Inside Claude Code From the Engineers Who Built It

- What made it work really well is that collicotas access to everything that an engineer does at the terminal, everything you can do, collico can do, there's nothing in between. - There's an actually an increasing number of people internally at Anthropic that are using like a lot of credits, like spending like over a thousand bucks every month. We see this like power user behavior. This is something that they teach in YCE. If you can solve your own problem, it's much more likely you're solving the problem for others. There's this really old idea and product called latent demand. You build a product in a way that is hackable, that is kind of open-ended enough that people can abuse it for other use cases it wasn't really designed for. Then you build for that because you kind of know there's demand for it. - You think the CLI is the final form factor? Are we going to be using collicoode in the CLI primarily in a year or in three years, or is there something else that's better? (upbeat music) - This podcast is sponsored by Google. Hey folks, I'm Omar, product and design lead at Google DeepMind. We just launched a revamped vibe coding experience in AI Studio that lets you mix and match AI capabilities to turn your ideas into reality faster than ever. Just describe your app and Gemini will automatically wire up the right models and APIs for you. And if you need a spark, hit I'm feeling lucky and we'll help you get started. Head to AI.Studio/build to create your first app. - Matt, Boris, thank you so much for being here. - Thanks for having us. - Yeah. So for people who don't know you, you are the creators of Claude code. Thank you very much from the bottom of my heart. It's, I love Claude code. - That's amazing to hear. (laughing) That's what we love to hear. - Okay, I think the place I want to start is, when I first used it, there was like this moment. Like I think it was around when Sonic 3.7 came out where I was like, I used it and I was like, holy shit. This is like a completely new paradigm. It's a completely new way of thinking about code. And the big difference was you went all the way and just eliminated the text editor and you're just like, all you do is like talk to the terminal and that's it. And you know, previous paradigms of AI programming, previous harnesses have been like, you have a text editor and you have the AI on the side and it's kind of like, or it's the tab complete. So take me through like that decision process, that process of architecting this new paradigm. How do you think about that? - Yeah, I think the most important thing is it was not intentional at all. (laughing) We sort of ended up with it. So at the time when I joined Anthropic, we were still on different teams at the time. There was this previous predecessor to Claude code. It was called Clyde, like CLIDE, I don't know. And it was this like research project, you know, it took like a minute to start up. It was this kind of like really heavy Python thing. It had to like run a bunch of indexing and stuff. And when I joined, I wanted to ship my first PR and I hand-wrote it like a, you know, like a noob in that. - In a time like I, I don't know about Eddie, these tools, like I don't know. - Thank you for admitting that on the podcast. (laughing) - I didn't know any better. And then I put up this PR and Adam Wolf, who was the end manager for our team for a while. He was my ramp up buddy and he just like rejected the PR and he was like, you wrote this by hand, what are you doing? He was Clyde. 'Cause he was also hacking a lot on Clyde at the time. And so I tried Clyde, I gave it the description of the task and it just like one shot at this thing. And this was like, you know, Sonic 3.5. So I still had to fix a thing even for this kind of basic task. And the harness was super old. So it took like five minutes to turn this thing out and just took forever. But it worked and I was just mind blown that this was even possible. And they just kind of got the gears turning. Maybe you don't actually need IDE. And then later on, I was prototyping using the Anthropic API and the easiest way to do that was just building a little app in the terminal 'cause that way I didn't have to build a UI or anything. And I started just making a little chat out. And then I just started thinking maybe we could do something a little bit like Clyde. So let me build like a little Clyde. And it actually ended up being a lot more useful than that without a lot of work. And I think the biggest revolution for me was when we started to give the model tools, they just started using tools. And it was just use this insane moment. Like the model just wants to use tools. Like we give it bash and they just started using bash, writing AppleScript to like automate stuff in response to questions. And I was like, this is just the craziest thing I've never seen anything like this. 'Cause at the time I had only used IDE. So like, you know, like text editing and a little like one line autocomplete, multi-line autocomplete, whatever. So that's where this came from. It was this kind of convergence of prototyping, but also kind of seeing what's possible in kind of like a very rough way. And this thing ended up being surprisingly useful. And I think it was the same for us. I think for me it was like kind of Sonic 4 Opus 4. That's where that magic moment was. Where I was like, oh my god, this thing works. - That's interesting. So like tell me about that, the tool moment because I think that is one of the special things about Cloud Code is it just writes bash and it's really good at it. And I think a lot of previous agent architectures or even anyone building agent today, your first instinct might be, okay, we're gonna give it a find file tool and then we're gonna give it a open file tool and you build all these like custom wrappers for, you know, all the different actions you might want the agent to take. But Cloud Code just uses bash and it's really good at it. So how do you think about what you learned from that? - Yeah, I think we're at this point in our Cloud Code actually has a bunch of tools. I think it's like a dozen or something like this. We actually like add and remove tools most weeks. So this change is pretty often. But today there actually is a search, there's a tool for searching. And we do this for two reasons. One is the UX, so we can show the result a little bit nicer to the user 'cause there's still a human in the loop right now for most tasks. And the second one is for permissions. So if you say in your like Cloud Code like settings.js on this file, you cannot read. We have to kind of enforce this. We enforce it for bash, but we can do it a little bit more efficiently for if we have a specific search tool. But definitely we wanna like unshipped tools and kind of keep it simple for the model. Like last week or two weeks ago, we unshipped the ls tool. Because in the past we needed it, but then we actually built a way to enforce this kind of permission system for bash. So in bash, if we know that you're not allowed to read a particular directory, Cloud's not allowed to aless that directory. And because we can enforce that consistently, we don't need this tool anymore. And this is nice 'cause it's a little less choice for Cloud, a little less stuff in context. Got it. And how do you guys split responsibility on the team? - I would say Boris sets the technical direction and has been the product visionary for a lot of the features that we've come out with. I see myself as more of like a supporting role to make sure that that one that like our pricing and packaging resonates with our users to making sure that we're shepherding our features across the launch process. So from like deciding, all right, like these are the prototypes that we should definitely add food to like setting the quality threshold for amp booting through to communicating that to our end users. And there's definitely some new initiatives that we're working on that I would say historically, a lot of quad code has been built bottoms up like Boris and a lot of the core team members have just had these great ideas for to do this, sub agents hooks, like all these are bottoms up. As we think about expanding to more services and bring quad code to our places, I think a lot of those are more like, all right, let's talk to customers, let's bring engineers into those conversations and prioritize those services and knock them out. - Got it. What is amp fooding? - Oh, amp fooding is-- - Oh, amp fooding? - Oh, it means dog fooding. (laughing) - Enthropic ads. - Yeah. - And I got it. - There are nicknames for internal employees is aunt. And so aunt fooding is our version of dog fooding. Internally over, I think 70 or 80% of aunts, technical and theropic employees, use collid code every day. And so every time we are thinking about a new feature, we push it out to people internally and we get so much feedback. We have feedback channel. I think we get a post every five minutes. And so you get really quick signal on whether people like it, whether it's buggy or whether it's not good and we should unshift it. - You can tell. You can tell