The Secret AI Prompt Tool Silicon Valley Engineers Are Using

Welcome to the Next Wave podcast. I'm your host, Nathan Lans, and today we have a killer episode. Everyone's talking about vibe coding, but the reality is for most things, vibe coding doesn't work right now. And even the guy who coined the term, Andre Carpathi, he recently posted that he's now trying to provide more context to models cuz he's realized that's what you have to do to get good results back. The reality is if you want to get the most out of AI coding, you've got to be showing the models exactly the context they need. Today

I'm going to show you the secret weapon that all the top AI coders are using that you probably haven't even heard of yet. Today I've got the founder of Repoprompt, Eric Provener, on here and he's going to show you exactly how it works and how you can use Repomprompt to take your AI coding to the next level. So let's just jump right in. Yeah. I've been telling people about Reprompt for like the last, you know, probably 6 months or so. I've kind of felt like it's been almost like my like AI coding secret weapon. Yeah.

Yeah. Yeah, I mean I I really want to bring you on after I saw that tweet from Andre Carpathi the other day. He used to be a Tesla AI now. He's like one of the best educators about how LLMs work and things like that. He had this uh this tweet saying, "Noticing myself adopting a certain rhythm in AI assisted coding. I code I actually and professionally care about contrast to vibe code." You know, he coined the term vibe code which everyone's been using. And he he then he basically goes on to talk about like stuffing

everything relevant into context. All this I was like he doesn't know about repop. I'm like, how did this like top AI educator in the world, top expert, everything totally has no idea about reprompt? I was like, okay, so I need to get Eric on the podcast at least try to help with that. Yeah, I appreciate that. Yeah, I mean, yeah, looking at that that tweet, you see exactly like that that flow that like got me started. Like you when when you start getting serious about coding with AI, like you start thinking like, okay, well, how

do I get the information to the AI model and like the UX on all these other tools is just not cutting it for it. Like you need a tool to just be able to quickly select, search for your files, like find things. There's probably some people listening right now like, "Oh, it's okay. Cool. Nathan's excited. What is it?" So, like maybe, you know, if you could explain like try to simplify it and I think we should then just jump into a demo and we can kind of just thing. Sure thing. Yeah. It'll probably be easier

if I just show the screen here. I mean, the first thing you're going to do when you're you're going to open up repo prompt is pick a folder. So, this here is like the entire folder. But generally, when you're working in a repo like this, you want to think through like what are the files that are are going through. And if you're using a coding agent like with cursor or whatever, what the first thing they're going to do when you ask a question is, okay, well, let me go find what the user is trying to

do. Let me search for files and pick those out. And with repo prompt, that's like on you as a user to to do that work. There are tools to help with it. So yeah, like I'll just show uh here with the website. This is my my other project. This is this is the website for repo prompt site. I was just iterating on some pricing. I just lowered the annual price there. And uh yeah, so help me update all the docs pages. So if I do that and then I just do gemini flash quickly to show

what that looks like. So it will actually search for for files using an LLM based on the prompt that you've typed out. So like you know a big part of using repo prompt is that you have to know you know what it is that you're trying to select here, right? And you know what I noticed a lot of users they were just putting everything in. They would just say like okay just select all and and that would be it. And you'd be like okay well I was doing it at first. Yeah. Yeah. I mean I

mean yeah I mean that's the easy thing to do. You're like okay well there's the codebase. Perfect. But you know there's plenty of tools that can just zip up your codebase and that's that's easy. But like the power prompt is you can be selective. You don't have to select everything. So I can just hit replace here and then okay well what did that do? Okay. Well that actually found all these files here that are related to my query. Put them in order of like priority of like importance based on what the LM judgment is. And

of course if you use Gemini Flash you're not going to get the best results compared to like using you know like a bigger model like Gemini 2.5 Pro. it'll be a lot more thorough, but you know, like that that's the thing. So, it'll it'll pick those out. It'll use something called code maps to help with that. You can see that the actual token file selection queries is just 6k tokens. If I were to switch over to the repo prompt codebase, which is a Swift codebase, you can see that the code the the codebase is a

lot larger, you know, so the map is 85,000 tokens. So, I I'll explain a little bit like what is a map really? Like what what do you do with that and and how does this help and what does it do? working with a codebase, you know, if if you've uh spent some time, you know, programming in the past. I know a lot of folks are are getting into just using AI and they're they're they're not super familiar with all the technicals there, but like vibe coding and Yeah. Exactly. Exactly. Well, you know, there's a lot

