I got the grilig of my life in building a pretty simple API endpoint using the Grilby skill. It asked me 35 questions. I kid you not. It was intense and annoying. And it forced me to think more.
Today's guest is the creator of his popular skill, Matt Pocock. Matt is a developer turned educator, well-known first total types of series, and now for his AI skills and educational videos. Today we cover Matt's unusual path of detect after years of being a voice coach and building his own DIY coaching software. Matt's popular skills, Grilby, Wayfinder, and why these skills became so widespread. Taking inspiration from decades old programming books to build better software with AI and many more.
If you want to understand which software engineering fundamentals approach is remaining very useful when working with AI agents, this episode is for you. This episode is presented by Triple Buffer, Vector and Full-Tech Search built on object storage. It's fast, cheap, and extremely scalable. This episode is presented by Linear, and I wanted to take you back in time to remind you how we used to get work done. Back when every line of code was written by an engineer, like you or me, a tracker's job was to keep people in sync without slowing people down.
Linear was built to be fast and low-friction, and you could tell. In last year's the Pymatic Engineer Survey, Linear was the most love tracker tool, and Jira the most disliked one for its sluggish performance. And Data coming from the Pymatic Engineer audience showed how Linear started to gain traction against existing tools, especially as startups and midsize companies. And since then, Linear grew up. They added all the stuff that larger companies need to manage work, projects, initiatives, roadmaps, and customer requests.
And large companies started to switch. For example, Helsker company Oscar helped move 600 engineers from Jira to Linear. Opening, I started with 100 seats and moved all 3000 staff without any mandate. Coinbase, Cash app, Brex and Ramp are all on the near. Many of them saw Linear as a way to consolidate a single tool that brings planning and building together.
So now, let's fast forward to today. When you have AI agents inside a company, those agents need context to work well. They need access to things like specs, customer requests, history. Oh wait, these are all already in Linear. So when agents arrived, Linear became the ideal context.
Today, 80% of enterprise works based in Linear have adopted agents. You can use agents like Codex, Cloud Code, Linear Agent, or your own agent. Coinbase and Ramp both build their own internal agents and describe Linear as the place that their agent goes and picks up the context before starting work. See how it works at Linear.Appslash, Pragmatic. Map is great to have you have a podcast.
Great to finally be here. I'm a huge fan of what so many of these. I feel like this is like the tiny desk of being a software engineer. You know what I mean? This is big stuff. So I'm glad to be here.
And it's also great to reconnect. Because about a year ago, we had lunch at Microsoft Build as well, which was really fun. But now it's going to jump into this. And with this, I wanted to ask about your background. You, unlike many people in tech and on this podcast, you didn't start out to study computer science, right?
Absolutely not. So for six years before I became a developer, I was a voice catcher. I was a singing teacher working in London and working next to where I went to university. I was teaching accents. I was teaching singing.
I was teaching voice. I did a masters in it. I spent a lot of time thinking that was what my career was going to be. You know, I didn't have any inkling of tech didn't sort of think about it at all. I sort of ran my own website and stuff.
But yeah, so I did that for a long time. And it's been an extremely important influence on my life. And I think my personality as well. Can you get a bit of a voice to come from? And what do you do as a voice coach?
Who are people who came to you for help and what kind of help? So I started as a singing teacher. I was in a band and stuff university. I sort of had a bit of experience doing singing. And so I sat at my own company, kind of university and doing that stuff.
And it was people who just wanted to sing better, who wanted to use their voice for choirs, who wanted to do it. It's a hobby. It wasn't anything particularly professional. And I went into a masters in it. And I started going to drama schools to teach people Shakespeare and stuff.
And getting people in who wanted to do public speaking. I did a couple of big gigs for consulting companies, going and teaching them how to deliver speeches and how to talk better. It was wild. And the reason I got out of it was because I realized in order to do it at a decent level, you had to live in London. I didn't want to live in London. I tried it for like two years.
I just hated it. I hated it. I didn't grow up in London. I wanted to get back to the countryside and where I was from. And that's what I did.
And so I learned how to be a developer. I was essentially self taught in order to have something I could do remotely. So basically you were looking at professions that you could do from outside of London that had a career or perspective. Or a future. Exactly.
And I was, I sort of taught myself how to build stuff and just sort of build basic stuff in JavaScript because I was interested in making my lessons better for my students. So I'd actually made sort of a little flashcard apps. I was working like the first app I ever built was the most ambitious thing I've ever attempted. It was like a web audio analyzer. So I could analyze the spectrogram of your voice to see which resonant frequencies were happening, whether you were, whether you're T1 and T2 were like properly balanced and things like that.
Extremely in depth ran terribly, but actually made my lessons that little bit better. And so I was doing pretty hard core stuff terribly straight away. And I realized, okay. I started looking at job postings. I thought, well, I could do a bit of JavaScript.
I could do a bit of SaaS, I could do a bit of this and bobs. And I just jumped into it. I, you know, quit my job, had a couple months off and eventually got a job. This was about 2017 where it was a little bit easier to get a job in the UK than it is now. And I just went from there.
I guess as long as you were also lucky because I was the peak. That was the time where the man was so high for engineers that people had a boot camps with a few months of experience. And I think people got all the chances from a lot of places where they drive and the motivation and the smarts, right? Yeah. And because I had this history of talking to people, that was an unbelievable advantage, right?
I could actually go into an interview and sound like a reasonable person. Instead of someone who comes straight from a CS degree who maybe didn't have those skills. So I had this bizarre ability of having zero technical knowledge or very little. In the beginning. But the ability to explain technical knowledge to people, right?
And so that, basically, all I needed to do was increase my technical knowledge a little bit. And I was very passionate about it. And that increased quite quickly. And then it was sort of seemed to be an unfair combination because I just rose through the ranks very quickly in various different companies. And I don't know.
I felt different from the other software developers I was working with. And how did you step up on the ladder? So like you decided, I'm going to do this. You taught yourself. You went to some interviews.
You got to give. I'm assuming I'm a small company, right? Yeah, a tiny company with a couple of really inspiring software developers who worked there, basically a guy. I wouldn't say something because he likes his own. Basically, a guy who lived in Sandals who lived in a canal boat for a long time, like a long hair, proper hardcore.
It was around the time that Microsoft bought GitHub. I remember him coming in almost in tears. Microsoft haters. Yeah, absolutely. Yeah, absolutely.
I remember first thing he got me to do was set up Sentos 6 on my Windows PC. It's pretty hardcore. It's really hardcore. I mean, it's a really, really, really, really, really wonderful guy and someone who taught me a lot straight away. And so I basically, that company ran into financial troubles and so I had to move to an agency pretty quickly.
And I got a higher job there. From then on, I moved to another agency and then another agency. So I'm just sort of bouncing around to agencies. And then I was working on open source, which is kind of the next part of the story. And with the agencies, what tech stack were you using at the time?
It was Tech Script. It was React. I was already back then. Well, I was pretty hardcore on Tech Script already. I was in my second job.
I think I was doing presentations on how important Tech Script was. We were working for an automobile manufacturer building a learning management system, right? Classic, boring agency stuff, right? And the front end team at that time was pretty small and we had a backend team in Portugal, right? So classic front end back end split.
The backend team were racing ahead. At the time I joined the front end team was really slow. We had a ton of bugs. The backend team kept changing their contracts without telling us. And we thought we need something to link us up a bit better.
Tech Script felt like the obvious thing. And once we shipped it, our velocity just went, you know, we were fast on the backend team. And eventually they took people off our team because we were so quick. So yeah, that was my history with Tech Script. That's kind of my origin story with it.
How did you get into open source? Was it at work? Was it on the side? It was. I would have been constantly playing around with open source on the side.
And I was interested in different things. By then I was into Twitter. I was sort of looking at people online thinking that someone I want to emulate someone I want to look at. And there was a guy who crossed my radar called David Koshit, who's the State Machine and Tech Script Guy on Twitter. A lovely, lovely guy.
I know a lot of, you know, my career to him really. And I was working on a project, this is, I think, in my fourth job where we needed a State Machine. It was a very complex application where you were on a video call with someone and you could navigate around a house in real time together using some sort of Matterport integration. And there was a lot of linking up that needed to be doing it across the network boundary. A lot of complicated States.
And so I used a library called X States at the time, X States version four, I think. That was a resounding success. And so I wondered, okay, how can I make this more type safe? And so I started to sort of build some tooling around it, have a fiddle built up to see either constructed around it. And that got me the attention of David and I became a member of the X States Quoting.
So I started contributing issues started having discussions about the future of the library. And it brought me into contact with just a level of developer that I'd never seen before. David and another guy called Mateusz Bzinski, who would embarrass Ray Conquiter. These are the most talented developers I've ever seen. Like this is another level.
And eventually David wanted to form a company out of it. He wanted to make a big bet on state charts and visual sort of programming as the future development. They got some funding and that was my first job where I was being paid American money, basically. And it was a huge set up for me. Which as we know, it's quite different from a European or local, even UK company paying.
Because I also covered some of the tri-model nature of software and new composition. Yeah, US companies especially in Europe and also in the US, they think about composition, different value generated differently, right? Totally. It changed my life, you know, in terms of the way I was thinking about money and the way I was thinking about flexibility. And it meant I was working on something I was passionate about.