that someone that is building stuff is using it all the time to build it because like it's ergonomics just makes sense if you're trying to build stuff and that that only happens if you're like, aunt fooding. (laughs) Yeah, and I think that that's a really interesting paradigm for building new stuff like that sort of bottoms up. I make something for myself. Tell me about that. - Yeah, and cat is also so humble. I think cat has a really big role in the product direction also. Like it comes from everyone on the team. And like these specific examples, this actually came from everyone on the team, like to do lists and sub agents that was sidd, hooks, dicks and ship that, plugins, daisy ship that. So like everyone on the team, like these ideas come from everyone. And so I think for us, like we build this core agent loop and this kind of core experience. And then everyone on the team use the product all the time. And so everyone outside the team use the product all the time. And so there's just all these chances to build things that serve these needs. Like for example, like bash mode. You know, like the exclamation mark and you can type in bash commands. This was just like many months ago I was using quad code and I was going back and forth between two terminals and just thought it was kind of annoying. And just on a way, I asked quad to kind of think of ideas, I thought of this like exclamation mark bash mode. And then I was like great, make it pink and then ship it. (laughing) It just did it. And like that's the thing that's still kind of persisted. And you know, now you see kind of others also kind of catching on to that. - That's funny. I actually didn't know that. And that's extremely useful 'cause I always have to open up a new tab to like run any bash commands. So you just do an exclamation point and then it just like runs it directly instead of filtering it through all the cloud stuff. - Yeah, and quad code sees the full output too. - Interesting, that's perfect. - So anything you see in the quad code view, quad code also sees. - Okay, that's really interesting. - And this is kind of a UX thing that we're thinking about. Like in the past tools are built for engineers but now it's equal parts engineers and model. And so like as an engineer you can see the output but it's actually quite useful for the model also. And this is part of the philosophy also like everything is dual use. So for example, the model can also call slash commands. So I have a slash command for a slash commit. Where I run through kind of a few different steps like differing and generating a reasonable commit message and this kind of stuff. I run it manually but also quad can run this for me. And this is pretty useful 'cause we get to share this logic. We get to kind of define this tool and then we both get to use it. - Yeah, what are the differences in designing tools that are dual use from designing tools that are used by one or the other? - Surprisingly, it's the same. - Okay, so far. - Yeah. I sort of feel like this kind of elegant design for humans translates really well to the models. - So you're just thinking about what would make sense to you and the model generally, it makes sense to the model too if it makes sense to you? - Yeah, I think one of the really cool things about quad code being a terminal UI and what made it work really well is that quad code has access to everything that an engineer does at the terminal. And I think when it comes to whether the tool should be dual use or not, I think making them dual use actually makes the tools a lot easier to understand. It just means that, okay, everything you can do, quad code can do, there's nothing in between. - Yeah, that's interesting. Yeah, there are a couple of those decisions. So no code editor, it's in the terminal so it has access to your files and it's on your computer versus like in the cloud and a virtual machine. So you get repeated, you get to use it in a repeated way where you can build up your cloud MD file or build slash commands and all that kind of stuff where it becomes very composable and extensible from a very simple starting point. And I'm curious about how you think about for people who are thinking about, okay, I want to build an agent, I want to build probably not cloud code, but like something else, how you get that simple package that then can extend and be really powerful over time. - For me, I start by just thinking about it like developing any kind of product where you have to solve the problem for yourself before you can solve it for others. And like this is something that they teach in YCE is you have to start with yourself. So like if you can solve your own problem, it's much more likely you're solving the problem for others. - And I think for coding, starting locally is the reasonable thing and now we have cloud code on the web so you can also use it with a virtual machine and you can use it in a remote setting and this is super useful when you're on the go, you want to fix that with your phone. - And this is sort of, we started proving this out kind of a step at a time where you can do at cloud and GitHub. And I use this every day, like on the way to work, I'm like at a red light, I probably shouldn't be doing this, but I'm like, yeah, yeah, yeah, yeah, yeah. At a red light, and then I'm like at cloud, you know, fix this issue or whatever. And so it's just really useful to be able to control it from your phone and this kind of proves out this experience. I don't know if this necessarily makes sense for every kind of use case. For coding, I think starting locals, right? I don't know if this is true for everything, you know? Got it. What are the slash commands you guys use? - Slash PR commit. Yeah, it's, I think the PR commit slash command makes it a lot faster for call to know exactly what bash commands to run in order to make a commit. - And what is the PR commit? Is slash command due for people who aren't familiar? - Oh, it just tells it like exactly how to make a commit. - Okay. - And you can like, dynamic, you can say like, okay, these are the three bash commands I need to be run. - Got it. And it was pretty cool as also we have this kind of templating system built into slash commands. So we actually run the bash commands ahead of time. They're like embedded into the slash command. And you can also pre-allow certain tool invocations. So for that slash command, we say allow, you know, get commit, get push, ghpr. And so you don't get asked for permission after you run the slash command because we have like a permission based security system. And then also it uses haiku, which is pretty cool. So it's kind of a cheaper model and faster. Yeah, and for me, I use like commit, commit PR, featured of, we use a lot to like create this one. It's kind of cool. We kind of like walks you through step by step, building something. So we prompt quad like first, ask me how to, what exactly I want, like build the specification. And then, you know, kind of like build like a detailed plan and then make it to do with, walk through step by step. So it's kind of like more structured feature development. And then I think the last one that probably is a lot, so we use like security review for all of our PRs and then also code review. So like quad does all of our code review internally and anthropic. You know, there's still a human approving it, like quad does kind of the first step in code review. That's just a slash code review, a slash command. Got it. Yeah, what are the things? I would love to go deeper into like the, how do you make a good plan? So the sort of the featured dev thing. 'Cause I think there's a lot of like little tricks that, I'm starting to find, or people at every start are starting to find that work. And I'm curious like, what are things that we're missing? So for example, one step in the, one unintuitive step of the, you know, plan development process is even if I don't exactly know what the thing that needs to be built is, I just have like a little sentence in my mind, like I want feature X. I have cloud just like implement it, just without giving it anything else and I see what it does. And that helps me understand like, okay, here's actually what I mean 'cause it made all these different mistakes or like it did something that I didn't expect that might be good. And then I use that like the learning from the sort of throwaway development. You just clear it out and then that helps me write a better plan spec for the actual feature development, which is something that you would never do before 'cause it'd be too expensive to just like YOLO send in an engineer on a feature that you hadn't actually spec'd out. But because you have cloud going through your code base and doing stuff, you can like learn stuff from it that helps inform the actual plan that you make. - Yeah, I feel maybe, I can start and I'm curious how you use it too. I think there's like a few different modes, maybe for me, like one is prototyping mode. So like traditional engineering prototyping, you wanna kind of build the simplest possible thing that touches all the systems just so you can kind of get a vague sense of like, what are the