind of the basics of it.~

~Uh, right on. So~ it's reinventing the terminal is a great way, I feel like, to target, especially like newcomers into the ecosystem, cuz it's a. Archaic, difficult piece of software to wrangle, especially for your first time, ~um,~ and bridging that gap.~ I,~ I know you folks have done some work on making ev even that intro, that first step into the ecosystem easier.

~Um, I know people listening, they're gonna think like, okay, so a terminal, it's kinda like my window into my operating system. There's a tight coupling there~

~Yep.~

~the,~ the ui, the ux, and the operating system. So where does Warp ~kind of~ draw those abstraction divisions?

~Yeah, so. So~ when you're thinking about clo and terminals and shells, and ~there's, there's a few pieces. ~There's ~a, like~ a few pieces of things to clarify. ~So,~ warp war is a terminal, which is the, ~um, you know,~ it's like the actual gooey application that you're running. So it's like a Native Mac app. ~Uh,~ when you run the terminal, what it typically will do is launch a shell.

And so the [00:03:00] shell is ~like,~ Typically bash or zsh or phish. Zsh is the default on Mac now. And the way to think of like the dividing line between those things is, ~you know,~ everything that has to do with like actually making like a Native Mac app, like painting pixels on the screen, handling character input, handling events, ~uh,~ like doing the actual user interaction at the level of the GUI is done through the terminal.

And what the shell does is it takes character input from the terminal. So you type characters in it interprets those. And then it, ~uh,~ runs a program. And those programs, these are all text-based apps, they output text back and then the terminal is responsible for the rendering of those on the screen. ~So,~ So that's like the traditional division, ~the, you know,~ when you're running the terminal app, you're really like running ~a,~ something that is emulating the, like a piece of hardware that isn't really made anymore.

So you're emulating, it used to be like a VT 100, like a physical terminal, like what you would see in the hacker movies. ~Uh,~

is a VT one? What is that? If I could interrupt quickly. I'm curious.

It's like a piece of [00:04:00] hardware that was made by, I think the digital equipment corporation, like d e c. And ~you know,~ the simplest way to think about it was like back in the day what this hardware did was it was like a keyboard. So you could put in characters. Those characters would ba go basically over like a line directly into a, ~uh,~ a shell or some environment that would process the characters and then all of ~like~ the.

Running of programs and like producing output would happen within that shell layer. And then the shell would output characters. And the other big responsibility of the terminal was to interpret that character output and print the characters to the screen. ~Um, and there's like, you know, it's the,~ in its simplest thing, it's like, it's just like,~ um,~ as characters would come back from the shell, the terminal just put them on the screen, but there's also like a sort of, ~uh,~ control language.

That, that has developed. So it's like the shell can write back things that are not characters to be printed to the screen, but they're things that control how the printing actually happens. So if you wanna ~like,~ move the cursor to someplace or, ~uh, you know,~ delete characters. And so there's like [00:05:00] this protocol that exists and that protocol.

It's not like super standardized, but the, ~you know,~ the kind of defacto terminal to shell compatibility is something that was, can shipped with this like old VT 100. ~Uh,~ and then there's other series of VT hardware terminals. But for ~like~ the purposes of ~like, uh, you know,~ a modern developer, I think the interesting thing is like when you're running the terminal app, or even running Warp to some extent, you're running like a piece of software that's actually copying the behavior of this old school piece of hardware Now.

~Okay. Thank you.~

~In,~ in warp. ~We, um, you know, we still,~ we, we are a terminal, we run shells, but ~I like,~ we've ~sort of~ taken the perspective ~that most, like,~ that's like almost an implementation detail as far as a lot of developers are concerned these days. ~Like what they, you know,~ I think a lot of developers, and I think like actually it's ~kind of ~the right mental model is like you're just using a command line interface or you're using like a rep.

And if you think about it like that, ~like we've,~ we've blurred some of the distinctions between terminal and shell ~in like~ in an attempt to produce a user experience that is a little bit more modern and accessible and ~so, Um, you know, warp,~ warp is [00:06:00] backwards, compatible with your shells, but we try to do more stuff in the terminal.