And I started while I was there doing a bit more advocacy for it. Because obviously the company is very small. I was doing a lot of development, but also I wanted to be an advocate for it because I believed in it, you know. And I still think state charts are incredibly primitive for certain kinds of work. I've sort of rode back a little bit on my belief of them, especially in the AI age.
But I was doing a bit more of that. And that got me the attention of a couple of guys at Versell. Because Versell at that time, Lee Robinson was the guy in charge of development education there. They had this incredible team, Dilba De La Vere, Lidi Halli, both of whom are now at all code, Lee himself. And I was working, I got a job there under Jared Palmer as like the, yeah.
Wow, that's your Palmer. That's your Palmer. Yeah, he's a really good mate actually. And he's the one who later moved to GitHub, he started our spearhead of SAC, DIFs or SAC PRs. And now he's at commission.
Yeah, he went into GitHub, ships, stack DIFs, left, refuses to elaborate. And he's now at commission, exactly. Yeah, but he's also in this relationship, yes. He's a great guy. And I worked under him for not very long at Versell.
So I was only there about three months. From there, I had, I got a funny contract at Versell because I'd already been floating this idea of sort of touch script and thinking about touch script and thinking about maybe making educational material. I had this urge while I was at stately the exit company to teach stuff. You know, I've been teaching for six years before. You know, I've been not teaching for four or five years in that point, maybe six years.
And I thought, I need to get back to this. Like, I miss it. You know, and I love making stuff. I love making content. I love teaching people.
And so that's what I started doing. I started doing it for advanced types. I got in contact with a lot of crazy typing tricks, a lot of really advanced types of stuff while I was trying to force x8 to be types safe, very, very hard job. I think of mostly a possible job. And so I made a couple of tips.
I made these two minute tips, post them on Twitter and they just went, you know, in a way I'd not felt before. And I realized, okay, there's a market here. And so one Sunday, I just made like 13, 15 of these two minute tips. I just queued them up next few weeks and my follow account went from, you know, four thousand to ten thousand to ten thousand. You know, it's just, you know, you've only felt that there was huge interest in this, right?
And the massive wave of something was, you know, some combination of the way I was speaking, the material I was delivering, that was clicking in a way that I'd not felt before. And that's only really happened twice in my career. So I was already floating idea, of course. And I knew I could do it well. I knew I could do really great course if I just had the right audience and if it clicked.
And so I went into Versell, I got a contract there for only three days a week for three months initially, which is very unusual. Is that what you wanted or this is like how, you know, Versell was probably testing the water, see how it goes? Versell, I want to be full-time straight away. You knew that there's a set of things. So let me kind of hedge my bets if I'm able to do, right?
Versell was this weird backup to what I, which is wild to me. For most people, this would be the dream job. I know. So it's a little embarrassing to say because obviously it's so many people's dream job, but I went into it going, okay, I need a stable nine to five for three days a week while I test this other thing out. But I mean, just to be fair, I think this is sensible, right?
Like at this point, if we just go back to where you are, like you've been a voice coach for a good part of your career, let's say six years. And let's say now for five years, you've been building software, you love doing it, you think you're good at it, you think you might be able to teach, but who knows, right? And at that point, saying, all right, let me take a gamble and like do this thing that might or might not work out. Whereas if you can pull it off when you have something stable and it gets traction, all that's different, right? You know, a lot of engineers have aspirations, especially because we're suffering, you can work remotely, you can take your idea, build a company and they're thinking, all right, should I just plunge, quit my job, should I quit my job?
So like, in some ways, I guess this is one model that is kind of unique and if you're able to pull it off. I mean, it was the most bizarre thing because it became very clear very quickly that I couldn't stay up to sell, basically. So we had about two months into my work of a cell. I was there actually over a very, not too much time because I was there when they released a backpack. I actually wrote some of the documentation, the initial documentation for a backpack and that's on the team.
Which was a lot faster built as to write. It was a build system, essentially at the time they were trying to rival Webpack, what they were working with. And I was there initially when they were building docs, I flew out to San Francisco, I was there for the next JSCON, for the announce date, big, you know, really fun experience. And I was there with everyone while they're getting everything ready for it. And so I do that and already in the back of my head I'm thinking, I've seen the newsletter from my top script stuff creep up, understand, okay, there's something really big here.
And when I made a pre-release sale, that just went crazy. I was earning, let's say, X in a VSL and that was like 30, 40 X or something. It was immediate. And X at VSL was already a really, really good composition. Absolutely. Very, very happy with that.
So I just, it was obvious. There was no other decision I could make. I loved working at VSL. I would probably go back at some point, but I just couldn't say. So I had to do this thing.
And then tell me about total type scripts. So you start to build two days a week on the weekends and then you did this pre-release sale. What's the... I try never to work on weekends, basically. I'm extremely radical about this.
I think it's something I mostly fail at because I'm quite an obsessional person. I like trying to make something work, but I'm not one of these guys who's doing, what's the SF thing where people go like 99 to 99, 6 days? It turns to my stomach. I just hate that stuff. I am trying to, with everything I do, build a lifestyle and build a life where I can spend most of it with my family.
That's my goal. And so just to pre-fix that, without all of my decisions after that, hopefully make more sense in that light. So Tental Times scripts I was working with. I got called Joel Hooks. Joel Hooks is extremely funny, extremely influential on me.
I've worked with him now for four years and he came up with Eket. He's worked with Kenzie Dodds on his courses, extremely successful course creator in the background. I basically reached out to him and I said, we'd like to make this course and he said, hell yes. And we went from there. Straight while I'm at the cell and also working with Joel.
We do this pre-release and as I said, just goes nuts. I've got a fully commit to this. We get to, I think about January 2023, February 2023, and we release the full course. I think I need to look at the charts from around that time, but it reaches seven figures extremely quickly. That's the revenue split between me and Joel's courses and expenses and that.
In terms of raw revenue, it was extremely exciting. But the seven figures, that's one million dollars, which is incredible milestone, right? Which is nuts. And I'm not changing. I realize, okay, I can wake up in the morning and this money's still going to come in.
This is something that I dreamed about for a long time when I was a scene teacher as well, making material that I could sell online. This is something I've been aiming for for a long time, high leverage work where I can do the work and then step back and go back to my family. And for the next couple of years, I worked on time scripts, expanding the course, selling a couple of supplementary courses. And that's basically where total time script was. And so that was the main portion of my success in the last four years has been total time script in building now.
Yeah, and total time script has been very inspirational, especially that you openly shared a big milestone when it hit two and a half million dollars of total revenue, which again, I think for many software engineering, that is, of course, we know this is before revenue share and their expenses involved as well. But it's something that is pretty clearly a higher earning potential than many great software engineering jobs. Not necessarily all of them, especially when we're looking at the US and some of the AI labs and whatnot, which was probably an exception. But the fact that there is a market and a business to be made of what I feel is a bit kind of an honest model in a sense of like, hey, I created this thing, people pay for it because they want to learn and they hopefully get value from it because otherwise they would pass for refund, right?
Exactly. We do like a very extended refund policy. I don't tend to want to accept any other forms of money. Either I don't like necessarily doing sponsored content, I'm not going to say never, you know, but I don't really, I've got to get up sponsors page, but I'm really trying to take it down. For a long time, I'm trying to take it down.
I like the idea of just being somebody who has, okay, these are the products you can buy from me, this is how you can support me. And hopefully this gives you, you know, 10x in terms of returns because this is a very lucrative industry, right? Like, and a lot of people have education budgets that they can spend. And if you want to spend some of your education budget on me, that's basically my model. And a lot of that money, a lot of the people taking the course, this is from people's education budgets, you know, this is companies coming and spending big on education.
And that was a market that Joel was very keen pushing me towards and realizing this, you know, I didn't have much of a sense for how much money was sloshing around in the industry, especially in that age and honestly still. But the fact that it just hit that milestone so quickly within a couple of years. I mean, life changing. Well, I mean, this sounds like an amazing story and it could be fairytale. I mean, we're like, you keep creating educational content for the rest of your life and it's highly in demand, but then AI happened. And as we know, it's changing a lot of how we work, how we find information, for example, like, I don't Google that much.
I actually work with AI agents or deep researchers, some of those things. Her story about educators, online educators were saying that their revenue and market share mind share is just falling down because people might not want to sit through courses or sessions when you can just turn to the bot. How did you see AI impacting the industry, how people learn and also your business as also you as a teacher? It's complicated because AI has changed the game. Right. It's changed how important knowledge is and specifically the types of knowledge that are important.
So when I'm teaching my courses, I sort of think of theirs being two layers. I'm teaching the syntax, obviously. I'm teaching the what's, but there's also the why behind it. And it's very hard to teach the why without touching on the what, if that makes sense. So the sort of medium is I'm going to teach you this syntax and maybe you might gather the sort of wisdom around it.
I'm teaching you knowledge, but I'm also trying to teach you wisdom. Knowledge is now very cheap to acquire. Very, very cheap. You can just look it up. You can build it up. I have a teach skill that can just take you and just teach you the knowledge that you need.
But the wisdom has gotten no easier to learn. It's still knocking about. You are still going to run into the same issues that you ran into if you didn't have that wisdom before, even with AI. So in terms of my revenue from Total Tactic, that's gone down. Obviously because I think people are not so interested in that material anymore.