systems? There's unknowns and just to kind of trace through everything. And so I do the exact same thing as you did. Like, cloud just does the thing and then I see where it messes up and then I'll ask it to just throw it away and do it again. So just, it escaped twice, go back to the old checkpoint and then try again. I think there's also maybe two other kinds of tasks. So one is just things that clock in one shot and I feel pretty confident I can do it. So I'll just tell it and then I'll just go to a different tab and I'll shift tab to auto-except and then just go do something else or go to another one of my quads and tend to that. All does this. But also there's this kind of like harder feature development. So these are, you know, things are, maybe in the past I would have taken like a few hours of engineering time. And for this, usually I'll shift tab into plan mode and then align on the plan first before it even writes any code. And I think what's really hard about this is the boundary changes with every model in kind of a surprising way. Where the newer models, they're more intelligent. So the boundary of what you need plan mode for got pushed out like a little bit. Like before you used to need to plan, now you don't. And I think this is general trend of like stuff that used to be scaffolding with a more advanced model. It gets pushed into the model itself and the model kind of tends to subsume everything over time. Yeah, how do you think about like building a agent harness that isn't just going to, like you're not spending a bunch of time building stuff that is just going to be subsumed into the model in three months when the new code comes out? Like yeah, how do you know what to build versus what you just say? It doesn't work quite yet, but next time it's going to work. So we're not going to spend time on it. Ooh, I think we build most things that we think would improve code codes capabilities even if that means we'll have to get rid of it in three months. If anything, we hope that we will get rid of it in three months. I think for now, we just want to offer the most premium experience possible. And so we're not too worried about throwaway work. Interesting. Yeah, and an example of this is something like even like plan mode itself. I think we'll probably unship it at some point when Cloud can just figure out from your intent that you probably want to plan first. Or, you know, for example, I just deleted like 2,000 tokens or something from the system prompt yesterday. This goes like, like, Sonic 4.5 doesn't need it anymore. But Opus 4.1 did need it. What about, you know, in the case where the latest frontier model doesn't need it, but, you know, you're trying to figure out how to make it more efficient because you have so many users that, you know, maybe you're not going to use Opus or Sonic 4.5 for everything. Maybe you're going to use Hikou. So there's a trade-off between having a more elaborate harness for Hikou versus just like not spending time on it, using Sonic, eating the cost, and working on more frontier type stuff. In general, we've positioned Cloud Code to be a very premium offering. So our North Star is making sure that works incredibly well with the absolutely most powerful model we have, which is Sonic 4.5 right now. We are investigating how to make it work really well for like future generations of smaller models. But it's not the top priority for us. OK. What do you think about-- you know, one thing that I notice is we get models often-- and thank you very much for this. We get models a lot before they come out. And it's our job to kind of figure out, is it any good? And over the last six months, when I'm testing Cloud, for example, in the Cloud app, with a new frontier model, it's actually very hard to tell whether it's how-- whether it's better immediately. But it's really easy to tell in Cloud Code, because the harness matters a lot for the performance that you get out of the model. And you guys have the benefit of building Cloud Code inside of the-- inside of Anthropics. So there's like a much tighter integration between the fundamental model training and the harness that you're building. And they seem to kind of like really impact each other. So how does that work internally? And what are the benefits you get from having that like tight integration? Yeah. I think the biggest thing is like researchers just use this. And so as they see what's working and what's not, they can improve stuff. We do a lot of e-vells and things like that that can communicate back and forth and understand where exactly the model's at. But yeah, there's this frontier where you need to give the model a hard enough task to really push the limit of the model. And if you don't do this, then all models are kind of equal. But if you give it a pretty hard task, you can tell the difference. What's that agent's to use? I have a planner sub agent that I use. I have a code review sub agent. Code review is actually something where sometimes they use a sub agent. Sometimes they use a slash command. So usually NCI is a slash command. But in secret, I use a sub agent for the same thing. It's a good question. Yeah, maybe it's like a matter of taste. Yeah, I don't know. I don't know. I think it's maybe when you're running a secrecy, it's kind of nice to fork off the context window a little bit because all the stuff that's going on in the code review, it's not relevant to what I'm doing next. But NCI, it just doesn't matter. Are you ever spawning like 10 sub agents at once? And for what? For me, I do it mostly for big migrations. OK. This is like the big thing. Actually, we have-- so this code review slash command that we use, there's a bunch of sub agents there. And so one of the steps is like find all the issues. And so there's one sub agent that's checking for a quad Md compliance. There's another sub agent that's looking through git history to see what's going on. Another sub agent that's looking for obvious bugs. And then we do this de-duping quality step after. So they find a bunch of stuff. A lot of these are false positives. And so then we spawn five more sub agents. And these are all just checking for false positives. And in the end, the result is awesome. It finds all the real issues without the false issues. That's great. I actually do that. So one of my non-technical cloud code use cases is expense filing. So when I'm in SF right now, so I have all these expenses. And so I built this little cloud project in cloud code that it uses one of these finance APIs. They just download all my credit card transactions. And then it decides these are probably the expenses that I'm going to have to file. And then I have two sub agents, one that represents me and when they represents the company. And they do battle to figure out what's the proper actual set of expenses. It's like an auditor sub agent and like a pro-dan sub agent. So yeah, that kind of thing. The sort of like opponent processor pattern seems to be like an interesting one. Yeah, yeah, it's cool. I feel like when sub agents were first becoming a thing, actually it wouldn't spark us. There was like a Reddit thread a well back. We're someone made sub agents for like, there was like a front end dev and a back end dev. I'm like a designer testing dev. Like there was like a PM sub agent. And this is like, you know, it's cute. It feels like a little maybe too anthropomorphic. Maybe there's something to this. But I think like the value is actually like the uncorrelated context windows. We have these two context windows that don't know about each other. And this is kind of interesting. And you tend to get better results this way. What about you? Do you have any interesting sub agents to use? So I've been tinkering with one that is really good at front end testing. So it uses playwright to like see, all right, what are like all the errors that are coincide and pull them in and try to test more steps of that app. It's not totally there yet, but I've seen signs of life. And I think it's the kind of thing that we could potentially bundle in one of our plugins market voices. Yeah. Definitely, I've used something like that just with puppeteer and just like watching it build something and then open up the browser. And then be like, oh, I need to change this. It's like, this is like, oh my god. Yeah, it's really cool. It's really cool. I think we're starting to see the beginnings of this massive, multi-massive sub agents. I don't know what to call this like swarms or something. There's a bunch of people-- there's actually an increasing number of people internally out on Thropic that are using a lot of credits every month, spending like over 1,000 bucks every month. And this percent of people is growing actually pretty fast. And I think the common use case is code migration. And so what they're doing is framework A to framework B. There's the main agent that makes a big to do this for everything. And then just kind of map reduce over a bunch of sub agents. So you instruct a lot of like, yeah, start 10 agents and then just go like 10 at a time and just migrate all this stuff over. That's interesting. What would be a concrete example of the kind of