~It's not, you know,~ most terminals are kind of dumb, is how you would talk about it. Like they're really just character rendering devices. Warp is smarter. Like ~we're,~ we're pushing more of the application logic into the terminal layer, and we have more of a vertically integrated command line experience.

And just to give like concrete examples, what does that mean? It means ~like,~ I. We've taken the, ~um,~ the input editor lives in warp as opposed to living inside the shell, which is how it typically works. And the reason we did that is because we can now in warp offer like a modern input editing experience, meaning you can use the mouse.

To edit text, which ~is,~ is crazy ~that~ that doesn't exist in normal terminal. ~Um,~ but even beyond that, ~it's like you get,~ you can do multiple cursor editing, you get syntax highlighting, you get auto suggestions, you get completions. And the completions are like visual completions that work with menus. And so there's a lot of ~like,~ stuff that we can do to, ~um,~ improve the command line experience if you're willing a little bit to blur the lines between the terminal and the shell [00:07:00] experience.

~I feel like that really cleared the lines for me of the lines that you folks are drawing. So like~ thinking about the text editing experience, be being at this higher level that you guys are creating is really compelling cause. If there's something like a dropdown menu I could get, ~I would never, like,~ I would need some crazy terminal plugins

~Yeah,~

~to, to make that work.~

I mean ~it's,~ it's ~kind of~ amazing what developers, ~um,~ have built in terms of like shell plugins to try to. Replicate gooey type interactions. ~Like~ there's a whole class of apps which you'll sometimes hear people talk about called T UI apps. TUI apps, which are like text-based UIs, ~which are like,~ you're essentially using the character grid in a terminal to make something that kind of looks like a gooey menu element.

And I don't know, to me at that point, it's like you probably wanna just have ~like~ the native menus and native interactions and ~like, Um,~ not do it using characters. However, I will say ~there, there, uh,~ there are some intrinsic advantages to that approach, which, ~uh, you know,~ warp will still, it still supports them ~if,~ if that's ~like~ the best way for it to work.

But ~like,~ for instance, one advantage to ~the,~ the text only approaches [00:08:00] that if you are ~like~ attached to a remote machine, ~Um, you can still, like,~ that doesn't even have any kind of ~like~ windowing logic text is a sort of universal wire protocol where you can drive all of this stuff. So ~there are,~ there are some advantages to ~the,~ the old school way of doing it, but what.

~You know,~ we're really focused on is ~like kind of~ from first principles, what's gonna make the most usable command line experience for developers. And ~you know,~ by usable we mean ~like~ you shouldn't have to spend aton and time configuring it. It should work mostly like your other apps work. It should be ~like~ intelligent and ~sort of ~help you accomplish the task that you're trying to accomplish.

It should be collaborative, meaning ~like, you know,~ most terminals, not most, ~like~ basically every terminal except for Warp is like a piece of local desktop software and we're trying to make warp. Something that works for teams, not just for individuals. ~So yeah, we're really,~ like I said, we're trying to reimagine this experience to, ~you know,~ help developers ship better software.

do you find yourself and your team at this ~like~ crossroads of, what is that experience and is it a safe experience for developers ~if this, like~ if they saw Warp as their first thing, ~do you,~ do you run through your mind about ~like,~ oh, is this the right.[00:09:00] 

Level of abstraction to expose them or is it too high? ~Like~ there are some folks that, ~you know, will,~ will say there's value to keeping understanding of the lower level primitives. And so yeah, ~I'm,~ I'm curious about where you guys have maybe run into that decision.

~Yes. That's a,~ that's a great question. ~So, You know, ~like, like I said before, so warp is backwards, compatible, and warp still works. Like you're running commands and getting output and you're running text-based apps ~like that, uh,~ that, that has not changed at all. We don't hide that from the user. I think a good example of what we ~do~ do though, is ~like~ we attempt to make.

The writing of those commands a lot easier. ~And so that could be through, like~ for example, ~like~ we have a feature in Warp, ~uh,~ that uses ai, that lets you ask for a command in natural language. But what we do is we don't just ~like~ go directly from natural language to execution. We go from natural language.