And I think people teaching that material are going to find it tricky because, again, that knowledge is really, really hard to come by. The only way that I've been able to not necessarily survive, but it took me a long time to figure out where I wanted to be in the AI space. Because I'm not a researcher from OpenAI. I don't have the credentials to talk about this stuff really, especially in 2020. Late 2023 was when I started looking at it.
And I was initially making courses about how you put AI into applications. I thought, OK, I've been building applications for a long time. It makes sense that AI is changing things a bit there. It started to be clear to me that was the wrong bet to make. I wasn't sort of seeing the returns I was expecting.
And I didn't, like the material was good and I feel proud of it, but I didn't think I wanted to make more of it. And around December last year, which is a date that many people cite, we know, or as we call it, the winter break where everyone came out of the uphill. Yeah, exactly. The pizza break, right? The open-floor break when Open 4.5 is out, people have a lot of time off and they just start slamming it and they realize, wow, OK, things are really happening.
And it happens to me too. And I realized, OK, the AI is now good enough that you can delegate to it. You can actually make structures around the agents. And the agents can handle the knowledge, the syntax, the sort of, I call it the tactical stuff. And you can handle the strategic stuff the long-term thinking.
I use that a lot. I know you had John Aus to how on this podcast, I've been really wanting to chat to myself. Like he's a huge influence on me. And he talks about the difference between tactical programming and strategic programming. AI has largely eaten tactical programming in my view.
And it's up to us to handle the strategic. And I realized, OK, in the strategic layer, there's a course I can create that. I was looking at rough loops at the time. I'm sort of Jeffrey Huntley was building this really cool stuff where you can loop the agents and get it to follow these goals. And I thought, OK, there's definitely material here.
I just need to find a structure within which I can put it and organize how to fill around. And I remember the Ralph Loops, you also made a video that became very popular on YouTube, everywhere where you basically said, like, all right, here's how I created a Ralph Loops. Like here's I have a project that has a lot of like to do is an even a great job in willing that video on the show notes below where you said, like, right, here's how usually we would try to get agents to work, like do a plan upfront, and then implement the stuff like the kind of traditional top-down planning. And you're saying the problem is that as you're implementing or even the agents implementing it realizes, hang on, I need to do more stuff. And then I'll do modify the plan and then enter the Ralph Loops where you gave the structures that you use at a time.
There was like an empty file. It keeps it as as there. And it kind of like eats to it, but keeps adding. And it was actually a pretty eye open up to me. It was like, oh, I'm a way to think about how to do these agents.
Actually, the couple of years I spent sort of trying to put agents into applications was really beneficial there because when you try to build an app that contains an agent, you're always thinking about data flow. You're thinking about how the data is going to get in, what shape it's going to be, what priority, whether you're going to put it in the system prompt or user prompt, you're working at a lower level than usually gets to with the harnesses. And so when I got to working with harnesses, it felt like, oh, this just feels very familiar. I just need to, you know, where is the state going to live? How am I going to pass the state into the agents? What shape is that going to look like? How am I going to compact it or clear it?
And this was still pretty early days of cool, cool code about me. I've got five, six months at that point. And it just felt very natural. And from there, I just got obsessed with these, I suppose we were called loops now, but really they're just processes, they're sort of different ways of stringing agents together. It's kind of like the diagram. Like if you can draw arrows from one thing to the other, that you can call that a loop, especially when there's a port that goes back where you call the workflow or whatever.
I would call it a finite state machine. It felt very similar to the stuff I've been working on in X-Tates, which is process-based, which is state-based event-based sometimes as well, where you have an agent at the bottom there that's calling and then back at top. And that's, I started just to see really good results from that. And I would do these experiments where I would try sort of building out my process and I would build out a feature of, you know, I have a few apps that I work on to extend what I do with a custom video editor. I have a huge thing that, a huge code base, I have a few open source projects as well.
And I was just building these little loops and little pipelines. And I would sometimes just drop it and go back to what the default setup was. And I just noticed a huge difference. Like, I just felt, wow, okay, the stuff that I'm doing here is really setting me up for success. And I started thinking, what's the best way that I can distribute that?
How can I share that with other people better? And that's where I started landing on skills as the distribution mechanism for this stuff. These are the skills for AI bots, harnesses where you can typically define them and now you can once you install them, you can invoke them with a slash command. Exactly. Skills really, they're just a folder of markdown files that can sit in your computer somewhere.
And the agents can either invoke them themselves, so model invoke skills, or you can have skills that the agent doesn't know about, but you can also invoke them. So user invoke skills. And I started seeing these skillsets pop up everywhere, like superpowers and, you know, claw code plugins that you can install. I think GStack as well. I realized, okay, maybe I can distribute what I have this process as a set of skills and see what people think of it.
And initially, I just put it up and I was doing other stuff. I was working on the course and I check back in and I realized, oh, it's got more stars than anything else I've ever done. I've not even really talked about it. It's just sat there on its own word of mouth, I suppose. I've done a little bit of documentation, but really not much.
And it's just exploded already. So I thought, oh, okay, maybe I should put a little bit more work into this. Maybe I should talk about them. And I did a talk at, I think, where we met last, which was A.I. Engineer London about in April.
That talk was entitled Software Fund Mental Stillmatem. And that talk is now at 2.1.2 million views or something. And I mentioned the skillset. The skillset is now at 230,000 stars. It is now the second most starred skills repo in the world.
I think somewhere like 20th to 25th of the most starred repos of all time. Wow. What's going on? So this obviously a hunger for this. So this was the second time in my career, just like when I was putting out the little title script videos where I felt, wow, there's momentum here.
There's something happening. And so I felt I had to double down on that. The skills, how did you write them? Is this trying to capture your workflow, your understanding of what works with Asians? Not just right now, but of course you're thinking about state.
You're thinking about how you were integrating AI into application, which again, didn't take off all that much, but you learned. So is this kind of like Matt's workflow, Matt's way of what works for me? Yes. That's what it is. I tried to think, first of all, people are going to use these skills and they're going to tinker with them.
So how do I make the simplest set of skills that people can audit very easily? I'm trying to think, how do I maximize people picking these up and using them at work? So for instance, the grille me skill, which is the most popular one. I don't know if you've used that. I don't know if you've used that.
I don't know if you've used it as well. It's annoying how damn it did grille me. I just asked it. I was like, I'd like to expose an API endpoint that can tell whoever has the authenticated token, have some basic authentication. Is this email a subscriber to my email list or not?
Because I want to connect it with one of the events that I'm doing, what to get priority subscribers. And that's very simple, right? And then the grille me thing, it starts to just really grill me like, okay, so what about all the other things that I'm doing? Like, okay, so what about authentication? Do you want the bear token or do you want it in Jason, which is not as safe as that?
Okay, well, that's decision making. And then we go through all of these decisions and it goes really low level, including like, okay, like, how do we enforce rate limits when it comes to rate limits? You want to exactly do it when you get like a thousand per day and not allow single more, which is more complexity? And I just realized it's been a long time since I've had such an involved design discussion with a team or an engineering team and you took it, we have it when someone has deep domain knowledge. And I was both annoyed by, you know, this is just simple.
Like, no, don't worry about that. But also impressed that this thing, this AI, this LM through a series of prompts is able to do all of this. I mean, everyone's got a grimy story. So I get so many of these accomplishments, people say, you know, this very, very simple skill. It's really just telling the agent to interview you relentlessly about the topic.
It's a very small skill. It just has this weird, emergent behavior with it where the models just start thinking a little bit outside the box and they start throwing ideas at you. I think I got it originally from like a Terri Ku works of cool code. He's saying, basically, I get the agent to interview you and then you'll see better results. So when you go back to a little skill and it's, I realized, wow, okay, it's just sort of 10 times better than anything I've ever used.
And that grimy skill was the first one that sort of reminded me of the discussions I would have at my first job, you know, with the guy with the sandals. It's this very senior engineer in the room really getting me to think about everything that I've done. It was the most familiar thing to me to actually working with someone like and a race race. It just felt like a really high quality developer was asking me these good questions and I thought, wow, okay. And then I started sort of taking that and going, like, how do I mind this agent for more software fundamental stuff?
How do I make it feel more like a proper developer, a real senior? How do I tickle the right latent space in order to get its behavior to change and challenge me in interesting way? Because if you can do that, if you can increase the quality of the conversation you're having with the agents, you can increase the quality of the outputs. What I liked about the grimy skill is it forced me to make decisions that I know what decision to make when I think about it, but it is my decision. So unlike when I tell the, when you do the slash goal command, like build this and it goes off and does this and it makes all decisions or most key decisions, what I like about grimy is I both make the decision, but also sometimes it reminds me about things that I didn't think too much about, or maybe it reminds me that I should do a bit of a research, for example, like it asked me, like, which authentication would I want to bear token or over post or even over getch?
And I'm like, hang on, like, I'm going to look up like what the differences are or ask a different session to to get to like it makes me a better professional. And I do have this belief that when you're working with AI, like as long as we're learning, I think we're fine. As long as we stop learning and outsourced to learning to this thing, trouble will be brewing maybe, you know, months or years down the road. Understand, there's two things there, right? I think what everybody underestimates my agents.