migration that you're talking about? I think the most classic is like Lint rules. So there's some kind of way where you're rolling out. There's no autofixer because it's like AST analysis can't really. It's kind of too simplistic for it. I think other stuff is like framework migrations. We just migrated from like one testing framework to a different one. That's a pretty common one where it's super easy to verify the output. One of the things that I found is-- and this is both for projects inside of every, and then just open source projects. It's like if you're someone building a product and you want to build a feature that's been done before. So maybe like an example that people might need to implement a bunch is like memory. How do you do memory? Because we have a bunch of different products internally, you can just like spawn cloud sub agents to be like, how do these three other products do it? And there's like possibility for just like tacit co-chairing where you don't need to like have an API. Or you don't need to like ask anyone. You can just be like, how do we do this already? And then use the best practices to build your own. And you can also do that with open source. Because there's like tons of open source projects where people are like, you know, they've been working on memory for like a year, and it's like really, really good. And you'd be like, what are the patterns that people have figured out? And which ones do I want to implement? Totally. You can also connect your version control system if you've built a similar feature in the past. Cloud code can use those APIs like query GitHub directly and find how people implemented a similar feature in the past. And read that code and copy the relevant parts. Yeah. Is there-- have you found any use for like log files of, OK, here's the full history of like how I implemented it? And like, is that important to give to Cloud? And how are you implementing that? They're making it useful for it. Some people swear by it. There are some people out in the topic where for every task they do, they tell Cloud code to write a diary entry in a specific format. That's just documents like, what did it do? What did it try? Why didn't it work? And then they even have these agents that like look over the past memory and synthesize it into observations. I think this is like the starting butting. There's like something interesting here that we could productize, but it's a new emerging pattern that we're seeing that works well. I think the hard thing about like one-shotting memory from just one transcript is that it's hard to know how relevant a specific instruction is to all future tasks. Like our canonical example is, if I say make the button pink, I don't want you to remember to make all buttons pink in the future. And so I think synthesizing the memory from a lot of logs is a way to find these patterns more consistently. It seems like you probably need, like there's some things where you're going to know you'll be able to synthesize or summarize in this top-down way. This will be useful later. And you'll know the right level of abstraction at which it might be useful. But then there's also a lot of stuff where it's like, you actually any given commit log and make the button pink, it could be useful for kind of an infinite number of different reasons that you're not going to know beforehand. So you also need the model to be able to look up all similar past commits and surface that at the right time. Is that something that you're also thinking about? Yeah, I think there could be something like that. Maybe I think one way to see it is this kind of like traditional memory storage work, like mehmex, kind of stuff where you just want to put all the information into the system. And then it's kind of a retrieval problem after that. Yeah, I think as the model also gets smarter, it naturally-- I've seen it start to naturally do this also with Sonic 4.5, where if it's stuck on something, it'll just naturally start looking like we talked about before, using bash spontaneously to just look through, get history, and be like, oh, OK, yeah, this is kind of an interesting way to do it. Yeah, one of the things that we were talking before we started recording one of the things that we're doing inside of every-- I feel like it has really changed the way that we do engineering, because everyone is Cloud Code like CLI Pilled. And we have this engineering paradigm that we call compounding engineering, where in normal engineering, every feature you add, it makes it harder to add the next feature. And in compounding engineering, your goal is to make the next feature easier to build from the feature that you just added. And the way that we do that is we try to codify all the learnings from everything that we've done to build the feature. So how to make the plan, and what parts of the plan need to be changed, when we started testing it, what issues do we find, what are the things that we missed, and then we codify them back into all the prompts and all the sub agents and all the slash commands so that the next time, when someone does something like this, it catches it, and that makes it easier. And that's why for me, for example, I can hop into one of our code bases and start being productive, even though I don't know anything about how the code works, because we have this built up memory system of all the stuff that we've learned as we've implemented stuff. But we've had to build that ourselves. I'm curious, are you working on that kind of loop? So the Cloud Code does that automatically. Yeah, we're starting to think about it. It's funny, we heard the same thing from Fiona. She just joined the team, and she's our manager. She hasn't coded in like 10 years, something like that. And she was winning peers on her first day. And she was like, yeah, not only did I kind of-- I forgot how to code, and Cloud Code kind of made it super easy to just get back into it. But also, I didn't need to ramp up on any context, because I kind of knew all this. And I think a lot of it is about when people put up pull requests for Cloud Code itself, and I think our customers tell us that they do some more stuff pretty often. If you see a mistake, I'll just be like, at Cloud, add this to Cloud MD so that the next time, it just knows this automatically. And you can kind of instill this memory in kind of a variety of ways. So you can say, at Cloud, add it to Cloud MD. You can also say, at Cloud, write a test. That's like an easy way to make sure this doesn't regress. And I don't feel bad asking anyone to write tests anymore. Right. It's just super easy. And I think probably close to 100% of our tests are just written by Cloud. And if they're bad, we just won't commit it. And then the good ones stay committed. And then also, I think lint rules are a big one. So first off, that's enforced pretty often. We actually have a bunch of internal lint rules. Cloud writes 100% of these. And this is mostly just at Cloud in a PR, write this one rule. And yeah, there's sort of this problem right now, but how do you do this automatically? And I think generally how Cat and I think about it is we see this power user behavior. And the first step is how do you enable that? By making the product hackable. So the best users can figure out how to do this cool new thing. But then really the hard work starts of like, how do you take this and bring it to everyone else? And for me, I count myself and everyone else bucket. I don't really know how to use them. I don't have this crazy T-Box set up. So I have a pretty vanilla set up. So if you can make a feature that I'll use, it's a pretty good indicator that other kind of average engineers will use it. That is interesting. Tell me about that, because that's something I think about all the time is making something that is extensible and flexible enough that power users can find like novel ways to use it that you would not have even dreamed of. But it's also simple enough that anyone can use it and they can be productive with it. And you can kind of pull what the power users find back into the basic experience. How do you think about making those design and product decisions so that you enable that? In general, we think that like every engine environment is a little bit different from the others. And so it's really important that every part of our system is extensible. So everything from your status line to adding your own slash commands through to hooks, which let you insert a bit of determinism at pretty much any step in caught code. So we think these are like the basic building blocks that we give to every engineer that they can play with. For plugins, plugins actually are-- so it was built by Daisy on our team. And this is our attempt to make it a lot easier for the average user like us to bring these slash commands and hooks into our workflows. And so what plugins does is it lets you browse existing MCP servers, existing hooks, existing plugins, and just like-- or sorry, existing like slash commands, and just let you write one command in caught code to pull that in for yourself. There's this really old idea and product of waiting demand, which I think is probably