Just showing you the command and giving you a chance to execute it. So as ~like~ a developer, ~you're learning,~ you're learning the command by doing this, and then you're executing it and you're seeing the output. And likewise, ~like~ we have a feature that if you have erroring your output, we can. Suggest fixes for it or [00:10:00] we can use AI to try to explain it, but we don't hide the output.

And so I actually think it's ~like~ a great way of taking sort of your average developer and trying to uplevel them into someone who can be a power user of the tool, cuz the terminal is a. Extremely powerful tool. ~Like~ if you can actually~ kind of ~master it. And so ~our,~ our attitude there is like, how can we provide ~like~ assistance, ~uh,~ not so much ~like~ hiding the complexity, but making the complexity comprehensible and making it faster, ~uh,~ for people to learn.

~Uh, ~and so that's ~like~ the approach there. And then conversely, a lot of the features that we have in work are not about, Learning, they're about, ~um,~ efficiency. And so if you're ~a ~already, ~like~ a pretty strong terminal user, strong developer, ~you know,~ that's a developer that we want using Warp. We're trying to make them be able to use it more quickly ~and, and get like,~ save them keystrokes, save them time, ~like~ save them wasted time from copying and pasting and sharing, like there's all of these sort of.

~Uh,~ UI frictions that add up that, ~you know,~ we're trying, even [00:11:00] for ~the,~ the most experienced developers, we want warp to be ~like, uh, you know,~ a 10 times better terminal. And so ~it's,~ it's really for both.

So if a developer wanted to use it today, you said it's on Mac, right? Something that I could just go download on Mac. Okay. And. Is there like ~a,~ a subscription associated with that? ~For me to be able to like run a, let's say a man page command.~

No. ~So there's no, um, it's,~ it's free software. ~You can,~ you can install it via brew or you can download it. There's a paid plan that we're just, ~uh,~ piloting, which is for small teams. And ~I, I can,~ I can ~sort of~ explain like the collaborative features, but it's about, ~um, you know,~ there is a cloud and team aspect to warp.

That, ~uh,~ is something that's new in the terminal. It's something that has ~like~ a kind of server piece associated with it as a cost that is what we're, ~uh,~ what we ask people to pay for. But no,~ most of you know the all,~ basically all of the single user. Terminal functionality that we've launched is free for, ~uh,~ individuals to use on Mac.

You can get it on warp.dev or you can do brew install dash Cask Warp, and you'll get it that way. ~Um, the, you know,~ one thing that people will notice, and this is the thing that we get a lot [00:12:00] of, ~like, uh,~ we'll say constructive criticism or feedback on, it's like warp. ~You have,~ you have to log into it. ~Um,~ so I don't want anyone to be surprised by that.

And the reason that we ask users to log in, Is that, ~um, you know,~ there is a cloud sort of component to this app, whether it's using the AI features or with Warp Drive, which is ~a,~ a new feature that we just launched. There's the ability to start storing all of your. Frequently use terminal commands and storing your settings and even collaborating with your teammates through Warp and all of that stuff requires the concept of user identity, which is why we ask users to log in.

~Um,~ so ~that~ that's, but there's no pay paywall associated with using, ~uh,~ using it right now.

~The,~ the way that we're taking classic software and putting it into the cloud has. Mis react mixed reactions from a lot of people, but it's indisputable that ~like~ when you have a collaborative workspace, you get, ~you know,~ a lot of higher level outputs that you wouldn't expect before. ~Um, and it's, it's inter,~ I would love to see, for example, statistics on my terminal.

What commands I'm running, [00:13:00] how they're being used, and maybe share that with somebody. ~So, totally,~ totally see where you're coming from with the log perspective. I, I'd love to hop into the warp drive stuff cause I know that's something that's just coming out before. Before we do that though, just wanna remind our listeners that this podcast is brought to you by Log Rocket.

And Log Rock Kind is taking a similar approach to the on the cloud front here, where you can get up and running really quick and get error tracking product analytics. User feedback, heat maps, all sorts of stuff into this general Cloud dashboard where you and your team can solve bugs, find them faster, and spend less time in the console.

So head over to log rocket.com today, and you can spend more time building and less time debugging in your own little local silo. So powers of cloud, powers of collaboration, try it out today. So Warp Drive, Zach is, that's a new thing that just dropped, right?

Yep. Two weeks ago we launched it.