Everybody is that there is a communication gap between you and the agents, right? There is a barrier there. You feel like because the agent is not a human and because you understand your hierarchy of values. You think that the agent will just pick up on them, right? There's this sort of feeling of, yeah, just trust the model, especially with the top tier models, you know, just trust the model.
But the agent, however good it is, however smart the model is, you know, even mythos, it can't read your mind. It can't read your mind. So you have to. There has to be some process of communicating your values to the agents because often when you do it like a goal, when you just go, okay, just spam me out some code, give me some slop. The agent is going to produce something that's totally misaligned for me because it doesn't understand what you think is important.
And so grille me is not only about implementation details, it's also about establishing, okay, do this. This is in scope. This is not in scope. Here's what I think is important. And so it's the agent getting to know you.
Matt just described the grille me skill. When I used the skill to design an API endpoint, the first questions I'd asked were about what the endpoint was and wasn't allowed to do and who it was allowed to do it for. Now, in my case, I had a decent idea of what I wanted, but it's generally a terrible idea to let an agent improvise authentication and authorization as they would often do. This brings us to a seasoned sponsor, WorkOS. How do you authorize AI agents?
The problem you have is how you want to control the scope of the agent. The tricky part is how permissions are static, but the job of what the agent does is dynamic. So teams pick between two bad options. Either you read the prompt, then approve every tool in the game call by hand, and then read the prompt again, then approve by hand again, until you eventually just stop reading the prompt. Or you just run a Yolo mode and letting it rip and hoping that whatever the agent does is not irreversible.
But there's a better way. WorkOS just launched airlock, intent-based access control for agents. You write the rules in plain English, for example, read repos and comments on PRs, anything touching auth or billing needs sign off, never push to main. Every call that the agent makes is judged against his task and allowed, denied or sent to a human. Every verdict is logged by airlock.
The neat thing about airlock is how there are no pre-granted scopes and there's no rules to assign. The task itself is what the agent can do. WorkOS airlock is in early access. Requested at workOS.com.com. I'd also like to mention a presenting sponsor, TurboBuffer. Matt and I are discussing a fundamental question.
How do you get agents to remember what's important? Here's an idea. What if, instead of building a complex memory system, you just let the agent search its entire history? This seems like it will be very, very expensive. But with TurboBuffer, it isn't. TurboBuffer's object storage native architecture means that the marginal cost of source sessions transcript is almost nothing, making it economical to index the entire chat history.
And because TurboBuffer names the area is a skill-virtually without limit, you can create a dedicated search index for every agent. Here's a good example of this. Entire another season's sponsor of the podcast, index is hundreds of millions of agents of agents of search and unless the coding agents retrieve what it needs to recall how and why an engineering decision was made. Entire showed that their agent was more accurate, used for your tokens, and took less time to find memories when it used TurboBuffer instead of get history and a CLI. Agent memory is a complex and a bold and used case, but perhaps there's a better lesson here. Maybe the best solution is the simple one. Just search every transcript with TurboBuffer. This is actually possible.
If Agent memory is something you're trying to solve, then please reach out to TurboBuffer at TurboBuffer.com slash pragmatic. And which other skills did you create? So from there, I thought, okay, how do I take that conversation and turn it into code? And I was immediately scared because I've been working with models just before they were good. And before the December winter where things got really good.
And so I felt the constraints from what I've been working with before. I knew that, for instance, the more context you put, you give to the agents, the worse it performs. I knew you had Dex Hall on this podcast. And Dex is a really big influence on me, especially as an idea of the smart zone. The dumb-sides on the dumb zone. So I'd hear of that just to kind of listen to that podcast and follow all that you should.
You have essentially the more context you give to the agents, every token is shouting for attention. And the more voices you put into that room, the harder it is to hear the important ones. And so the model starts losing the connections between things and making mistakes because of that. And you can think of that as a slow decline, but there is a portion of the context window where it's better and where it's worse. And so you have the smart zone, which is currently, I would say about the first 150,000 tokens of front-tier models.
Of a 1,000,000 token window. Yeah, of any, any size token window size doesn't matter the context window size. It's all about raw amounts of tokens, raw amounts of attention relationships. And then the rest of it will slowly degrade more and more and more and more. And so I started thinking, how do I take work that's bigger than 150K tokens, which is not very large, and portion it out over multiple context windows, multiple sessions.
And this took me a lot of tries, a lot of different, filling around with different approaches. The Ralph Loops was one version of that. Ralph Loops are designed to make the most of the smart zone because they essentially just give the Ralph loop a goal, and they say, do the smallest possible change that will get us further towards that goal. And then clear your context. Clear your context.
Start fresh. Exactly. And you're not technically starting from fresh because you've got the code base, right? You're just a little bit of state saved in the file system and in the environment, but not in the model essentially. So that's the idea.
So I started thinking, how do I take that Ralph loop idea, but make it a little bit more stable and turn that into skills. And so what I realized I needed was two different types of documents. You need a document for where you're going, which is the destination document. I used to call that a product requirements document or a spec is what I call it now. So that's the specification that declares when you've reached the end.
And then you need to break that spec down into individual tickets, one ticket per session. And so I'd have a very simple skill just to spec and then to tickets. And so you take that grilling session that you've had and you turn into a spec. Now that spec can work over 30, 40 tickets. Let's say you can have really massive, great big chunks of work that are all tied into that spec.
And so that's the main idea. You just grill, you turn that grilling into a spec, and then you just run some kind of implementer loop over the tickets until you've got a huge chunk of work. After a grilling, do you get user input as well or throughout this process? Or it depends. I was mostly designing this to be run for the user, like to be away from keyboard totally, because there's this idea of like the day shift and the night shift to be heard of this.
No, no, no, it's great. Basically the optimal way to work with agents is to plan during the day shift and then get the agents to work during the night shift. Right. And so hopefully you wake up in the morning and you've got beautiful clean code to look at. And that's what I was trying to optimize my process around because I was really sick of what I, and what I still do to an extent of just switching between terminals, context switching all the time, just going, blah, blah, blah, blah.
What I wanted and what I'm trying to optimize for is to just get a good chunk of planning done and then let the agent work for a couple of hours, and then I can do other work, decent chunks of time, 15 minute chunks working on one thing, planning on stuff, and then I can really the code and do that. So that's what I was trying to optimize for all the time when I was doing Ralph's loops. And that was the big thing that I found in December is these guys are good enough to delegate to, and so I can run them AFK. And then you have a different skill as well, which is a bit more ambitious called the Wayfinder skill. Can we talk about that?
Absolutely. So in the exactly the same way that implementation I noticed needed to be split out of multiple sessions. Sometimes you're grilling something and you're hitting the limits. You're going to grill something, build me a strike clone or something, right? You're going to hit the limits there.
There's no way you can plan that in 150k tokens. And so I thought, how do I break that up so that I can run grilling sessions that can be infinite left? How do I split up grilling so that it can work like that? And so I came up with this idea. Again, I'm thinking about the flow of information, essentially like, what does it need to perform well in a grilling session?
It probably needs to understand exactly what the purpose of that grilling session is, but it also needs to understand what's been decided so far, needs to understand what other grilling sessions might be happening at that moment. And I came up with this idea of a map. And the map would be the sort of center points of everything that was needed for all the decisions that you were coming up with. And once you've got a map, you realize, okay, there are certain things I can like, as I'm trying to find my way to a destination, there are certain things I know I need to decide, certain points that kind of let milestones on the map. And there's a fog of war.
And that lovely metaphor just sort of carried me through designing the rest of the skill, right? Because you've got your map, you've got your fog of war, you vaguely know where you're going. And every time you have a grilling session, it opens out more points on the map. And so you sort of figure out where you're going. And so this is kind of like a direct-to-desiclet graph where you're walking down until you reach your final destination.
And so you've got the map, and then each individual session in there are tickets on that map. And I realized, okay, grilling is good, but what if you need to prototype? What if you need to do research? What if you need to do like an arbitrary task, like provisions of infrastructure or something? Well, those are different types of tickets on the map.
And we find it basically just guides you through this process. I've had maps that have, you know, 50, 100 tickets or something until I finally reach my destination. I've actually been using it for course planning as well, so non-technical stuff, which is really grab and using it to build a garden office in my garden, right? You know, a lot of these skills we say, okay, these are great for engineering. Then you realize, okay, engineering is just a discipline that what are we doing here?
We're just discussing something. We're doing things in real life, like looking around websites and stuff. You realize how easily that can map onto other domains. So that's kind of maybe we can touch on that a bit later, which is I am thinking how transposable this stuff is into different disciplines and into different areas of life. So we find it has been great.
Yeah, but if we think one interesting thing about engineering and software engineering, when Hill Wing was on the podcast, he interviewed engineers like we thought our real engineers chemical engineers, mechanical engineers, civil engineers, and so forth, and so forth. And so he was a mechanical engineer, civil engineers and tried to find out his software engineering, real engineering. And then he found that it probably is, but he said that one interesting thing with software that is very different to every other engineering profession is the materials that we work with every single place mechanical engineering, civil engineering, chemical engineering, you have a material that has a threshold of things. You don't know exactly what it's like. You know that it'll be like it can take about this much load, et cetera.