the main way that I personally think about product and think about what to build next. And it's a super simple idea. It's you build a product in a way that is hackable, that is kind of open-ended enough that people can abuse it for other use cases. It wasn't really designed for it. Then you see how people abuse it, and then you build for that, because you kind of know there is demand for it. And when I was at Meta, this is how we build kind of all the big products. I think almost every single big product had this nugget of waiting demand in it. You know, like, for example, some of like Facebook data, it came from this idea that when we looked at who looks at people's profiles, I think 60% of views were between people of opposite gender, so kind of like traditional setup, that were not friends with each other. And so we're like, oh, man, OK, maybe there's-- maybe if we want a dating product, we can kind of harness this demand that exists. That's interesting. And for Marketplace, it was pretty similar. I think it was like 40% of posts in Facebook groups at the time were by cell posts. And so I go, OK, people are trying to use this product by themselves. We just build a product around it that's probably going to work. And so we think about it kind of similarly, but also we have the luxury of building for developers and developers love hacking stuff, and they love customizing stuff. And it's like, as a user of our own product, it makes it so fun to build and use this thing. And so yeah, like we just build the right extension points. We see how people use it, and that kind of tells us what to build next. Like, for example, we got all these user requests for people like, dude, code is asking me for all these permissions. And I'm out here getting coffee. I don't know that it's asking me for permissions. How can I just get it to ping me on Slack? And so we built hooks, Dixon built hooks, so that people could get pinged on Slack. And you could get pinged on Slack for anything that you want to get pinged on Slack for. And so it was very much like people really wanted the ability to do something. We didn't want to build the integration ourselves. And so we exposed hooks for people to do that. The thing that makes me think of is you recently released, you kind of moved or rebranded how you talk about Cloud Code to be this more general-purpose agent SDK. Is that driven by some latent demand where you sort of saw there's a more general-purpose use case for what you built? We realized that somewhere to how you were talking about using Cloud Code for things outside of coding, we saw this happen a lot. We get a ton of stories of people who are using Cloud Code to help them write a blog and manage all the data inputs and take a first pass in their own tone. We find people building email assistance on this. I use it for a lot of just market research because at the core, it's like an agent that can just go on for an infinite amount of time as long as you give it a concrete task, and it's able to fetch the right underlying data. So one of the things I was working on was I wanted to look at all the companies in the world and how many engineers they had to create a ranking. And this is something that Cloud Code can do even though it's not a traditional coding use case. So you realize that the underlying primitives were really general, as long as you have an agent loop that can continue running for a long period of time, and you're able to access the internet and write code and run code pretty much you can, if you squint, you can kind of build anything on it. And I think at the point where we rebranded it from the Cloud Code SDK to the Cloud Asian SDK, there was already many thousands of companies using this thing. And a lot of those use cases were not about coding. So it's both internally and externally. We kind of saw that. It's like health assistance, like financial analysts, legal assistance. It was pretty broad. What are the coolest ones? I feel like, actually, you had a NOAA briar on the podcast recently. I thought the obsidian mind-mapping note-keeping use case is really cool. It's insane how many people use it for this. It's a particular combination. I think some coding or kind of coding adjacent use cases that are kind of cool is we have this issue tracker for Cloud Code. The team is just constantly underwater trying to keep up with all the issues coming in. There's just so many. And so a lot de-dupes the issues. And it automatically finds duplicates. And it's extremely good at it. It also does first pass resolution. So usually when there's an issue, it'll proactively put up a PR internally. And this is a new thing that Enigo on the team built. So this is pretty cool. There's also like on-call and kind of collecting signals from other places, like getting like sentry logs and getting like logs from BigQuery and kind of collating all this, plus just really good at doing this. Because it's all just bash in hand. And so these are all kind of these internal use cases that I saw. Is it, so when it's collating logs or de-dubing issues is that-- you have clouds continually running in the background. And is that something that you're building for? It gets triggered for that particular one. It gets triggered whenever a new issue is filed. So it runs once, but it can choose to run for as long as it needs. Got it. What about the idea of clouds always running? Ooh, proactive clouds. I think it's definitely where we want to get to. I would say right now we're very focused on making quadcoding incredibly reliable for individual tasks. And if you think about-- if you think about multi-line autocomplete and then single turn agents, and then now we're working on quadcoding that can complete tasks, I feel like if you trace this curve, eventually you go to even higher levels of abstraction, even more complicated tasks. And then hopefully, the next step after that is a lot more proactivity. So just understanding what your team's goals are, what your goals are, being able to say, hey, I think you probably want to try the teacher, and here's the first pass at the code, and here are the assumptions I made, and are these correct? I can't wait. And I think probably right after that is, Claude is now your manager. [LAUGHTER] That's not in the plan. So everyone on the team was super excited that we were talking today. And they gave me a bunch of questions. I don't want to make sure I hit all the questions. Oh, here's a good one. Why did you choose a gentech rag over vector search in your architecture and are vector embeddings still relevant? So actually, initially, we did use vector embeddings. They're just really tricky to maintain, because you have to continuously re-index the code, and they might get out of date. And you have local changes, so those need to make it in. And then as we thought about what does it feel like for an external enterprise that's adopted, we realized that this exposes a lot more surface area than security risk. We also found that actually, Claude code is really good, and Claude models are really good at a gentech search. So you can get to the same accuracy level with a gentech search, and it's just a much cleaner deployment story. That's really interesting. If you do want to bring semantic search to Claude code, you can do so via an MCP tool. So if you want to manage your own index and expose an MCP tool that lets Claude code call that, that would work. What do you think of the top MCPs to use with Claude code? Puppetier and playwright are pretty high up there. Definitely, yeah. Century has a really good one. Asana has a really good one. Do you think that there are any power user tips that you see people inside of Anthropic or other people who are inside of organizations that are big Claude code power users that people don't know about what they should? One thing that Claude code doesn't naturally like to do, but that I personally find very useful, is Claude code doesn't naturally like to ask questions. But if you're brainstorming with a thought partner, a collaborator, usually you do ask questions back and forth to each other. And so this is one of the things that I like to do, especially in plan mode. I'll just tell Claude code, like, hey, we're just brainstorming this thing. Please ask me questions. If there's anything you're unsure about, I want you to ask questions, and I'll do it. And I think that actually helps you arrive at a better answer. There's also so many tips that we can share. I think there's a few really common mistakes I see people make. One is, like you said, not using plan mode enough. This is just super important. And I think this is people that are kind of new to agent of coding. They kind of assume this thing can do anything. And it can't. It's not that good today. And it's going to get better. But today, it can one-shot some tasks. It can't one-shot most things. And so you kind of have to understand the limits. And you often understand where you get in the loop. And so like, something like plan mode, it can like two, three acts success rates pretty easily, if you like land on the plan first. Other stuff that I've seen power users do really well is companies