~And ~it's all about. Collaboration and sharing efficiency to my understanding,~ I'm that, ~I'm just glossing over it here,~ but.~

~Yeah, so, ~so [00:14:00] what what it is,~ is, um,~ you could think of it a little bit like Google Drive for your terminal. ~Um,~ the idea is that there are certain things that you do in the terminal or that your teammates do, ~um,~ that would be useful to ~sort of like~ save and have apl a way of looking them up and then sharing them.

And ~so, you know,~ just to give concrete examples like we have. ~Um,~ some commands that we run on our team to like white list ips to access certain parts of our production infrastructure. And we have some commands that we run, ~uh,~ if you are doing like cherry picks to do a release. And so what Warp Drive ~is,~ is it's a spot directly inside of, ~uh,~ warp where you can save these templated commands, which ~we,~ we call them workflows.

~And so,~ It's like a command, a description, a name. You can parameterize them, you can give the parameters, descriptions, and then you save them. And then by default, those things are saved for you. So wherever you access warp, ~uh,~ you have you, you can pull them up in like your command entry, like search dialogue.

But then you also have the ability to make a team. And start building up directly in the [00:15:00] terminal, ~um,~ like your team's, like terminal knowledge base. And so these shared commands are the first thing that we're launching in Warp Drive. We're working on ~a,~ a second thing, which is a, like a terminal notebook format essentially, where.

~Um,~ you can actually put, ~like, you know,~ a good thing would be like, ~how do you,~ how do you do a release or ~like,~ what are the terminal tools that you should use when firefighting? And so you could put that type of documentation directly into the tool itself and have it shareable by your team. ~Um,~ and yeah, this is something that we launched.

Couple weeks ago. It's in a, it's in a preview mode, but we've, ~you know,~ we've actually, we've gotten a bunch of traction from existing users and some new users on it even. And, ~uh, you know,~ I think ~it's,~ it's a cool first step in the direction of how do you make the command line, ~um,~ a collaborative space, not just a space for an individual developer.

~Workflows and, uh, shared commands, shared learning. Right now, and,~ and I know you folks also just got ~a,~ a series B from Sequoia Capital, right? This big event. ~So,~ so ~what's,~ what's coming then? What are you excited about that isn't yet in Warp Drive?

Yeah, so there's a [00:16:00] lot of stuff that I'm really excited about in general. ~So, um, you know,~ warp drive, ~uh,~ I think this sort of like evolution of it is that it should be one spot for you and your team, all of their terminal related stuff. And ~so,~ Workflows and notebooks are two initial things. ~Uh,~ I think some concept of ~like~ a project makes sense.

So some way of organizing, like on a per project basis, what is all of the terminal stuff? ~Um,~ launch configuration. So the ability to set up your terminal in a specific way is the thing that we think should be shareable. Even more fun stuff like theming, I think makes sense is like a shareable thing across the team.

~Um,~ so basically ~any,~ any ~like, uh,~ Thing that you might do could go in here. And then I think there's also more like clever stuff that we can do that's more interactive, where it's like, you can imagine ~a,~ a shared feature that lets you, like, hey, if you experience some particular terminal error, you could have ~like~ a trigger for ~like,~ oh, here's how someone else on my team fix that and like surface that.

~So, you know,~ leveraging the, ~um,~ the communal knowledge and like activity of the team is really exciting. [00:17:00] So, A second, like big bucket of things that we're really excited about is, ~you know,~ again, this is so much hype about it, it's ~kind of~ hard to distinguish, but ~it's,~ it's AI in the terminal, which I think actually for the terminal use case ~is pretty,~ is pretty compelling.

~Um,~ and so we already have a couple features that, ~uh,~ that, that are AI driven. So one is what I mentioned before, which is ~like,~ you can go from natural language. To a command. ~Um,~ and then we also have a sort of deeper chat based interface where it's like, ~uh,~ you can, ~you know,~ if you call me use case, be like this, I have some error.

I'm trying to figure out how to do something on my server. ~How do I, uh,~ how do I do that? And the AI directly in Warp will walk you through it. It's able to ~sort of like,~ Seamlessly understand your terminal outputs and then suggest things for you to run. And so you don't have to do any context switching while you use it, which is pretty cool.