But in software, the material is software, which is it just works like a program, take out non-termistic, which maybe brings us to more engineering. And then we're a code, you run it a thousand times and it does the same thing a thousand times, whereas in other fields, it doesn't. And he said that that he sees a big difference. But now I guess with LMs, maybe we have this thing where we have a thing where you run a thousand times and it will have these, this variety and most of engineering cast. So who knows if if both what works with LMs will be useful at other engineering, where again, they are head is fire and so or we can take some approaches from other engineering professions that will maybe work nicely with working with this material called AI.
I totally agree. What I think is interesting about software engineering and the way the reason agents are good with it is it's all the inputs and all the outputs are text based. Everything. So the inputs, code, documentation, instructions for the agent on what to do, all text based. And the output is more code is test suites is touch checking results, linting, all that stuff is text based.
The thing that agents really struggle with is anything that's non-text based. You see these amazing demos of people one-shotting a perfect UI first time. But what about if you have an interaction problem in that UI? What if you're hovering over something and the animation doesn't look right? How are you going to get that to the agent?
I mean, you can record a video, I suppose, and it sort of pauses on certain frames, let's say, but it's actually not that good in terms of vision just yet. And so anything that's non-text based is just garbage from the agent. It just can't handle it. And so I think in those sorts of professions, if you can turn, I seem to be doing simulations, right? I seem to be doing some kind of, you know, I don't know if you have a linter that can work on architectural diagram.
I'm sure you have some variety of that, right? Some simulation. If you can make that text based, if you can take the interactions that you have in your day-to-day life and turn them into text, which mostly they are anyway, then the agent's going to do a pretty good job. That's something I'm trying to do currently is take all of the services that I use and plug them into agents, right? Make them available to the agents. But yeah, the more we can make our work agent friendly, the better results we can get.
Certainly back to AI as a whole, and then what has changed. It has changed so many things, but one thing that comes up with AI is often researchers and people working with AI companies is no prior. So AI, you should go with everything that we know before because this thing is different. Start from scratch. The approaches might not work. In fact, let's assume they don't work and come up with new approaches.
Having been a developer before AI and actually, like you were really interested in building quality, great software. How much do you think AI has changed of everything, including the fundamentals? This is something that I thought too. I thought, right, AI has changed everything. I'm going to throw the baby out with a bathwater, right?
I think we just need to look at everything in a new way. I started doing that a lot. I was especially looking at life spectrum developments, you know, which is, I have some mixed feelings towards, I think it's a strange term, it encompasses too much. And I thought, okay, right, maybe English is the hot new programming language, right? Which went viral, that's a simple one under coverage posted.
Exactly. Like, maybe I can just write a spec and that specification is going to be persistent, it's going to be something I can edit and just get the agent to change it as it goes. And as I experimented with it, I tried it a lot and I was just getting worse results than if I coded it by hand. And it wasn't getting better as well. And I noticed that every time I would run this loop or change the spec, see the code change, the code would get worse.
I'm not supposed to look at the code, of course, but I'd look at the code and it was garbage. And I thought, how is the agent going to perform well in here? How is it going to work? Because the feedback loops are so important to the agent. If you have a bad test suite, the agent is going to get bad signal from it, just like a human would.
And I thought, how do I improve the test suite? How do I get this set up, not like tuning up garbage every time? And I just opened a book that I had on my shelf that I think was still wrapped in plastic the first time I took it out, which was a pragmatic programmer. Which is everyone told me to read it. Everyone said, you know, this is the best book ever. You just got it. And I bought it and I didn't read it for some reason.
And I opened it and it had a whole chapter, a whole section on software entropy. And software entropy is the concept that, you know, entropy is the idea that things go towards a more disordered state, that is more likely than going into an order state. And I realized, okay, software entropy is inevitable. What I'm seeing here is the agents are producing software entropy at a higher rate than ever. And I started looking more into that book. And almost every line I read, I thought, wow, this feels like it was written for today.
You know, you should go back to that book. These ideas of like, don't outrun your headlights, always work within your feedback loops, programming by coincidence, traceable. It's so many smart ideas. And I realized this book has been out for 25 years, right? This is probably in the agents' priors. Maybe if I just mentioned some of these concepts, especially the ones that are really pithy, like traceable, it's for instance, which is the idea that you should always get feedback really quickly on the work that you're doing.
I can say the idea of the traceable is right. Like a traceable that leaves a mark. You implement a path that works like an important piece of a software instead of building a database layer and the application layer and I don't know whatever layer, like building all three and putting them together. Like just build one part of each, but they should work together. That was the problem I was seeing with agents. You would get it to build a piece of software, even with Ralph Loops. And it would build the entire database.
And then it would build the entire application layer onto that. And it would build the entire React component library. Only at the end would it start actually plugging things together and getting feedback on what it was doing. And it was maddening because things in the database will affect what you show on the front end. You only really know where the things are actually making sense when you see it crossing those integration layers. And so another concept is vertical slices, right?
Instead of these horizontal slices across these different deploy units, you have a vertical slice where it gets feedback on what it's doing straight away and builds out from there. And so I just started using these phrases in my prompts when I was talking to the agent. And I started noticing that it was saying those phrases back to me. It was repeating them back to me. It was saying, okay, I'll turn this into a traceable. It's because this is a traceable. I'll do this.
It was using the words I was using in its own reasoning traces. And so this is what I call a leading word, like vert, let's say, which is a fancy literary term, where you lead the agents just with a simple phrase that you repeat a couple of times in the skill and the prompt to change its behavior. And so traceable is a fantastic one. And I just started diving into different books, all the books I could find to try to mine them for leading words. And another one was John Asdau's book, a software design, where I picked up tons of great stuff like deep modules, which is a massive one for me.
It's interesting to consider if these agents have all these, we've been trained on those books just still available for print and, of course, there's arguments of what they're doing with those books, whatnot. But if it's in their training data and the agents as they're trained, they connect all these different concepts and, yeah, these, I guess, leading words could invoke those concepts. And I wonder if this is much different to when on a topic talk with a professional and you're trying to describe as an amateur, what you want to a professional that says a word that fellow professional gets it. And it was just jargon, jargon. And one, and it's not very inviting when you join a company and there's jargon, but it just, we use it because it makes things faster, easier, fewer misunderstandings.
Definitely. That was something that idea led me to, because obviously you've got these leading words that are in the agents' priors, right, like traceable is all that stuff. What about describing my application? What about describing my code? How do I get the agent to, because the agents are just awfully verbose, right?
Especially opus 5, for some reason, people really go after that model for being verbose and it really is. And I thought, how do I get it to be less verbose? How do we start talking a common language between the agents, this communication barrier again? And it led me to DDD, domain driven to the design. Eric Evans incredible book where he talks about ubiquitous language, a language again really deep in the agents' priors and understands it really well.
I sort of started toying with the idea of maybe changing grille meat a little bit, because grille means very simple skill. But what if we, while we were ideating, while we were thinking about the application we were going to build, what if we were also building a domain language? Or if we were also deciding on the right terms to use, and this turned into a skill called grill with docs, which is terribly main skill. But it's going to create this domain language as you go. And if you get the agent to use the domain language, the difference is night and day.
Because suddenly you're speaking the same language, you're able to describe the things you want to change in so many fewer words. Like I had this app that I'd sort of work on, there's this complicated interaction where there are ghost lessons and real lessons, and what happens when you turn a ghost lesson that's inside a ghost section, inside a ghost course, into a real lesson that means the ghost section needs to become real, the ghost course needs to become real. How do you explain that? Well, that's the materialization cascade, right? And you came up with these.
Yes, with the agent, right? The agent's actually really good at coming up with these terms. And so I have a domain modeling skill. And we talk about jargon, but really it's domain language. And if you can integrate that and integrate that not only with the way you talk about the app code itself, then you've got a studio going, right?
It's very, very exciting. And it means that the agent can navigate your code base a lot easier. It can find the functions that mention that specific domain terminology, just with a simple grep. It's just gorgeous. So that's something I've been really integrating with every part of my setup is diddity.
But this is so interesting because in an effort to make these agents work more efficiently or do workflows that just mean that you can produce better software with fewer mistakes. And these things, you start to go back in time, found this book that is now, I think, what is it, like 20, 30, 40 years old? And you're even still going back and finding jumps from us. I'm sure at some point you'll get to Mythical Man month. Yeah, I've got it already. I've got it already. I've got it already. I've got it already. I've got it already.
Which is now more than 50 years. And you're trying to find the right words to describe things, which is very curious because when I talk with Kent back on how they used to program with Word Cunningham, as they were coming up with the concept of actually just domain design patterns, they had a desires with them. And they would look through trying to find the right word that has the right meaning. And they had it on their desk right now. This feels we're going back to the fundamentals, the how that people have been asking themselves and every now and then people write it in books.
And it kind of spreads as wisdom. And now we're back to where we started, which is what you're trying to teach the wisdom part. It's wild, right? Because AI is so different to humans, you need to optimize it. Imagine you essentially had a human who wakes up every morning and cannot remember who they are, right? The guy from Memento.
This is Memento driven developments, right? We are trying to optimize our code bases for new starters. So we're trying to have the most healthy code base that we've ever had. Because if you, a human can work around a bad code base. They just develop memory. They just slam their headings the wall again and again until they've got there.
But an agent can't do that. It starts fresh every single session. And so you need to optimize your code base for that person. That leads you down into really interesting parts. And it turns out that software fundamentals have been saying we've been trying to do that for the entire time, right?