of files and you want to like try and get an understanding of the files. So, Rever has this code map feature and what this will do is it will basically as you add files, it'll index them and extract what's called like it's it's a map, but so it's like a highle extracted view like an index of your codebase. Exactly. Yeah. And this works for a bunch of languages, Swift being one of them. And if I were to select everything and kind of show you, you know, what that looks like. So these are the the map

files. So I'll open that up. I'll open VS Code just to show you. So you've got like all the the file definitions and uh you know, just just the you what's private, what's what's public, and you can scroll through and and so you give this information to the AI model, and it's able to be very selective. Uh well, it's able to have a lot more information in a lot less context. And so the context builder uses that data to help you find what what the relevant files are based on your query. So it has like

a kind of peak inside the files without having all of the details and it's able to kind of surface that relevant information for you so that you can use that in a prompt. So yeah, I've I've gone and I've explained a little bit about like all this ways of picking files and stuff like what what do you do with that, you know, right? Like that's like the the big question, right? I think it's also good to explain like why it's important because I I think one thing like one thing I love about repo prompts so

when I first started using it I had been like using just like a custom script I had created to like take my codebase and like and then like put you know the relevant you know context in there which a lot of times I was just doing all of it. I was literally putting all into a single file and I' I'd copy and paste that into chat GBT when pro when it first came out and I think I tweet about this someone and someone told me like oh you got to try repo prompt. So when I

tried repo prompt, the fact that I could like see how much context I was sharing with the model was amazing. And and it seems like that's super relevant too because you know at least from the benchmarks I've seen you know everyone talking about how much context you can put into their LLM and now you're getting like 1 million and I think Llama was four or something crazy high like that. I think it's 1 million. Oh no it goes to 10 I think. Llama. Oh yeah 10. But I think on the benchmarks though, right? I think

on the benchmarks for Llama 4, as soon as you went over like 128k context, like nowhere near the 10 million, like the quality just like dropped like like a rock. Yeah. Well, you know, like until Gemini 2.5 came out, like pretty much all the models and I would say even including like 01 Pro, like you would really want to stay below 32k tokens in general. I just I find like over that you're you're you're you're just losing a lot of intelligence. Like so there's this concept of like effective context, you know, the effective context window.

Like at what point does the intelligence stop being like as relevant for that model and for a lot of smaller models and local models, it's a lot lower and you probably want to stay around 8K tokens. But like for for bigger models, 32K is a good number. And it's only now with Gemini that you're able to kind of use the full the full package, the full context window. And 03 is quite good too on these benchmarks too for context utilization. But but yeah, so you're using this context, you've picked out your files. Say you you

want to use as many as you want 100K like what do you do with that? Uh and so I have a question here. I'm just going to paste it to 03 and you'll see like what what is 03 getting out of this out of this? So it's getting basically this file tree. So it's getting a directory structure of of this uh of this project. It's getting basically the highle code maps of the of the files that I haven't selected. So basically when it's set to complete everything that I haven't selected gets kind of shipped in

and then you have the files that I did select and and so then the context is able to to go ahead and all right there you Oh, so 03 gave like a really nice response here with tables. Loves tables on how to do that. And so this is like a great way to kind of just get this information into 03, get the most out of this model. And 03 is an expensive model. If you're trying to use it a lot, like this is a great way to kind of get more value out of it, move

fast, and and get good responses. Hey, if you take a look at my web presence online, it's safe to say that I'm a bit AI obsessed. I even have a podcast all about AI that you're watching right now. I've gone down multiple rabbit holes with AI and done countless hours of research on the newest AI tools every single week. Well, I've done it again and I just dropped my list of my favorite AI tools. I've done all the research on what's been working for me, my favorite use cases, and more. So, if you want to

steal my favorite tools and use them for yourself, now you can. You can get it at the link in the description below. Now, back to the show. When I demoed this to Matt like many months back, the thing that like blew him away. He was like, "It can do that." Was just showing him that like you're you're putting all that context together and then you're giving it instructions and now you got one button to copy that and then you go to your LLM and then you paste that and you're pasting so much information and I

think the average person doesn't like people who are just using chat GPT or even people who are coding with cursor like people are just testing it out they don't realize that you can do that that you can literally just copy and paste all of that context in there and that the LM gets that and it understands what to do. So I've just gone ahead and and I've got Claude open. So Claude, you know, in contrast to Chad PT, Claude is very good at following instructions. Like it's the best model at following instructions, I find. And

and so I've gone ahead and just asked me help change the dark mode color scheme and I pasted that in to Claude like this with this XML copy button. And I think this is another thing that that like repop does quite well is that it's got tools to kind of send information into the LM, but it's also got tools to to go ahead and write an XML plan and and it's going to create this theme selector and it's going to add these files and and change files for me. And what's cool with this is that