that have really big deployments of quad code. And now, luckily, there's a lot of these companies so we can kind of learn from them. Having settings JSON that you check into the code base is really important. Because you can use this to pre-allow certain commands. So you don't get permission prompted every time. And also to block certain commands. So let's say you don't want web fetch or whatever. And this way as an engineer, I don't get prompted. And I can check this in and share it with a whole team. So everyone gets to use it. I get around that by just using dangerously-skit permissions. [LAUGHTER] Yeah, we kind of have this there, but we don't recommend it. It's like it's a model. It can do weird stuff. I think another kind of cool use case that we've seen is people using stop hooks for interesting stuff. So stop hook runs whenever the turn is complete. So like this is in did some tool calls back and forth with whatever. And it's done. And it returns control back to the user. Then we run the stop hook. And so you can define a stop hook that's like, if the tests don't pass, return the text, keep going. And essentially, it's like you can just make the model like keep going until the thing is done. And this is just like insane when you combine it with the SDK and this kind of programmatic usage. You can-- this is a stochastic thing. It's a non-deterministic thing. But with scaffolding, you can get these deterministic outcomes. So you guys started this CLI paradigm shift. Do you think the CLI is the final form factor? Are we going to be using Cloud Code in the CLI primarily in a year or in three years? Or is there something else that's better? I mean, it's not the final form factor. But we are very focused on making sure the CLI is the most intelligent that we can make it and that it's as customized as possible. You can talk about the next form factors. Yeah, I mean, cast asked me to talk about it because no one knows. This stuff's like, it's just moving so fast, right? Like no one knows what these form factors are. Like right now, I think our team is in experimentation mode. So we have CLI. Then we came out with the ID extension. Now we have a new ID extension. That's like a GUI. It's a little more accessible. We have @ClawedInGithub. So you can just @ClawedItAnywhere. Now there's @ClawedOnWeb and on mobile. So you can use it on any of these places. And we're just in experimentation mode. So we're trying to figure out what's next. I think like if we kind of zoom out and see where this stuff is headed, I think one of the big trends is longer periods of autonomy. And so with every model, we kind of time how long can the model just keep going and do tasks autonomously? And just, you know, in dangerous mode in a container, keep auto-compacting until the task is done. And now we're on the order of like double digit hours. I think it's like the last model is like 30 hours. So like this. And you know, the next model is going to be days. And as you think about kind of paralyzing models, there's kind of a bunch of problems that come out of this. So one is what is the container this thing runs in? Because you don't want to have to like close your laptop. I have that right now, because I'm doing a lot of Disp. I don't know, I've only heard, I've only read it, but DSP wire Disp prompt optimization. And like it's on my laptop. And it's like, I don't want to close it. I'm like in the window like with my laptop open. Because I'm like, I don't want to close it. Yeah. Yeah, that's right. Yeah, we like visited companies before like customers. Everyone's just like walking around with their like plot codes. Is this right? So I think like one is kind of getting away from the mode. And then I also think pretty soon we're going to be in the mode of like, clouds monitoring, quads. And kind of, I don't know what the right form factor for this is. Because as a human, you need to be able to inspect this and kind of see what's going on. But also it needs to be quad optimized. Where you're optimizing for kind of bandwidth between like the quad to quad communication. So my prediction is terminal is not the final form factor. My prediction is there's going to be a few more form factors in the coming months, you know, maybe like year or something like that. And it's going to keep changing very quickly. What do you think about, you know, I teach a lot of quad code to a lot of every subscribers. Thank you. You're welcome. Doing your work for you. And I think that like one of the big things is just the terminal is intimidating. And just like being on a call with the subscribers, being like, here's how you open the terminal. And you're allowed to do this, even if you're non-technical. It's like a big deal. Yeah, I-- one of the people on our marketing team started using call code because she was writing some content that touched on call code. And I was like, you should really experience it. And she got like 30 pop of some on her screen where she had to accept various permissions because you'd never use a terminal before. So I completely see eye to eye with you on that. It's definitely hard for non-engineers. And there's even some engineers we found who aren't fully comfortable with working day-to-day in the terminal. RVS code GUI extension is our first step in that direction because you don't have to think about the terminal at all. It's like a traditional interface with a bunch of buttons. We are working on more graphical interfaces. So call code on the web is a GUI. I think that actually might be a good starting point for people who are less technical. Yeah. There was this magic moment maybe a few months ago where I walked into the office and some of the data scientists had anthropic right next to the quad code team. And the data scientists just had quad code running on their computers. And I was like, what is this? How did you figure this out? I think it was Brandon, who was the first one to do it. And he was like, oh, yeah, I just installed it. I work on this product so I should use it. And I was like, oh, my god. So he figured out how to use a terminal and knows, yes, he hasn't really done this kind of workflow before. Obviously, very technical. So I think now we're starting to see all these kind of code adjacent functions, people use quad code. And yeah, it's kind of interesting. From a latent demand point of view, these are people hacking the product. So there's demand to use it for this. And so we want to make it a little bit easier with more accessible interfaces. But at the same time for us, for quad code, we're laser focused on building the best product for the best engineers. And so we're focused on software engineering. And we want to make this really good. But we want to make it a thing that other people can hack. So sometimes quad code will write code that's a bit verbose. But you can just tell it to simplify it. And it does a really good job. Interesting. And so how and when are you doing that? So you're using a slash command or you're-- I just say it. I just say simplify it. Like sometimes you're like, hey, this should be a one line change. And all right, five lines. You're like, simplify it, and it understands immediately what you mean, and don't fix it. Yeah, I think a lot of people on our team do that too. It's interesting. Why not then, if you're saying that all the time, why not then push that into a slash command or the harness or something like that to make it just happen automatically? We do have instructions for this in the quad empty. I think it impacts such a low percentage of conversations that we don't want it to over rotate in the other direction. And then the reason why not a slash command is because you actually don't need that much context. I think slash commands are really good for situations where you would otherwise need to write two, three lines. But for some-- like even for a plan mode, you actually can use a few words for some time. But it actually takes two or three lines to capture the entirety of what you want in plan mode. For simplify it, you can just write some fly it and it gets it. Yeah, yeah, that makes sense. Cool. Yeah. OK, now we can-- [LAUGHTER] That's interesting. Yeah. But this stuff-- it still feels disso early. Yeah. We were talking about before they're recording about where are we on the adoption curve and it's still-- The house in curve or whatever. You don't know what we're going to do with that term. And it just feels like we're first 10% still. This stuff is going to change so fast. It's going to keep changing. Even when I talk to researchers outside of Anthropic who abuse code code, they also get stuck on things like this. Not realizing that they can just tell the LLM to simplify it. And I think that just goes to show that even for people who are working in this industry, they don't always realize that you can just talk to the model. That's the thing is, I think that there's this underlying expectation that using AI shouldn't have to be a skill because it just does whatever you say. And you're like, well, I mean, whatever you say is going to matter for what it does. So if you can say things better, it's going to do better. [LAUGHTER] I mean, it