I am fully like, I don't know, Eric Evans, Pild, I'm fully like software fundamentals build. We are sort of changing the rules a little bit, but maybe we're just emphasizing rules that we knew we were supposed to do, but maybe we didn't. And I find that really fascinating. It's definitely a lot of fun. Okay. Like, I think it's easy enough to follow with this train of thought, why fundamentals matter?
But which fundamentals? And if I'm an engineer, especially maybe someone who has been just kind of like heads down coding, more tactical coding, how do I go about and find those fundamentals that matter and go back to them? What do you found work? This is really tough question, right? It's really tough because strategic programming has always been really hard to learn.
The reason for that is that the feedback loop on it is really long. You would often find like people who quit their jobs after six months, their strategic mistakes never catch up with them, right? Maybe that strategic mistake takes nine months to come back at you. I think of strategic learning, strategic programming is kind of like you've got a huge mixing desk in front of you with loads of these different sliders. Maybe one of those sliders is like the amount of deployable units that you have. You turn it up, you've got more microservices, right?
You turn it down, you've got a monolith. How do you make that decision? Where do you put that slider? Because it's kind of like you're mastering something, you're mixing some music, but you can't hear what's wrong until nine months later, right? Until the mistakes kind of get you. So I think the only thing that can make that feedback loop faster is moving faster.
AI now lets you move faster, right? And so your strategic mistakes will come back at you quicker. Because AI is just able to produce so much code. And so what you need to be thinking about is that your code is the environment that the agent operates in. And you should always be thinking about improving that environment, thinking about how to do it better.
And obviously that requires a bit of tactical knowledge, right? You need to understand what code is and how it fits together and what the memory constraints are and all that stuff. But in order to get better at strategic programming, you just need to be thinking on that level all the time. And I would say reading these books as well because just having the language to explain that and understanding the difference between applying strategic techniques and not is the whole game. I mean, up to pre-AI for senior developers, senior engineer, staff engineers, they were the people who often you didn't see a senior engineer under five years of experience because you typically needed even in a fast space environment.
You needed that much time to get the feedback loops to make the mistakes, make your own mistakes. And by the time people got to staff engineer oftentimes around 10 plus years of experience, some people did earlier, but they often just had paddle-scalls all over them. And they would, you know, someone started and you probably would go in and they would just make a tweak. And it wouldn't be clear why. And they were like, trust me on this, we're avoiding disaster and production or on call or whatnot.
But all of this came to a lived experience. Now, AI speeds things up. It also makes it easier to fix mistakes. I'm wondering how this might change. Like, on one end, I could see how it could just speed up experience. Like, you can, in a year, some people, some teams will ship more projects than they have in four years or about the same as in, let's say, three or four years before.
So you get a lot more experience. But I wonder if sometimes the mistakes that you make are just not as serious because you can fix them quickly. And now I wonder if the learning is not as strong. Because again, like some of these battles are so these war stories are, it was just really bad outage. We lost a lot of money because we didn't have an idempotent key.
And now of course, now you know what I'd be pulling C as a, it's not an easy concept, but it's important if you've been hurt by it and so on. If you're a company right now and you want to train the next junior developer, like, because this strategic programming knowledge is so valuable now, because you can use it as such high leverage, are you really going to employ someone without it? Like, why would you? Like, I was asking, I didn't see a Uncle Bob of the day. And his recommendation was, okay, you just hire someone and you treat them as an agent for a while.
You just delegate to them. You keep them that tactical mindset for a while. And still, they mistake start coming up at you. But that's such an enormous waste of money for people, right? Like when software engineering, when the tactical stuff is gone below minimum wage in a lot of countries.
So I don't know is the answer. I only know that the strategic stuff, the understanding of the code, the understanding of the long view has gotten more valuable than it's ever been, right? Because you can just get so much leverage out to it. I asked about interesting things I'd like to know from you. And this is very related to this.
This person asked like, how do you convince non-engineering stakeholders that investing in software fundamentals are important, even if they might be used to speed on productivity on paper? I think the question here is, if some people have, okay, look, we do want to get the fundamentals right, which means we want to take a bit slower, think about their decisions, maybe educate ourselves as well, as opposed to just turning it out. I mean, you could have asked the same question 10 years ago, right? Okay, we'd still be relevant. Except we'd still have software fundamentals, we'd be able to ask about the paying off-tech duct.
Exactly. And it's the same thing, right? Like, we have been having the same conversation, which is quite satisfying to me because you need some sort of metric for figuring this out. And it's a little easier to figure this out because agents allow you to move faster. And the first step to this is getting observability in your organization over every single agent on what it's doing and what its success and failure rate is.
We've never been able to have that with developers before, you know, that's kind of invasive for developers. I'm afraid agents. It's like, it's okay. It's okay. We're paying for this service, right? We need to understand how well we're optimizing for it.
The first step there is actually getting a harness or observability around your agents, the entire organization to work out what's working or not. And you probably need someone whose job it is or part of their job is to look at that data and figure out what we're doing. Maybe some reposing your organization have better success rates than others. And so you take the lessons that are in there and you pass them out. I also think that most organizations need to gather around a common set of skills.
You need a common software workflow process so that everyone can contribute back to it so that you can experiment with things. You can A, B test things. You know, you can have one team doing one set of stuff and one team doing another set of stuff and then you ask them afterwards. And so everyone working with agents in any kind of organization needs this experimental mindset. You need to be thinking, how do we get more juice out of these tokens that we're spending and observability is the first step there.
And I also wonder if there's a human feedback loop in the sense that, I mean, just talk to your colleagues. Like, you know, we do have rituals team meetings, company-wide meetings for a reason. Like, there, share, here's what's working for me, here's where it didn't work, here's where I'm learning. Like, in the end, we are in charge of setting up the rules, deciding how we use them, where we use them, where we don't use them. And where we say, like, no, this needs to be humanly to take 100%.
Like, we're not even getting any involved, which, again, will be different everywhere. And it's not only that, like, a lot of this stuff now you don't need to be human in the loop for, right? You don't actually need to delegate that much time in order to build up a better code base. I have loops that essentially, every morning, it will run my improved code base architecture skill and give me a proposal for the something that I could improve in the code base. And then I can just press a button, I can say, okay, turn that into tickets, and then let's ship that.
That is pretty easy to do, and it's pretty easy to stream that in with other work. And so I think that, I don't know whether you need, like, 20% of your time focusing on the factory that's built your software as well as the software, because I feel like that's a massive, incredible investment into your future leverage. And not only your leverage with your work, but also your team's leverage and understanding and getting better at those skills. But of course, you need results, and you might need to hide that work for a bit before you actually reveal it to this is what we've been doing. Well, and this is down to your environment, but yeah, I know it's going to matter to you if you come back single.
By the way, guys, I also did this. Yeah, exactly. I want to ask you about your specific kind of how you use tools. First one is coding agents local or in the cloud, and you recently posted a pretty provocative tweet, which is, I'll quote you, I'm moving away from my local Devs setup, make zero sense to me now. A lot of people ask me, how do you make your skills collaborative?
How do you have a collaborative grilling session? And the answer to that is that you need more than just your terminal and you, right? We're in a phase now where every dev has like 100 terminals available to them, and that seems crazy. It feels like you need those 100 terminals available to your entire organization. You need to be able to collaborate in a shared space.
You need to be able to ask someone, tag someone in to your grilling session and say, okay, do this. And so it makes a lot of sense for me to have a lot of those interactions in the place where you already work in Slack or in Discord or in teams, whatever, or linear. And that is really the thing that's driving me to explore this. I didn't work with a team particularly, but I understand the value of that and I've been trying to build that into my flows. So on the train over here, I'm in Discord chatting to my heads and box, you know, building stuff for my course or fixing bugs that students are coming across.
So I can see less value now in just doing things locally when I have this setup that I can port forward into, let's say, and like see the dev server as it's making changes. And I don't know, it just feels like it makes way more sense to me than having a very, very expensive laptop that can do this stuff, because I wasted compute. And especially because on that remote box, I can set up schedules. I know the box is always going to be on. I have like a morning stand up with my agent where I get it's schedules my day for me and it understands all of my Discord chats and all that.
Yeah, having that remote feels like it makes just so much more sense for me. And the only thing I do locally now is debugging issues with the remote box. I think I wonder if there's a question of how easiest to replicate some pretty complicated local setups in the cloud, but once that becomes possible, it's probably a matter of when not an if. Yeah, and if anything, people having this similar issue with local setups, right, with just 1000 Git workshrees just spanning their hard drive. And with how do I have a work tree that I've got to run like five Docker containers in order to get my local dev set up?
Well, that's often a little bit easier in the cloud because you can just provision the resources that you need on demand. And by the way, we're seeing that companies like Ramp, Stripe, Uber, that have platform teams that manage to take a local dev's full setup and put it into the cloud machine that you can now invoke local dev. So on cloud machine that you can now invoke with an at Slack or a website, they're seeing people use these agents far more except for front and work, which you still want to have that feedback loop. There are a few exceptions where you really want to have that like a local dev set up for latency or whatnot, but they're also seeing like 70, 80% of devs are just voluntarily going for the cloud. Yeah, I mean, I think you can just tunnel through and just get the if it's running a dev server and you just have that appearing on your local machine.