I can just go ahead and use Claude with my subscription and then have it modify all these files. So, it's based creating all these files and it can search and replace parts of files, too. So I don't have to re-update and re-upload the the whole thing, have it output the complete code. So a lot of models struggle with, you know, people noticing like, oh, this model's really lazy. It's not giving me the whole code. But like this kind of circumvents that issue because it lets the AI just kind of get an escape hatch and just

do what it needs to do here. You know, LLMs are are are beasts that seem to really like XML. So if I were to hit copy on this prompt here, I can open that up here. Uh, you'll see like, oh, first of all, all of the parts of the prompt are formatted with these XML tags. So there's the user instructions, you've got the file contents. If I scroll up, you know, they're all kind of separated out. So everything's structured in like an optimal way for the model. And you know, when you're outputting XML, the thing

that about XML is that it's like it's it's not like a a nested structure. JSON is, you know, what people are used to looking at, but it's it's something that like models struggle to output reliably. If you're using like chatpt or, you know, cloud on the web, like it's just not it's just not going to like respect that formatting the way you want it to. I think of it as like a information language. It's like that AIS can really simply understand. It's probably the best one, right? And it's it's easy for them to follow. So

I have like examples in my prompt on how to output this kind of stuff. Do you do you check everything or I try to if I if I care about what I'm doing like that's the you know what Karthy's saying is like you got to you when you're when you care about what you're coding uh it's very important to to to go back and forth. And you know sometimes when I'm when I'm coding like this like I'll I'll iterate like so I pasted this question right uh with 03 and often what I'll do is I'll

read through the answer and then I'll I'll change my prompt and then paste again into a new chat and and try and like see where the result is different because basically I look at like here's the output. Okay, I I actually don't care maybe about this copy link button. Okay, then I'll put specifically put a mention in my prompt to say like let's let's kind of just focus on this part of the question and kind of reorient it. And and that's the nice thing with this is that I can just hit copy as many times

I want. If you're if you're paying for a pro sub, like there's no cost to trying things. There's no cost to to hitting paste again. And except for that, you know, the paste limit uh with with 03 here. Yeah. So, you know, you just try again. You just paste again. Let the model think again and try things and and I think that's that's like a really important way of working with these models is to experiment and try things and and see, you know, where how how does changing the context, what files you have selected, your

prompt. I use these these also these stored prompts that come built in the app. So there's the architect and engineer and and these these kind of help focus the model. They give them roles. So like if I'm working on something complicated, the architect prompt will kind of focus the model on on just the design and have it kind of not think about the code itself. Uh whereas the engineer is just the code like don't worry about the design just kind of give me the code. Maybe you should explain like when you say engineer prompt it's

literally you're just adding stuff that you copy and paste into the LM saying like you're an expert engineer and this is what I expect from you. expect for you to give me XML. That's your job. Do it. And that's that's literally how the LM's work. Like, okay, I'll do it. Absolutely. Yeah. Giving them roles is is crucial. Telling them who they are, what their what their job is, you know, what their job description, you know, what do I look for, like giving them a performance review, evaluation, uh, all that stuff. Like I I find like

the more detailed you are with your prompts, the more you can help. Like they kind of color the responses interesting way. So just adding the engineer prop you see like it spent more time thinking about it but actually that's partly because it's trying to add call tools that are invalid but um just as a contrast to the old response. So here this time it kind of said okay this is the file tailwind here's the change and this is this is the change that I'm going to do in a code block and so like just having

it kind of answer. So you know for the longest time before I had any of these XML features I was just kind of using repo prompt and like getting these kinds of outputs and then just copying them back into my codebase manually and kind of reviewing them. That was like really the antithesis of vibe coding where everything's kind of automated. Yeah. So, you know, like I I think you know the the working on repo prompt it's it's really like there's there's building your context. That's like the biggest thing. Just picking what you want. You want

to frontload that work and and you know in contrast to using agents you're going to have those agents kind of run off, do a lot of work, call a bunch of tool calls. You see like 03 kind of thought for 15 seconds, thought through some tools to call. Um it didn't really make sense. It just kind of kept going and and ended up doing this. And if you've used, you know, cursor a lot, you know, you'll see like often using 03, it'll call tools. It'll like read this file, read that file, read this file. But