~And then, you know, I think like~ as we think a little bit bigger, where does that go? ~Um,~ I think exploring the. The sort of, ~um,~ use cases of AI agents within the terminal are super interesting. ~So, you know,~ everything in the terminal today works as~ sort of~ a command at a [00:18:00] time. I think there's like a pretty cool concept of ~like, Hey, can, can I,~ instead of just being like, can you gimme the command?

You could ask an AI to help you actually solve a problem or ~like~ set something up or do something at a higher level of abstraction. And that might result in the production of several commands or many commands. And some of them might need your approval. Some of them might be autonomously doable. But the idea of having ~like, you know,~ ex a super experienced ai, ~like, you know, uh, uh,~ c l i Buddy who's always there to help you solve things, I think ~is,~ is a really compelling direction to go.

So yeah, I think like intelligence. Collaboration. And then honestly~ finally,~ I still think there's a lot of stuff to do just with the improve the core user experience of the terminal. So like looking at use cases that developers like might encounter a bunch of friction in, like for instance, ~like. You know,~ if you do the thought experiment of like, how do you make working with logs much better in, in the terminal environment, ~um,~ I think that there's a whole bunch of stuff that we could do around being smarter at how you work with logs.

~Um, ~so there's, there's a lot to do. ~Um,~ I, what [00:19:00] we really would like to do long term and in addition to building these types of features is actually turn warp. Into something that other developers can contribute to more and can build on more. ~Uh,~ we have some extension points right now, for instance, like developers can contribute themes, they can contribute workflows.

~Uh,~ but we would like, ~you know,~ as we, as the product gets more mature to turn warp a little bit into more of ~like~ a ecosystem or platform where, Developers who are interested in making the command line experience much better. ~And I, I,~ I actually think that there's a lot of them, ~like~ there's some massively popular open source projects that are just around, ~uh,~ improving the terminal.

~Uh,~ we wanna enable those developers to tailor the warp experience, ~uh, for,~ for our community of users.

~And, and~

~Thank.~

~excuse me. Zach, you've, like~ we mentioned at the beginning of the podcast, you've worked on, ~you know,~ these collaborative Google projects, ~uh, you know,~ people editing, live, people editing together, and you're the guy to bring that sort of interactivity into this space. So are you foreseeing a future where we have something akin.

To [00:20:00] seeing like Matt's pointer up here ~and,~ and the text he's highlighting.

So the answer is yes. So ~I,~ I think we started, ~um,~ with asynchronous collaboration. So meaning, like we start, we started like building up a library or knowledge base. We are now, ~um,~ exploring and actively working on ~the,~ the actual ~like~ live two people, one session, group of people, one firefighting session, which I think is like a pretty, it's gonna be a pretty cool, ~uh,~ experience once you have the ability.

To do that, and I like the types of use cases I think it unlocks or things like. ~You know,~ you used to be able to like, go over, not tap on someone's shoulders to ask them how to fix something. I think having the ability to ~like~ do a joint debugging session live via a shared terminal is cool. I think, ~uh,~ in firefighting situations, having the ability to see what commands are being run, ~uh, in a, you know,~ in a higher fidelity way than like zoom or screen sharing, ~which is,~ which is the fallback.

~But, you know,~ time and time again, ~it's like the, uh,~ I think it's been proven that it's much better to have a native sharing experience where you can like independently scroll and select texts and [00:21:00] copy paste and ~like~ suggest commands and that, that type of thing. So the answer is yes, ~we're,~ we're working on that.

As part of that, we're also, ~you know,~ working on like, how do we. We're Mac only right now, which is ~not,~ not the goal or plan. The goal is to make Warp available, ~uh,~ on Windows, Linux, and the web. And so we're actively working on expanding, ~um,~ platform support. We started on Mac just cuz it's ~like, you know, with,~ with a startup it's better to.

Sort of start focus and make sure that the product is good, ~uh,~ for some deep use cases before, like increasing ~the,~ the sort of breadth of surface area that you have to maintain. ~But, um, you know,~ we're working ~on,~ on more platforms and I think the web platform is especially interesting because it makes some of these, like collaboration and sharing use cases so much easier if you're.