How is that different from having it locally, right? I think I think I'm not experimented with that, but that's when I talked about that and said it will maybe front-side is a good exception. That was the immediate response that I got and makes sense to me. I want to ask about planning and requirements of your big believer in grille me and planning out front or getting a plan and then having the agent work, but there's a devil's advocate here. Agents are so fast at implementing. You can actually even have like several, like a few edges implement different architectures.
What about the approach of like, well, they're fast at implementing. So I might not need to do as much upfront planning. I can just course correct as I go. It depends what type of work you're doing, right? Because I believe that you shouldn't be using grille me for everything.
Essentially, you need grille me for pieces of work where the actual thing being done is going to be quite large and hard to row back from. If you feel like, okay, this feature, maybe it's a whole new page, maybe it's a big feature, this code, you think if the agent gets it wrong, then the wrong code is going to be in its context window influencing everything that comes afterwards. And actually going back and editing the stuff afterwards and doing the alignment after the fact is going to be expensive. Whereas for those cases, it makes sense to align first to answer all of the tricky questions, like your JSON cookie or whatever, your authentication token first and then do it. But for some cases, like simple box fixes or just like move this button three pixels to the left, it's obvious that you don't need to align before that.
You can see the thing, if it's just like a five line change or something, you can align afterwards. And so that's how I think of it is that where you can, you should shift right as much as possible. And actually, there are actually certain features that I have a little in my video editor, I have a button that I can send feedback to it and often use this for very simple tasks where I send the feedback, it goes into a GitHub issue, this immediately gets picked up by an implementer agent, gets just worked on immediately. Then a code review agent comes in and reviews the code. And then at the end, I get to see this actual thing being fixed and I can do my alignment then.
And that's what really well for things that are very easy to specify things that I don't need to grill on. So those are the choices you've got. Is it a small enough thing that I can align afterwards? Then don't use grill me. Does it fit into a single session?
Then use grill me. Does it span multiple sessions? I need to align over the entire thing, then use wayfinder. Interesting because it's not all that different to where some tech companies landed years before, which is on the PRD, the product reference document. If it's something trivial, just build it.
If it requires the team, like it's a team level scope, I mean right up here, at least send out the team, maybe CC some other teams, but it's not a blocker. And if it's something bigger, then it's a blocker. Like we need to wait for feedback. And then basically the way we would say it is like look, if it's like a one-month project, like spend two days, it's not a bad thing to spend one or two days planning it because we're going to save time on it. But if it's a one-day project, forget about it.
If it's a one-year project, I mean what are we doing? It should be a smaller one. Totally. And I want to, like, there's a bit of criticism I hear just from outside the room when you say that, which is that it doesn't sound like waterfall. What we're doing?
When I'm talking about wayfinder and when I'm doing any kind of building up any kind of spec, I do a lot of upfront aggressive prototyping before we get there. That's something that comes up again and again and again. It's like this is just waterfall. What are we doing? But agents give you this ability of just churning out slop, right?
Sometimes you can use that to your advantage because a prototype, right, just getting a sense of what it should look like. You can build out three or four different versions and just choose your favorite and iterate on it and just keep churning, churning, churning. That can be a really powerful setup that we've not really had before, right? It was always expensive to produce prototypes. Now it's the cheapest that it's ever been and that's an essential part of writing specs to me.
It's actually producing these prototypes. Yeah, but also like the water for criticism. I think Brady Buch might have told me this as well as I can't forget like we should not criticize waterfall because for example, a lot of big tech, the largest companies from like Amazon, Microsoft, Google, you name it. They are kind of doing mini waterfall like pre AI. They've been doing pretty many waterfalls, which is let's do a plan.
Let's agree on it. Let's build it. Let's ship it. And this is all done in like two weeks a month, two months, three months, three months is kind of the extreme. The creativity was saying the problem was never this waterfall.
The problem was the planning was literally taking like a year, like one year and then the implementation taking three years and by the time it was ready four years later, it's not what we wanted. And that was the problem. It's like the problem is not like having like a one or two month project or one week project with a waterfall. The problem was always that this what we're talking years and he said that the industry has not seen waterfalls for decades now. And so here we're using this term, which is a bit like we're criticizing or mini waterfalls were criticizing that one.
It's actually that's not about thing. That's not about thing. It's a scarecrow that we're punching or something. Yeah, it's a pinata, which stopped existing. It might exist in some crazy like enterprise projects.
No one is known about and regulate industries, but I feel even there is probably kind of stuff. I think it's like if we're hitting the pinata, I think it's actually a useful thing to have up there. It's like a useful ghost or useful cautionary tale, right? Because what which one fits the agent's accept more closely is going to be agile, right? Because the cost of labor has gone down so much.
We can just make changes very, very quickly. I don't know. That feels like the right metaphor to me. So I don't mind hitting on waterfalls, even though no one really does it anymore. One other thing that just is just one out of style.
We didn't hate it, but test driven development TDD. What is your take on using them for agent accept? When I saw what Ken Beck would talk about how this could be a great fit for many reasons, but I still don't see people really using it. I see people writing tests, the agents also like write tests after the fact, which is how most people work. But I think you've got an advocate for TDD, right?
Yeah. So I have a TDD skill which I recommend using. And this is quite timely because I have been thinking about it. TDD optimizes for having a very small working memory, right? You write one test and that test is supposed to fail.
And it means that even if you get distracted, you go for coffee or something, you go for a long walk. When you come back, the test is still failing, reminding you of where you are in the implementation and guiding you to the next thing. Agents don't need that. Agents, the thing that's great about agents is that they have a much larger working memory than humans, right? They can actually hold a lot more in their heads than humans can currently, which is very useful.
But they don't have an infinite working memory. And TDD, it's sort of aiming at the wrong problem, I think. But the thing that agents really do need is that they need to have feedback loops. So they need to see what they're doing and how it's interacting with the environment of the code. They need to probe it all the time.
And having an agent that builds the failure first, it's also very hard for an agent to cheat that. So not only are you forcing the agent to build its own feedback loops, the agent is providing proof to you that the thing is actually working as it goes. And even if I'm not using TDD directly, I'll wear it right in the same test as then fix it and refactors. I will often say provide proof that your change does the thing it's put to do. Give me TDD evidence, right?
That it would fail without this change. And that's been really good for just improving the feedback loops essentially because another thing with TDD that agents get wrong is they will often just write crap tests. They'll often just write, especially tautological tests where the test is just asserting the implementation itself. It's just like a duplicate of it. It writes a constant and then it says, expect this constant to be this value.
I mean, what's the point in that test? It's just asserting implementation. So yeah, I have a mixed relationship with TDD. I do still recommend it just because it gives you so much more confidence in what you're building from a human perspective. But yeah, I'm starting to see the counter-alkants.
Let's talk about tech-dubbed. Jair Friedman, a white combinator, brought a tweet that I called from him. I think that used to be something you just had to live with. It's just a large code base no longer and to which you replied, yes, now you can live with it even in a tiny code base. That's good.
I read out loud, actually. You really gave for you. I was sent to that one. Yeah, it's just so easy for agents to produce rubbish, right? But even really smart, powerful agents, because they're unable to think strategically, they're just focused on what they're doing right now.
It's very easy for them to produce tech-dats. What is tech-dats? Right. Tech-dats is anything that makes the code base harder to make modifications to over time. A good code base is one that's easy to change, easy to make a change in that doesn't result in cascading failures.
Right? So a code base with a solid test coverage and a good test suite is a code base that's easy to change. Yeah. But it's so easy for agents to just make a code base worse over time. No.
And it's a really hard problem. And it's one that you need a strategic mindset to think about. One thing that I found works really well is automated review. So you have one implementer agent to do the thing, and then you have another automated review agent that sort of imposes your coding standards that looks for these tautological tests improves the quality of the test suite over time. But then how do you know if the automated review agent is doing a good job?
And so even in tiny code bases, even in one line changes, the agent can produce a wrap. And so I think it's just something we need to live with and something we need to be in a constant battle against. No, it's also not a bad thing. We bring a bunch of value when you understand what good code looks like, what big when you go right, what tech-dats. And it's also a problem that we've always had.
You know what I mean? Has it gone away? Has it gone away? This is what I feel like. We just have the same conversations we've had for 20 years.
It's just there's this new elephant in the room. I want to ask you about living in the UK and AI. This is a question that also came from one of the readers now that you're based in the UK and outside of London. But you're now educating about AI. Is being further away from Silicon Valley and the HQ of the labs making things easier or harder for you?
I'm really just trying to plan my own furrow. Really? What I realized quite early on is that I have no power to predict the future. Because I'm so far away from things, I'm just a person in the field working with this stuff. I have no way of knowing what's coming.
I don't know whether the model is going to prove I don't have privilege access to stuff. So I'm just trying to focus on what's working right now. And because of that, I think that's narrowed my scope a little bit. That means I can just try to get my stuff working. And it's sort of quite surprising to me that it's working as well as it is.
Because I don't have this privilege access. I'm just trying to make this one approach work. So I think, yeah, you're probably right. I probably will be able to do this stuff if I lived in San Francisco, but then I'd have to live in San Francisco. I don't want to do that.
That's miserable. That's a great set of it. My parents are just down the road. I've got my son growing up on the countryside. So it is what it is.