if you just give it the files up front and and you just kind of send it off to work with your prompts, you right away you get a response and you're like, okay, well, does this make sense to me? Am I able to use this? Um, instead of a little bit more work, at least right now, but it's yeah, I think you get a lot better results. So it's Yeah. Yeah. Yeah, just frontloading that context, being able to to think through and iterate on that and and that's the whole philosophy around it is just like

thinking through like making this easy. The context builder helps you find that context. Uh you know eventually I'm going to add uh MCP support so you can query documentation find find things related to to to your query as well and just spend time as an engineer sitting through what do I want the LLM to know and then what do I want it to do and then make that flow as quick and as painless as possible and like that's kind of everything and I think you know going forward and you know as you get serious coding

with AI like that's that's what the human's job is in this loop this engineer's job is figuring out the context. I think that's that's the new software engineering job. I'm just like I said before, I'm just so surprised that a lot of people haven't talked about this cuz like for me like right now cursor is good for like something very simple like okay change some buttons or change some links or change whatever you know but anything complicated repo prompt I got like way way better results. So I I'm curious like you know have you ever

thought about like this being used for things outside of coding and do you think it would be useful for anything outside of coding? I I think of like a in a corporate context of they have all this information they're trying to use like let's say it's for um you know and there's different divisions of a company like let's say one's like a marketing division uh it feels like there's probably like you know instead of like just trying to feed all the company's information to the LLM probably they should have like the relevant like okay for

this kind of content what is my relevant marketing guidelines brand guidelines other information that's useful to the LM to tell it like writing style and things like that. Yeah. I mean I've I've gotten academics reach out to me tell me they're using it for their work. Uh there's folks in different fields for sure. Like in general, like you know, if you're working with plain text files, you know, repop can service those use cases for sure. Like it's it's all set up to to kind of read any kind of file and then apply edits to any

kind of file too. Like I don't differentiate, you know, distinguish or anything like if it's if I can read it, then I'll I'll I'll apply edits for you. And and and yeah, like I think a whole bunch of work is around just like gathering context and kind of iterating on stuff like even you know in in doing legal work. I I do think though like there is, you know, a flow that is still missing from this app. It's just that like kind of collaborative nature. I think there's still some some work that needs to kind

of be done to kind of make this a more collaborative tool, make this a tool that that kind of syncs a little bit better with different things like for now like developers use Git and like that's that kind of collaboration bedrock. Yeah. I mean, yeah, that's something I think too is like yeah, repo prompts super useful but you kind of you have to be a little bit more advanced like than the average vibe coder, the average person using an LLM. And uh I'm kind of curious like why did you not go the VC route or

like and where's Rainbow Prompt at right now? Like what is your kind of you know where is it now and what's your plan for it? Yeah. I mean I you know I I've had a lot of folks you know kind of you know bring that up to me and they're kind of thinking through like you know why not VC or whatever. And I think like I mean it's it's not something that the door's closed on forever. It's just I think right now it's it's working. I'm able to build and you know I I'm able to

to kind of listen to my users and you know pay attention to what they need. And I think you know it's it's it's just not super clear to me like where where this where this all goes you know like this is an app that is like super useful and it's it's like helping me and I'm I'm able to build it. But like is it is it something that necessarily makes sense to like have like you know $100 million invested into it to grow a huge team to like build maybe. I I don't know. So right

now you're still working at Unity. Is that right? So you're like working at Unity and then building this at night while like having a family. So I was in paternity leave for a huge part of the last year. Um and that's kind of like a lot of the time I was able to build this out on. So I so you know raising a child you know that's that's a lot of work too. So yeah a lot of it right now is still just like you know nights and weekends spending a lot of time on it.

You know I feel like everything with AI right now is like who knows what's going to happen like yeah in a year everything could be different in 5 years. who the who the hell knows right and in that kind of situation um I think there's a lot of risk like right now because AI is such a big wave that's why we call the show the next wave right it's such a large wave of transformation happening that you are going to see the largest investments ever I think in history as well as the largest acquisitions ever

and I think these are have yet to come like people think oh it's already these have yet to come we're like like in the early part of this this transition and so with in that situation I I I think you definitely should consider it because there's going to be so much money thrown at the competitors and I think people like Kerser Windsurf maybe Windsurf's getting bought by OpenAI possibly uh for crazy amount of money. Um they'll probably try to integrate Wind they'll try to probably integrate repo prompt type features into their products at some point.