~You know,~ if your sharing is based on a, ~uh,~ on just ~like~ a link based model rather than on, ~uh,~ like na, like native app sharing, which is doable, ~but, but,~ but ~not as,~ not as frictionless.

~There. I mean, one thing that I'm immediately seeing of, just from my own personal experience, being an Ops server monkey is my first job. I don't know. I don't know if you've heard of SaltStack.~

~What is, no, what is SaltStack?~

~SaltStack was kind, it was kind of like Ansible sort of, not really, but it, you know, you can make these things called salt modules and you could, it's in the shared repo, they're in Python, and then you could down take reference a salt module.~

~They got some inputs and outputs. Fix your ops. We had like 1500, uh, VPCs, believe it or not, as crazy as that sounds. And so like having this modular system to run and debug, I'm like, all this could be replaced with workflows. We could have like some shared environment and we could have these workflows that define our playbooks almost, and, you know, run through them.~

~Uh, I'm, I'm curious to think, uh, to hear from your brain, ~what other sorts of team dynamics do you see already being disrupted [00:22:00] and, ~you know,~ maybe hope to be disrupted in the coming quarters as you unleash these new features?

Yeah, so the, ~you know, the, the,~ the one big trend is ~like,~ how do you ~like, sort of~ standardize and centralize environments that developers are working in? And there's a bunch of different, ~um,~ approaches to doing this. ~Like~ there are things like code spaces and GI pod and that are very like verticalized, ~um, and.~

~You know, I think we're interested, um, in the, like how do you provide all of the, like, surrounding information for an environment, uh, and make that, you know, sort of. Diaspora of informa, like the,~ the spreading of information across multiple systems right now, away from the tools where the information is used ~is,~ is like a, I think, a big time waster and it makes it harder to standardize on things.

~And so, you know,~ our feeling is like by bringing the information into the place that you use it, bringing the documentation, bringing all of the like actions you might take, bringing the ability to search across knowledge, bringing, being the, bringing the ability to like, Realtime collaborate on all this stuff will, ~um, you know,~ make the productivity of working in a consistent environment.

~Uh,~ a lot easier. And ~so,~ yeah, the, ~uh, you know, the, the,~ the sort of disruption there is like going from a world [00:23:00] of like siloed individual setups where people waste a lot of time looking for stuff or figuring out how to do it to a world where people are, ~uh,~ natively in the tool on the same page, and able to share knowledge and contact and collaborate either asynchronously or in real time together.

That's the big shift.

Yeah, I can almost see people that would traditionally not sit next to somebody who's operating or debugging, sitting there. Or maybe they don't need to physically sit there. ~Right?~ Cause we're, you're coming out with sharing features and actually learning how the infrastructure works, how the environments work.

~Yeah. I mean,~ the other cool thing about this is it's not limited to ~like, um, sort of like. ~Teams at companies or something like that. Like I think there is a powerful world where if you are using like a. An open source tool, for instance, that has a c l i, it has a bunch of documentation. We wanna make it possible for that stuff to be like, ~you know,~ sh basically inserted directly in the terminal.

So there's like a kind of community ecosystem that we wanna foster as well. So not just [00:24:00] like private team, ~uh,~ like knowledge sharing. ~Um,~ so we're trying to set up the infrastructure right now to make ~all of that,~ all of that possible. And I do think that's, this just will be a very. Different way of, ~uh, you know,~ developers using the command line and actually, ~you know, it's like,~ and developers just doing their day-to-day job.

It's like they should be doing it in a space that is very tightly connected to all the other developers on their team.

What are some security fe security thoughts, I'll call them security thoughts, security concerns or ~um,~ vulnerabilities that maybe people have pointed out and brought up that you guys have historically tackled and are currently thinking about.

~Yeah, I mean, so there's,~ I think that, ~you know,~ the main thing ~that's~ that's different is that, ~um, you know,~ warp is like a cloud connected, cloud native app. And ~so, you know,~ when, if you're working in normal terminal, everything is local. Now we've been like very careful to be like users, choose what goes.