And, yeah. You're an educator at heart. How have you seen the business of teaching or educating software engineers change? And also how people want to learn? If you've observed that any trends from before, like, already when you started, I feel you were on, at the time where online courses and learning over video became a lot more popular, as opposed to, let's say, a decade ago where it was maybe tutorials and before that it was books.
Obviously, they still exist, but they're just different preferences. Yeah. It was around COVID time that sort of video tutorials really took off. I think you've wanted a much richer learning experience and I was kind of just after that wave, I suppose. I think that people's way they've learned doesn't, hasn't changed that much, right?
And their desire for certain types of materials hasn't changed. I think it's very sexy, the idea that, you know, an agent can just come in and teach you everything. And that sort of works in some contexts, but really what you want is curation, right? You want a human to have come in, understand the flow of the information. I always think of information as kind of like a graph, right?
You have a piece of information that's dependent on another piece of information, dependent on another piece of information. And that turning that graph into a linear path is how I think of my job, right? I'm just trying to teach you like find Dijkstra's algorithm through the graph so that you can learn it in the most sense of way. And that level of curation is just not something that, again, that's strategic, right? That's not something that AI is particularly good at.
So, I mean, I've obviously made this huge pivot from TypeScript, from tactical stuff really to this strategic layer. And it's working okay for me. I really can't speak for other folks doing this work. And I know that lots of people are not having this level of success, I suppose. So, I think what it shows is that agents have just changed the game in terms of what people value and what people prioritize.
And the industry has shifted in seven months faster than it's, I think, ever done. You know, it's a huge shift. It doesn't mean we need to throw away our working practices, but it does mean that what we need to focus on is difference. And I feel like I've been able to move with that quite well, whereas I think others just haven't because they're focused on different things. And I wonder if in your case, it's also with total TypeScript, and even before with TypeScript, a lot of things you shared, you were helping people use the very popular tool at the time.
TypeScript was gaining market share. There were migrations happening from JavaScript from Python to TypeScript. And so on. And so developers wanted to get really a lot of them, or the top 10%, or top 20%, you name it, wanted to get really, really good with TypeScript, and they were looking for efficient ways of doing it. Now, AI is here, it's changing how we work as software engineers, and I think it's really the building software is valuable, but there's a question of how do I use these tools more efficiently, which is more pressing right now than how do I write TypeScript efficiently, especially with agents.
So, I wonder if you've kind of just a little bit how you pivot it from voice acting to what you can do from outside of London to a thing that you could do outside of London, which we're still teaching. Which right now is again, it's on so many people's minds. I think I've just been lucky, basically, of choosing the right thing at the right time. It would have been very easy for me to, and I actually took quite a fair bit of convincing to move into AI. Like back a couple of years ago, it was Joel, my business partner, who was pushing me to actually go, you've really got to try this.
It's actually pretty good. And you can use it for all sorts of stuff. And it took about three months of me actually trying it and failing and trying it and failing before I realized, okay, this is great. I just feel quite fortunate that I've landed in the right place at the right time. And I try not to narrativise it.
I try not to think, well, well done, Matt, you've been so smart, you know, making the right time. I've been doing several mistakes as well. And I could have easily found myself in a different zone. And I think I would just get back to being an engineer. That's why I love to.
Putting us back into the shoes of when you were someone just turning out in the industry today for people starting out in the industry or the career junior folks, what would you recommend them for tactical things to do? Like they will know, like, look, I want to get that experience. I want to get that judgment, that taste, those fundamentals. You'll need to get repetitions. And if you found yourself in those shoes, how would you approach?
Like I want to be a builder, a software engineer with all these AI tools, whatnot, which is now confusing because now there's a mix of the way. Use these tools just to do stuff for me. Do I get the fundamentals, which is slow and so on? Yeah, I mean, I would love to be a junior right now. I would love to be in the exact position I was in like 2014 where I was building these tools for my students, right?
I actually got really nostalgic for it on the other day. I thought, I'd love to get back into some teaching because just the ability to, like, I could finish a lesson and then just prompt the agent. Okay, this tool didn't quite work in that way. I could maybe modify it a little bit and, you know, see it working. I just think the right thing to do is to use these agents as much as possible because that's how people are going to be working now.
And I think the thing that I find valuable about my skill set is you're constantly in touch with the changes that are happening. Grille me, not only you're having a discussion with the senior developers, but you're having a discussion with a senior developer, right? That's beneficial for the developer, but it's also beneficial for you. Keeps you thinking about these deeper ideas and the absolute rubbish that I was tuning out, you know, with my Spectrogram Analysis tool, that would be so much better if I had an agent to work with. It ran like a pig, you know, like it was performance was actually terrible.
If I'd have been able to say, okay, this frame rate has dropped to 10 frames per second, how do I fix that? It would have seen the six nested for loops and gone, okay, maybe you should do something different there. So I think that there's never been a more empowering time to work on this stuff as long as you're interested in not only the code you're producing, but also the process of creating the code. So I've been a better time to be a kind of naval gazing programmer just constantly thinking about your own processes and being introspective. This sounds like you're motivated that you should be able to learn really fast compared to me before.
Absolutely. It's just about being curious about being adaptable and that's the people that I see who are thriving in this new environment. Other than people who are thriving 10 years ago because they're just interested in this work, interested in making better software and interested in their own process. And I'm interested in making better software as I want to ask you about gardening. A software engineer on XLOR and posted out a quote every team user gardener, someone quietly watching stream a PR, flowing into your code base and the least of the smells, the linspirations creeping like ivy across your careful of haunted garden, a steady hand in telling the weeds staff, which otherwise engulf the garden and to which you're applied, I'd argue the only thing your team needs are gardeners.
You probably do need a couple of other people. Yeah, but more specifically, once you ask about this concept of gardening, I actually really love how Lauren described the weeds taking over the garden and getting them out. I think I made it sweet a while ago that we are, this is when I was thinking about Ralph and sort of the agent sort of looping over stuff. We are essentially just Ralph's platform team, right? That's what we are now.
And we are the, we are our agents platform team. We are trying to build the environment for them to succeed. That's exactly how you should be thinking about it again. And that garden metaphor is nice because, you know, it's very easy for the garden to itself just to suffer entropy, right? To gather weeds and to do all that stuff.
So understanding and diagnosing that stuff before it becomes a problem in your own code bases, an essential skill and might be the essential skill, right? As long as you can queue up work for agents, as long as you can build these loops now that we're starting to see these processes where agents improve the code base based on bug reports and feedbacks, that feels to me like really cool work and noble, interesting work as well. We talked about some great standout software engineering that you learned from. You got an inspiration from today. What scale sets experience approach do you think makes a great software engineer?
I'll use an example, which is Lars Rammel of the works at VSL on the AICK. We had a check with you all day. And he is building an entire software factory for his extremely popular open source library that gets a ton of issues. We're talking about planning again. We're talking about gardening.
We're like thinking about the processes of software development. And I suppose if I had to put it in a word, it would be introspection. It would be looking at yourself and the ability to take what you do and put that into something the AI can work with. You're essentially trying to put your process into words. And that's what I've been doing with the skills.
That's what I've been trying to do with the automation type in creating as well. It's not just look at what I'm doing and think, how could I do this better? And also, how could I encode this into this strange animal that I have in front of me? How can I make it work like I want to? And that attitude has been really helpful for me.
It's something that I value in Lars and I value in all the people that I work with when they approach agents. And then, as closing, what is the book that you would recommend or multiple books? I'll go with a pragmatic programmer, a philosophy of software designed by John Asselow, and I'd say the first three chapters of DDD, the Eric Evans book, the E-Bixus language one. That one in particular, it's really great for the ubiquitous language concepts, the domain modeling, the actual coding into code, I'm not such a huge fan of. But those three are the big three.
Awesome, Math. Well, thank you. This was really interesting and really fun. Great to finally be on the book, Casio. Meet the famous guy himself.
It's great. But we've met before. It's great to be here. It was so nice to sit down with Math and I have to say, knowing that he was a voice coach and actor, makes me understand how he talks so smooth and how he's pleasant to listen to him. Probably the most amusing part from this conversation was how, as Math was searching for how to work better with AI, it wasn't modern approaches that he found to be useful.
Instead, he went back to classic software engineering books, the pragmatic programmer, the philosophy of software design, and the domain driven design. There's some irony as to how the best practices documented 20-plus years ago, like tactical versus strategic programming in this book, not only do they still work, but they become more important when writing code with AI agents. A really important point I want to emphasize is the importance of leading words with AI. When Math started to use terms like tracer bullet or vertical slices, the model started to follow his ideas better in planning. And if you think about it, this makes sense because software engineering literature is part of L and training, so these terms are also part of the model's priors.
Just as interestingly, using the right words for describing your problem is not a new concept. For example, when I had Ken back on the podcast, he talked about how 30 or 35 years back, when him and work Cunningham had a thizzles on their desks, they used it to try to find the best words for the specific thing they were describing. This was just another full circle moment on how words do matter. Finally, I appreciated Math's push on how you should want a clean codebase, not just because it's easier for humans to navigate, although I think you really want to do it for that as well, but also conveniently, agents do not have a long-term memory, and they will look at your codebase for the first time on every new run, and it's much easier to get around inside a well-structured codebase than one that is really messy. Check out the show most below for an interview with Jon Osterhout, the author of a full-