I would assume the best two routes for you in my opinion would be either to go really big and go the VC route or to go more like, hey, who knows what's going to happen with it. I just want to like get my name out there and I can leverage my name for something else in the future and like open source it. That's like my my kind of thought on strategically what I would do is like either go really big or open source it and make it free and just put it out there and say,

you know, and get some reputation benefit from it. Yeah. I mean, well, there there is a free tier. It's not open source. Um, but but there is a free tier. You know, the thing the thing about open source actually is something I've thought about a lot and and the the big issue with it right now, especially as people are are building AI tools is that like it's never been easier to fork a project and kind of go off and just build it as a competitor. Like if you That's what I'm saying though. Like like I

think that you're like kind of getting up on like the monetary side of it. You're literally just making the reputation like hoping to get a reputational benefit from it versus monetary. But even that reputational side like like if you if you've looked at Klein like Klein's a big tool you know that that that came around actually started around a similar time as me working in repop. Um and if you're not familiar the folks listening on the call like Klein is a is a is a AI agent that sits in VS Code and and and it's

pretty cool but the thing that is not so cool about it is that it eats your tokens for lunch. Like that thing will churn through your your your wallet like faster than any other tool that exists just cuz it goes off and reads files stuffs the context as big as possible. You know, if if like cursor is trying to like be conservative with with what it serves the model to save on their inference costs, like client's the complete opposite and will maximize your costs as much as possible. Sounds like an open source project. Yeah. Yeah.

But it's really powerful. It's very powerful. It does a lot of good. A lot of people really enjoy using it because it has good results for certain things. But yeah, that cost is very high. And but the thing that I was trying to bring up with this is that like so client was actually forked a few months ago by by another team of developers and it's called brew. Brew is the the uh the alternative. And a lot of folks, you know, and if you look at open router and some stats, like RU is actually surpassing

Klein. And so the the you know, that fork is is now overtaken the original. And you know, that's the kind of space that we're in where like different teams will kind of take your code, take it in their direction, and then all of a sudden they've overtaken you and you know, you kind of lose track of, you know, where things are going there. So like it's it's it's a crazy space and and it's never been easier to open pull requests with AI. You don't you don't understand you don't need to understand the code. You're like,

"Oh, I have this open source project. I'm just going to fork it and add my features and kind of go and and you know and like it's a tricky thing but like you know having a free version and and kind of trying to ship and and try and grow a community of users who are passionate who like can talk back to you and and you know I mean that's kind of the route I've taken right now and it's it's it's kind of been working so far. I was in beta for a long time but yeah

you know leaving beta I you know it's it's like you know it's still new figuring out where to go next with and and it's Mac only right now. Is that is that correct? Yeah that's that's true. it is only Mac is Mac only and and I think a part of that is is that I started off you know just kind of trying to think about like you know how do I build this in a good way like I started with Electron actually building this out and and and the problem is like I immediately ran into

issues trying to build for different platforms and like I spent a bunch of time debugging just getting SVG icon rendering you know all these little things that are just like rabbit holes and you're like okay well you're so abstracted from the base of like what's happening and you spend a lot of time just solving build issues that it's like I'm just going to go ahead and do build native and just just run with it and and have better performance doing so. Like, you know, if you open an IDE like VS Code, you you you you

open up like a huge repo. What actually happens is is it'll load the file tree and it will it will just kind of lazy load everything. Like not everything needs to load because if you're opening an IDE, you know, as a coder, traditionally, you only have a couple files open at a time. Maybe maybe you have a dozen, but like that's that's at the most like you're not you're not going to be processing 50,000 files at the same time. But an AI model can you know if you give it to Gemini like an Gemini will

will want all those files. It will want as much as you can give it because it can read all of it. And and so like you need a tool that is built different that is kind of organized in a way where it's kind of thinking first through that performance of mass data processing um that you need to kind of do. So yeah, it's it's it's a whole different way. That's why that's why it's native because like I I want that performance processing all these files. there's all this concurrency happening where you're like in parallel editing

these files like processing them and doing all this stuff like you know it's it's very hard to do if you're just you using JavaScript or TypeScript. Yeah, I assume so. I had thought about that like when I use repo prompt it seems like you've done a really great job of building it. It works really well and it it is all just you like right now. Yeah, it is just me. Yeah, I've been working on it a lot. That's crazy. That's crazy. Yeah, it's it's come a long way. I iterated a lot on it, you know,