changes with every model though. That's the hard part. Prompt engineer was a job, and now famously, it's not a job anymore. And there's going to be more jobs that are than not jobs anymore of these little micro skills that you have to learn to use this thing. And as the model gets better, it can just interpret it better. But I think that's also like, for us, this is part of this kind of humility that we have to have building a product like this that we just really don't know what's next. And we're just trying to figure it out kind of along with everyone else. We're just here for the ride. And that's why it's cool that you're building it for yourself, because I think that's the best way to know that is just like-- and this is what we do, too, is like you're sort of living in the future, you're using it all the time. And it's pretty clear what's missing. You're like, I just want this thing. And you can just do the next thing rather than being like, let me ask some enterprise product manager at some gigantic company, like what kind of AI feature do you want? And they're like, I don't know. Put a little chat bot on the side of my IDE. And you're like, OK. Yeah, this is like the luxurious thing about building DevTools, right, your own customer. I think it's also really a unique thing about AI, because it sort of reset the game board for all software. So we have Kora, this email assistant. We have Sparkle, which organizes your files. And it's like anything that you do for something that you want to use on your computer. If you're building it with AI, there's a good chance that hasn't been done before, because the whole landscape has been reset. And so it's a uniquely exciting time to build stuff for yourself. Totally. I think it totally opens the point you feel, too. It's like any individual can now build an app to fill their need and then distribute it to everyone else. I've been prototyping all these random pet projects. But I just moved into a new apartment, and it's empty. And so I've been building this shopping advisor assistant unlike the Quad Agents SDK, because who has time to read all the reviews and look at all the options and find their pricing and everything's really hard to discover. And so it just asks me a bunch of questions, and I tell it what I want. And it shows me a bunch of photos of different sofas and options and what people say online. And then I tell it what I don't like. And it literally feels like working with a shopping assistant. It's been really cool. I also have my little email response agent that like drafts responses for me, but I don't use email that much, so-- Oh, and I knew it wasn't you responding. [LAUGHTER] That's my seven days delay. [LAUGHTER] The agent's just doing a very thorough job. The agent SDK is cool, too. Yeah, the agent SDK is cool. Yeah, it always just feels amazing, like how much we're able to build with such a small team. Yeah, so I feel like there's-- The other thing that's really cool is that I think people are just shifting their mindset from docs to demos. Like internally, our currency is actually demos. It's like, you want people to be excited about your thing. Show us like, show us 15 seconds of what it can do. And we find that everyone on the team now has this kind of indoctrinated demo culture, for sure. And I think that's better because there's a lot of things that you might have in your head that if you're a great writer, maybe you could figure out how to explain it. But it's just even then it's just really hard to explain. But if someone can see it, they get it immediately. And I think that's happening for product building, but it's also happening for all sorts of other types of creative endeavors, like making a movie, for example. You had to pitch it. But now you can just be like, I made this sort of video, and you can kind of see the glimmer of the thing you're trying to make for very cheap. And so that means you don't have to spend time convincing people as much, you're going to be like, here, I made it. Yeah, and also as a builder, you can just make it, and then make it again, and then make it again. Until you're happy. I feel like the flip side is like, you used to make a doc, or way forward something, or I would draw stuff and sketch or figmar or whatever. And now we'll just build it until I like how it feels. And it's just so easy to get that feeling out of it now. And I think it's like, you could see it visually before, or you could describe it in words, but it's like, you could never get the vibe. And now the vibe is really easy. Yeah. And you build plan mode like three times. Yeah. Yeah. Because of this. Like you built it, and then you threw it out and rebuilt it, and then threw it out and rebuilt it. Yeah, we're like 2Dos, like Sid built a original version, also like three or four, he built like three or four prototypes. And then I prototype maybe like 20 versions after that, like in like a day. Yeah, I think this is like a lot of pretty much everything we've released, there was at least a few prototypes behind it. How do you like keep track of and carry forward the things you learn from prototype to prototype? And especially if it's like, you know, some one person is prototyping it, and then you're like, I'm going to take it over, I'm going to do 20 more. Like how do you maximize what you get out of that? You know, it's like, there's maybe a few elements of it. One is the style guide. So there's like some elements of style that we discover. And I think a lot of this is like building for the terminal. Or like, we're kind of discovering a new design language for the terminal and kind of building it as we go. And I think some of this you can codify in a style guide. So this is our quantum D. But then there's this other part of it that's like kind of product sense, where I don't think the model totally gets it yet. And I think maybe we should be trying to find ways to like teach the model this kind of product sense about like this works and this doesn't. Because in product, you want to solve the person's problem in the simplest way possible and then delete everything else that's not that and just get everything out of the way. So you kind of you, you align the product to the intent as quickly as possible. And maybe the model doesn't totally get that yet. Yeah, it's never, it doesn't really feel what it's like to use quad code. Like the model doesn't use quad code. And so I think like when, you know, a quad code can like test itself and it can kind of use itself. And like we do this one developing. And it can see like UI bugs and things like that. I don't know, maybe we should just try prompting it though. It could like, honestly, a lot of the stuff is as simple as that, like when there's some new idea, usually you just prompt it and often it just works. Maybe we should just try that. A lot of the prototypes are actually the UX interactions. And so I think once we discover a new UX interaction, like shift tab or auto except, I think, for us to figure it out. That's eager actually. Oh, eager. Yeah, we went back and forth. And we went back and forth. We went back and forth. We did like doing prototypes for like a week. Yeah, shift tab felt really nice. And then one of the now current plan mode iteration uses shift tab because it's actually just like another way to tell the model how agentic it should be. And so I think as more features use the same interaction, you form a stronger mental model for what you go where. Yeah. All right, like thinking, I think, is another really good one. First, we were before we were we squat code. Or maybe it was like the first thinking model was like 37. I forget what the first one was. But it was like, it was able to think. And we're like brains from like, how do we like toggle thinking? And then someone was just like, what if you just ask the model to think in natural language? And it knows how to think. And we're like, OK, sweet. Let's do that. And so we did that for a while. And then we realized that people were accidentally toggling it. So they were like, don't think. And then the model was like, oh, I should think. I just started thinking. And so we had to kind of like tune it out. So you know, don't think, didn't trigger it. But then it still wasn't obvious. But then we made a UX improvement to like highlight the things that started. And I was like, that was so fun. And it felt really magical. When did you alter think? It's like rainbow. Yeah, yeah. It was OK. And then with Son of 4.5, we actually find like a really, really big performance improvement when you turn on extended thinking. And so we made it really easy to toggle it. Because sometimes you want it. Sometimes you don't. Because you kind of, for a really simple task, you don't want the model to think for like five minutes. You want it to just do the thing. And so we used tab as the interaction to toggle it. And then we unshipped a bunch of the thinking words. Although I think we kept alter think just for our extent of ethical reasons. It was such a cool UX. Interesting. Do you think there's some new metric that's about what you deleted? And I think programmers have always felt like deleting a bunch of code feels