Into warps cloud. So it takes ~like~ an explicit action, [00:25:00] meaning like there's no, ~um,~ warp does not store any commands. You run, we don't send them off of your local machine. You have to do something that explicitly uses a cloud feature in order for warp, ~uh,~ warps cloud to see these things at all. So for instance, like you have to like, Ask AI for help with something, or you have to explicitly save a command to Warps Cloud.

~And so, you know,~ for that reason, ~like, you know,~ the sensitive stuff that is ex is flowing through terminals is generally not something that Warp has any access to ~on the,~ on the server side. And so ~like~ that drastically limits the, any of the sort of like typical security concerns that people would have going forward.

I actually think we have a chance to make, ~um,~ The terminal much more secure. Like historically, like cloud native apps are, have better security than~ like,~ totally unmanaged, not ~like,~ like desktop apps. And so for instance, ~you know,~ a feature that we could add for enterprises is the ability to ~sort of~ control the configuration of people's terminals.

~Uh,~ where it's ~like, you know,~ right now what people will [00:26:00] do is they'll ~sort of like,~ They'll be like, okay, ~I wanna,~ I wanna configure my terminal to work some way. And so they'll Google for something and they'll get some snippet of script that they'll stick in an RC file and then, ~you know,~ all of a sudden you have untrusted third party code running against your local file system, which is, I think, ~kind of~ the opposite of what most companies would want.

~Um,~ so I think there's like a central administration feature, ~so,~ We're actually right now, ~uh,~ doing a pretty cool like hack week feature that I'm working on with a couple engineers on our team where we do something where we try to. ~Um,~ detect if you're printing anything that's sensitive to the terminal console and like Alli, it meaning ~like~ if you accidentally print out an API key or you print out, ~um,~ like IP addresses that you don't want leaked, and if you're in a situation where you're like screen sharing your terminal and you don't want other people to see that stuff, ~uh,~ we're gonna have the ability for you to ~like, sort of~ have the terminal catch that for you.

~And so, like~ the short answer is ~I get the,~ I get the trepidation around. Cloud-based models. ~Like~ we had this with Google Sheets and Google Docs of course, like people were [00:27:00] scared to put their stuff there, but at the end of the day, there's so many other risks around ~like~ unsecured local software. Are you losing your laptop?

~Like~ around ~like~ not having like central administration for these tools that I think the cloud security model will actually eventually be a feature for us, ~not a,~ not a concern.

And building on, ~you know,~ the personal security here ~and, ~and storing your ~commands, ~commands you don't store. Do you feel ~like~ there's a, a market as well for maybe enterprise customers, ~like, like~ you already alluded to, to ~sort of~ limiting the scope of that which. Certain employees in certain departments can access and how they can access it.

~For~ for sure. ~I mean, this,~ this isn't our short term focus, but I think it's, ~you know, to,~ to be clear, ~we're,~ we're like a venture backed business, ~obviously.~ We're trying to build a sustainable business model, and so ~we're, we're,~ we're always looking for ~like,~ what are ways that we could add value that a business buyer might wanna pay for?

And the notion of ~like~ a business buyer being like, okay, I wanna centralize. ~Um, uh,~ the way that, that our command line works and [00:28:00] they wanna do that for security reasons. They may also might wanna do that just for productivity reasons. Like having everyone have ~like~ a onboard into a standard environment, I think is ~like,~ pretty compelling.

~Um, But,~ but for sure that's the thing that we're looking at. I would say today our focus is much more still on like individuals ~and,~ and recently on small teams trying to, ~to in~ increase their productivity. ~So we,~ we tend to build first with ~like~ the developer in mind and really wanting to build something like that.

The person who's actually using the tool will love, but, ~uh,~ longer term we are also definitely thinking about ways that we can, ~uh,~ build something that's attractive for, ~uh,~ someone who's buying things for the enterprise.

~And Zach, we're, we're kind of already running up on time here, so ~I'm gonna throw you a question out from left field that isn't necessarily related to Warp here, but we're, we have an audience of web developers ~and,~ and, ~um,~ react ~leveler~ lovers and beyond. I didn't forget this. Felt people don't worry. L listening to this cast right here.

~So, in your, uh, development with Warp and with the team,~ What are maybe some web specific things that you just, you're crazy about right now ~that you think~ that you think are cool that maybe have moved the needle for your teams?