but like you know the first version was super jank. It's just, you know, I But that but that's the power of dog fooding, too. Like if you're not feel like folks listening, dog fooding is when you like kind of use your own product to iterate on it and build with it and you kind of like make it make it a habit of making sure that you're a number one user of your app, you know, your own product to make sure that you see all the stuff that sucks about it. And for the longest time, like,

you know, it's really sucked and and and just that struggle and that that pain of of using it and forcing yourself to to feel that pain, like that's what makes it good. That's that's where you're able to kind of feel feel those things that the other the users using the app will feel and and that's when you end up with something that that is great in the end. So where do you think reput prompt is going like long term? It it's weird you know like in December like OpenAI announces 03 and they're like oh it

beats all the ARC AGI tests and you're like well is this AGI like what is this like? And then and then it shifts and it's like okay I mean like it's a better model. It it lies to you. It's it's it's not like the messiah you know. So, so it's hard to say like I I don't I don't know like where we go. Like I have ideas on like you know using LLM so much. I have ideas on like where the future is. One year from now I I think like you know I I'll have

to adapt this product and and keep iterating on it to kind of stay relevant. So it's going to keep changing but like I think that the flow of kind of I'm kind of pushing towards of that context building I think that remains relevant for a while longer and and what improves is the layers of automation around that work. So I think like long term I still think that is kind of the vibe that I want to go towards. Though of course like you you'll have to you know have an agent that kind of like does

more research for you on your behalf that will kind of run off a little bit. Being better at doing parallel work you know some folks using cloud cloud code like they they have this thing called the git work trees that they use where basically they they split the repo into like parallel universes and they let like cloud code kind of try things on these parallel universes and iterate and go off in the distance. That kind of workflow is going to be important. very expensive workflow, but it's it's one that that like I think is is

one of those ideas that you'll kind of want to explore and go down that rabbit hole. So, I think just like integrating MCP, just like embracing that like universality of all of these different tools. So, for folks listening if they're not sure what is MCP is another acronym. Yeah. Yeah. I mean so so the idea the idea there is is like basically traditionally if you use like claude or open AI they have tools and those tools you know one of them could be like search the web or one of them could be like read the

files on your thing or look up documentation or these kinds of things and and there's this protocol MCP that like creates like an abstraction layer so that like any client app can can implement this protocol and then users can bring their own tools. So if a user comes in and says like oh I want to use and there's this new one that's really cool it's called context 7 where basically they've gone ahead built a server that fetches the latest documentation for whatever programming language you're using and and we'll kind of pull that in as context.

So you can say okay great fetch fetch the latest you know angular docs or whatever docs you you care about and then you can bring that in. So that kind of work where you're like doing that that that context retrieval that's super important or like you Stripe has one too where basically all the docs for their their tool is set up and you know you just plug in the Stripe MCP and then all of a sudden if you're trying to vibe code your way through integrating Stripe like that's super easy that the work is kind

of handled. You can plug in your API keys onto it so it can even talk to the back end for you. Uh that whole work is kind of automated. So it's really like it's all about like having tools for for folks using these models to kind of automate connecting to different services in this like universe of all these different you know services. Yeah. I kind of think of it most I mean it's different than XML but for me the way I think of it cuz I'm like I'm a person who's like more on the business

side slightly technical code a bit uh you know I think of it as almost more just how XML is like the information language that AI can understand. The MCP is like the same thing with any any service you want to use or tool. It's a way that you for the AI to know how to work with those things uh very clearly. Yeah. And and funny enough you mentioned XML because that's actually one of the things that I do a lot with reform is is like parsing XML and like I've show I've shown with that and

I think one strength there that I have that like a lot of other tools are kind of ignoring. So traditionally when you're when you're working with these language models as a developer uh and you say like and you see and you can see this if you use chattv you be like hey like search the web it's going to use the the search tool and you'll see it say tool search and it'll go through. But what happens when it's doing that is that basically it it calls that tool, it stops, waits for the result, and then

continues. And when it continues, it's basically like if you think about it like in in this weird way like how tool calling is it's it's funny like I think of it like you have this the the the robot is is kind of being reboot as a new session uh with that new context because basically every tool call is a new query. So you're giving back the old information but you're not necessarily talking to that same instance. It's like it's like a different AI instance that is answering your question from the new checkpoint. So like that's

like a weird thing. So you know as you're as you're making all of these tool calls if you if you use cursor you know it'll make like a 100 tool calls in in in I think they have a limit of 25 actually but by the end of it you know you've gone through 25 different instances of these models and then you get a result at the end and you're like well you know it's like weird like what actually happened you there's some data loss like weird stuff you know we don't know how all of this