really good. But there's something about-- because you can build stuff so fast, it becomes more important to also delete stuff. I think my favorite kind of diff to see is a red diff. This is the best. And whenever I see one, I'm like, yeah, forget on. Another one, another one. But it's hard because anything you ship people are using it. And so you've got to keep people happy. And so I think generally our principle is, if we ship something, we need to ship something even better that people can take advantage of that matches that intent even better. And yeah, I think this is kind of back to how do you measure quad code and the impact of it. And this is something like every customer asks us about. And I think in turn away at Anthropic, I think we doubled in size since January or something like that. But then productivity per engineer has increased almost 70% in that time. Measured by. I think we actually measure it in a few ways. But kind of peers are the simplest one and the main one. But like you said, this isn't capture the full extent of it. Because a lot of this is like making it easier to prototype, making it easier to try new things, making it easier to-- these things that you never would have tried because they're way below the cut line. You're watching a feature. And there's this kind of wish list of stuff. Now you just do all of it because it's so easy. And you just wouldn't have done it. So yeah, it's really hard to talk about it. And then there's this flip side of it, where more code is written, so you have to delete more code. You have to code review more carefully and automate code review as much as you can. There's also like an interesting new product management challenge because you can ship so much that you end up-- it ends up not feeling as cohesive because you could just add button here and a tab there. And a little thing here, it's much easier to build a product that has all the features you want, but doesn't have any sort of organizing principle because you're just shipping lots of stuff all the time. I think we try to be pretty disciplined about this and making sure that all the abstractions are really easy to understand for someone, even if they just hear the name of the feature. We have this principle that if we've brought to the team that I really like where we don't want a new user experience, everything should be so intuitive. They just drop in and it just works. And I think that's really set the bar really high for making sure every feature is really intuitive. How do you do that with a conversational UI? Because when there's not a bunch of buttons and knobs, and it's just a blank text box to start, how do you think about making it intuitive? There's a lot of little things that we do. We teach people that they can use the question mark to see tips. We show tips as quad code is working. We have the change log on the side. We tell you about, oh, there's a new model that's out or like we show you at the bottom. We have a notification section for thinking. I think there's just like subtle ways in which we tell users about features. I think the other thing that's really important is to just make sure that all the primitives are very clearly defined. Like hooks have a common meaning. In the developer ecosystem, plugins have a very common meaning in the developer ecosystem, and just making sure that what we build matches the average developer would immediately think of when they hear that. There's this also this progressive disclosure thing. Any time in quad code, when you run it, you can hit Ctrl-O to see the full raw transcript, the same thing the model sees. And we don't like to show you this until it's actually relevant. So when there's a tool result that's collapsed, then we'll say use Ctrl-O to see it. So we don't want to put too much complexity on you at the start, because this thing can do anything. I think there's this other new principle which you've just started exploring, which was like, the model teaches you how to use the thing. And so you can ask quad code about itself, and it kind of knows to work up its own documentation to tell you about it. But we can also go even deeper. For example, slash commands are a thing that people can use, but also the model can call slash commands. And maybe you see the model calling it, and then you'll be like, oh, yeah, I guess I can do that too. Yeah, yeah, yeah, yeah. Interesting. How has it changed when you first started doing this? Quad code was this sort of singular thing, the singular way of thinking about using it through a CLI. And other people had stuff like this, but it felt like this shift. And now there's a whole landscape of everyone is like going CLI, CLI, CLI, CLI. Like how has that changed? How you think about building, how it feels to build, and how are you dealing with the sort of pressure of the race that you're in? Think for me, imitation is the greatest flattery. So it's awesome, and it's cool to see all this other stuff that everyone else is building inspired by this. And I think this is ultimately the goal is to kind of inspire people to build this next thing for this just incredible technology that's coming. And that's just really exciting. Personally, I don't really use a lot of other tools. So usually when something new comes out all, well, maybe just try it to get a vibe. But otherwise, I think we're pretty focused on just solving problems that we have and our customers have and kind of building the next thing. Cool. Sweet. I love this part of the interview, too. Do we answer all your teams' questions? Oh, do we get through all my teams' questions? Let's see. I think we did. I'm curious also how you would answer the unshipping question. Because also, if you're doing this AI-driven development, you should be a lot. You have a small team, so it's a lot of operational load. The reason I asked that is because I don't think we do a good job of that. And I have this feeling that some of the products are a little bit messy because of that. And I think particularly for Cora, there's just a big product surface area, and I can do a lot of different things like it. We have an e-moss system, so you can ask it. Tell me about the trip I'm taking, and it'll go through all your emails and summarize the trip. Or we have this feature that it automatically archives any email that you don't need to respond to immediately. And then, twice a day, you get a brief that summarizes all the stuff that you probably need to see, but you don't need to actually do anything with it. And you just scroll through it, and you're done. And there's just all this complexity that around, for example, how are emails categorized? And now we have a whole view of how all these categorization rules and you can order them in whatever, but it's just complicated and hard to communicate. And I want to retain a lot of the power and flexibility, but also you can't look at a screen and be like, I have no idea what's going on. This is way too complicated. So I'm just like, I'm processing all that stuff. So the kind of deletion, unshipping idea feels like an interesting cultural principle that we haven't really explored. Yeah, it's really hard. I think there's like a social cost to it, too. We're like, you kind of want to be the person who tells your co-workers to unship there. Yeah. It's definitely tricky. It's more than just the code. I definitely run this at Instagram, honestly, because I think Facebook does a terrible job at unshipping. And we have this problem where every time we-- I think even like unshipping pokes was like really spicy, because there's a bunch of these old timers. They're like, no, pokes, you're never going to take it away. But like, if you look at the data, no one really uses it anymore, but for sentimental reasons, they were kind of tied to it. And so like, for Facebook, it always-- maybe nothing ever got unshipped. It always got moved to like a secondary place, like an overflow menu, a summer that no one looks at, like a graveyard. And I think Instagram was just very principled. There was like, very strong product and design point of view. Those like, if this thing isn't used by like half of people, 50% of wow or whatever, we're just going to delete it and deal with it. And then we'll figure out some next thing that's used by more people. I love it. Well, thank you. This is amazing. I'm really glad I got to talk to you and keep building. Thank you for having us. Yeah, thanks. [MUSIC PLAYING] Oh my gosh, folks. You absolutely positively have to smash that like button and subscribe to AI and I. Why? Because this show is the epitome of awesomeness. It's like finding a treasure chest in your backyard. But instead of gold, it's filled with pure, unadulterated knowledge bombs about chat GPT. Every episode is a roller coaster of emotions, insights, and laughter that will leave you on the edge of your seat, craving for more. It's not just a show. It's a journey into the future with Dan Shipper as the captain of the spaceship. So do yourself a favor. Hit like, smash subscribe, and strap in for the ride of your life. And now, without any further ado, let me just say Dan, I'm absolutely hopelessly in love with you.