Yeah, [00:29:00] so I guess we didn't really touch on this, but Warp is built, ~uh,~ in Rust primarily, but as I mentioned, we're gonna do a, ~uh,~ actually we're working on it right now, a web rendering of it. ~And so the,~ the world of web that we're in, which is pretty interesting, is the world of like web assembly. Web, gpu, WebGL.

~Um,~ and so trying to figure out, ~uh,~ all of. Like how that tool chain works and going from like a, ~you know,~ really high performance native app written, ~uh,~ for the desktop and Russ to having that same type of performance, but still having ~like, you know, uh, uh, not,~ not like doing an insane amount of low level work to make it work on the web is what is where we are at right now.

~Um,~ so that's where most of our like engineering stuff is going. For people who are more like web oriented, we have a lot of them who also use Warp and ~you know, the,~ the value there is in ~like, Uh,~ the nice like workflows and completions that we have for a bunch of web tools. So ~like, you know,~ node and pm all, all those tools.

~Any,~ any like web-based cli you're using Docker, we actually have really great [00:30:00] integrations with. ~Um,~ and so yeah, that's ~like,~ that's ~kind of~ my take on web stuff. But yeah,~ we're,~ we're primarily in built in Rust, which has been a pretty cool experience actually.

Is this the first time that you've led a team into using rust? In, in, in seriousness.

~So, um, we, we,~ it is the first time ~we,~ we chose Russ primarily for performance reasons. It was informed a lot actually by my time at Google. And, ~um, you know,~ it was the goo Google Sheets. Google Docs is built, ~uh,~ in JavaScript. More or less, ~like,~ it's actually more complicated than that. It's actually a lot of it is like trans pile from Java to JavaScript, but like the thing that runs in the browser is JavaScript.

~Um,~ and then there is some more advanced web tech like Canvas, ~uh,~ or SVG renderings that we do. ~Um,~ but we always have a lot of performance problems. And I think, honestly, I think Google Docs, like Google Sheets still has these exact performance problems. And so ~we, uh,~ for warp. ~Because I knew we would like,~ one of the performance challenges is just like rendering huge amounts of scrollable Text is like a notoriously performance intensive [00:31:00] thing.

~Uh, so we,~ we ended up picking Russ, ~uh, as,~ as a stack Russ Plus we go directly to the, ~um,~ To the GPU APIs on Maxa to metal, and we're gonna be doing the same thing on the web. ~Uh,~ and that's actually, ~you know,~ there's a learning curve to Russ, but I'm like a huge fan ~of,~ of Russ at this point. I think it's a great language,

~One. One of my favorite things about Ross is it, I feel like it just combines the functional and non-functional pieces, and like this way, I didn't think it was possible. Very impressive. Yeah.~

~definitely. It's um, Yeah, it's, I mean, the, the, the memory model is, I think the thing that is hardest to get used to, but it has the huge advantage of that, uh, you know, by default with no garbage collection and like, no, like not reference, counting everything, like you just kind of get good performance.~

~But it really, it did bend my head, you know, at the beginning, like how to think about what was going on. But now I'm very used to it.~

~Go read the RU book. It's great. Uh, Zach, is,~ is there anything that you'd want to, ~uh,~ share with our listeners ~as,~ as we're wrapping up, anything you wanna plug?~ Maybe there's a feature about what we didn't mention that's, that excites you or, or anything beyond.~

~um, uh, well,~ I'd love for people to try Warp. ~Uh,~ it's still early. I'd love people to give us feedback, like we read every piece of feedback and try to respond to all of it. ~Um,~ we are, ~you know,~ still growing ~our,~ our team, so I ~I,~ I'm sure you have a lot of. Super talented engineers, product people, designers ~who,~ who listen to this podcast.

~Um, so that's, you know, we're, we're,~ we're still hiring if people, ~uh,~ are interested in what we're doing. ~Um, no, ~other than that, thank you so much for having me. This has been fun.

~It's,~ it's been a pleasure, Zach, and just to drive it home. If you're listening, you wanna check it out. Warp dev there. They got the dev

Warp Dev. Yes, thank you.

That's it. Thanks, Zach. It's been a pleasure.

Thanks Paul.[00:32:00]