yeah it does seem like that could create like reliability issues right because like you know with LM like sometimes they give you amazing results and other times it's like what is what is this and so every time you're doing a new tool it sounds like you're you're almost recreating that the chance of it going wrong in a way. Exactly. Yeah. You're you're aggregating this issues but you don't even know where that information it could be different servers that are actually processing all these different tool calls and you know maybe you know it's weird sometimes you'll

have like oh that server has some like chip issue on its memory and like that actually causes some weird issues where cloud is actually really dumb today. Um but on the other one it's it's a lot smarter cuz their chip the memory chip is working fine you know you don't know. So that kind of thing. So the the way that I've kind of gone about this is that like the way I call tools is you you have your XML and the AI will just answer in one instance and and and it'll just give you the

whole thing and then I parse it and it can call a bunch of tools in there. It can be like hey like I want to call this this do this and this and then I just parse that and then bulk call the tools and then get the results and then we go another instance with the results and you can kind of go back and forth like that. So like not have to wait on each single one. You're actually just bulk sending them out getting that data. It's a lot more efficient. you're able to to process

say like 25 queries you know get read 25 we'll bring them all in you know let's work from there and see how it goes and so that kind of thinking so I think there's there's a lot to to kind of play with in in terms of you know how you're even getting this data back and forth from the LLM because at the end of the day it's all text you know text and images maybe um some video in some cases but like really text just for your coding like that's that's the thing that that you're

working with and you can do a lot with text manipulating it and playing with it to to kind of get good route what do you think uh like I've heard you know YC and others I think Gary Tan said that I can't remember if it was 80% but I think he said like 80% of the code for the the startups going through YC right now is AI generated that number could be wrong like where like do you think in 3 years from now like like in 3 years from now do we still have like normal

engineers who don't use AI at all is that a real thing well first of all like I think saying a percent like that of how much of it is AI generated it's a bit misleading because like I can basically like no but like I can go ahead and like every line of code I could basically like type it like in pseudo code to the AI model and like have it paste it back in uh as like a fleshed out JavaScript function and say 100% of my code's written by AI, you know, and you know, if

you're it really depends on how how your workflow is, what your pipeline looks like. I do think fundamentally the job of an engineer has changed. It's already done. It's already completely different. Like you can't you can't work the same way, but it depends what point in the stack you're working on. Like I I work for some folks who do some like really low-level, you know, graphics, you know, work and and I I talked to someone about like how they they can't really use AI models because the AI models just hallucinates everything like they it is

just not trained on anything that they work on. So it's just useless for them. But then if you work on work with someone who's a you know web developer, well 100% of the code like like 98% of the training code is web web code and web framework code. And so it's like okay well yeah 100% of that work can be done by AI. It's really easy. But I think like as we move forward more and more, you're going to want to have AI models thinking through hard problems for you because it just happens much faster

as they get better at math at solving like you know connectivity architecture. Like architecture is something that like these 03 and 01 Pro and hopefully 03 Pro is just excel at. They're they're very good at finding good ways of like organizing your code and helping you plan how you connect things together. The the job is is fully changed. Like I think from from today on like if you're not using these tools, you're not learning how they work. Like I think that that's like an issue cuz like I don't think you know a traditional engineer who

spends his whole career just typing code up like that doesn't exist anymore. But what does exist is someone who understands code and who can read it and and who understands you know what questions to ask. And if you're if you're prompting like you know about the code if you're if you if you understand you know the connections that that's where you're going to get the best results. And that's why like a tool like ruber prompt is so helpful because you're able to do that and feed the right context in. But if you're just saying like

make the button blue or like move it over here, I mean that works for to some extent. You know, if as long as your your your instructions are simple enough and you know what you want, you can get there. But like at a certain point, you fall off and and and you know, that's when it stops working. And and maybe that point where you fall off gets further and further as the models improve. But I don't think that like in the next 10 years, we get to a point where that point stops existing. One thing

that we didn't talk about that I was kind of curious to talk about was like what what do you do at Unity? like and and like a crazy backstory like so I used to be involved in the game industry and I used to see David Helguson the creator of uh you know Unity around parties all the time. He's super nice, super nice guy. Uh like I had some kind of like image in my mind of him was like this uh eccentric uh you know uh European billionaire who had like the popped up cuffs and he'd

always like you know uh he'd go around with like loud techno music with his you know the roof down and uh it's crazy how big Unity is now. Yeah. Yeah. I mean, so what I do there is like I so I've been I've been doing uh kind of XR XR research and XR engineering and uh so I work on a toolkit called the XR interaction toolkit and and basically what what that is for is it's a it's a framework for developers to to build interactions with XR. If you're if you're putting on an Oculus Quest

or your a hollow lens or or you know like uh whatever you know Apple Vision Pro you want to basically interact with objects in your in your scene in your world you know like in in AR if you're walking up and you want to pick up a virtual cube like how do you how do you process that interaction of you grabbing the cube and picking it up and looking at it so that's like I've done a lot of research on that that interaction of like playing input hand like I I've written specs that are adopted

for for like the industry in terms of hand interaction. So like you know just tracking your hands. How do you how do you grasp something? What what should you be doing if you want to poke a button that's like not there? Like what does that look like? Uh so that kind of stuff. That's that's what I do there. That's amazing. That's like that's a really complicated engineering work. I'm I'm constantly this actually one reason I even like you know helped start this podcast was I'm constantly thinking about where AI is going and wanting to stay

ahead and also think about what does it mean for me and my family like with your child. Uh you know have you thought about that yet? Like what course what what do you think they should learn and like and and how I have no idea. Okay. Okay. This this everyone right like like what do you even teach your children? Like is it is it important to learn to code? Is it do we teach them logic morals probably all of this and more and being flexible and super fluid and Yeah. Yeah. You know it but it

is funny on that topic like I I look at engineers coming out and learning to code with AI around and I think they're at a disadvantage. You know, it's it's unfortunate that like, you know, if you're starting to code today and you have AI to lean on, you just don't have that struggle. You just don't have the pain that like I had to go through when I started to code when, you know, when engineers who've been in the field for so long that had to struggle and and and and not get the the dopamine hit

of a fixed problem right away, like to to study it and understand how it works, like that just doesn't exist anymore because the AI just solves it for you. And I think that's true in code, but it's going to be more and more true in every field. And so I think like you know if you know if you want to really get into something I think like there's going to be a need for people to have the restraint to kind of put aside these tools to struggle a little bit. I think there's a ton of

value in kind of using them to learn and grow. But there's also like that restraint that you need to form to kind of have the struggle because that's where the learning is and and it's it's really tricky and I and I don't know how you you solve that now because it's it's too easy not to struggle now which which is a big problem. Yeah. I've heard uh Jonathan Blow I don't if you know know of him the game designer he he talks about exactly what you're saying that you know it's in the future like yeah

sure AI could get amazing at coding in the future but it's also going to create an issue where like just like you said people are not going to learn to properly code he was already complaining about before AI the code was shitty and then now with AI it's like okay now we're kind of screwed I guess because like we're in a situation where like no one knows what's going on and like you're you're entirely dependent on the AI for for everything. Yeah. I But that that's the thing like I think maybe that's the middle part,

you know, where where we're at this point where it's like the AI is just not quite good enough to kind of solve all the problems and you still have problems to solve and you still have people that need to kind of work with the machines to kind of figure out how to go. Maybe at some point in the future it all of it is moot and I I know some folks think that and and maybe it doesn't matter. But like I think you know there's going to be some discomfort in the middle where where you

know the machines are not quite good enough to solve every problem. we lean on them as if they are and then, you know, we're kind of atrophying a lot of skills that we we've heard. I' i've he folks, you know, I haven't driven in in a Tesla with FSD, but I've heard folks say the same thing there where like if they're using it all the time, they actually like suck at driving without it. And it's like, you know, like more and more that's going to kind of be a thing where where like, you know, that's

that that is the thing where we start to go like to like where like almost living in like one of those sci-fi novels, right? Like everything being super super safe. You know, I live in Japan. Everything, you know, I used to live in San Francisco. Everything's super safe in Japan. There's one reason I like it, but you do lose some some freedom in in that. Uh, so Eric, it's been awesome. And, uh, maybe we should, you know, tell people where they can find you and, uh, where they can find repo prompt and Yeah. So, so

I'm, uh, you know, puncher with a V on X. So, it's like PVN C on X. Uh, and most most social, my handle all over. Um, so you can reach out there. My DMs are open if you have questions. And repop.com. So, you can just head over there and, uh, find the app. free to download and uh nice Discord community too. If you want to hop over there and send me some messages and tell me what you think like please do. Cool. I think I'm on there too which I'm not very active but yeah. No,

it's all good. Yeah. Well, thank thanks for having me on Nathan. It's been great chatting with you and it's been great. Yeah. Yeah. Appreciate it. [Music] [Music]