Behind the product: Replit | Amjad Masad (co-founder and CEO)
01:04:08 Share

The idea of what is that making software today is very difficult. We want to make IT easier. People view this as a developer and their pocket. Essentially, we have thirty four million users globally. There is people everywhere learning to code and rap, building startups, building personal software.

person tools for people, building products, the product managers, founders. Like, what skills do you see matter.

matter less? Typically your bottle act, where your ideas not fitting in need to be made? And do you be made quickly? Now you open up that. Like, so now I actually makes you things is a lot of easier. Actually, you become limited by how fast you can generate ideas.

I think people are unaware just how far things gone.

I imagine whatever five years from now, someone, a billion company, zero employees, where it's like the support is handled by A I, the developer is handled by A I, and you're just building .

and creating this thing, man, the future is wild. Today, my guest as amjad massad. Omid is the cofounder of replant and A I powered software development and deployment platform for building and shipping software is one of the fastest growing developer communities and AI products in the world.

There's a lot of talk these days about how AI is changing, how products we built our products teams are going to Operate, which functions will be more and less valuable over time. But I feel like very few people have actually seen what modern A I tools can do and a fully grasp how much you can get done with a very little technical scale e now and in the future. And so i'm going to do an experiment with this podcast where i'm going to do a series of behind the product episodes where we go deep on important products that product builders should be aware of and should probably start playing with in our conversation.

I'm job does a demo of what you can do with replay today, which is gone to blow your mind. And then we spend most of the conversation talking about the implications of this on the future product development, on the future product management and on future of startups and founders. It's a very exciting time.

It's also a very scary and destabilizing time for a lot of people. And my thinking is the more you are aware of what's possible today and where things are going, the Better position you'll be to throw in this very wild and crazy future that is coming very fast. If you enjoy this podcast, don't forget to subscribe and follow IT in your favorite podcasting, uber, youtube.

It's the best way avoid missing future episodes and helps the podcast mendous sly. With that, I bring you i'm job. I'm job. Thank you so much for being here. Welcome to the podcast.

It's smart pleasure .

I thought would be great to start with. Just haven't you explain what is replay? What's the vision or is this going? What job does he do for people?

The idea behind rapid is that making software today is is very difficult and we want to make IT easier. Uh, one of the reasons for the difficulty is that, that is very fragmented. So you would need to download what's called that ID. It's basically a code editor. You need to download ad the run time, basically python or office script t need to figure out a package manager to figure your kind of open source packages.

And once you've done all of that, you need to figure out how to deploy IT, how to share IT, how to and so it's it's a very hard process, and that's one of the ways where people get stuck and never learnt how to code because IT just feels like this conversation IT process. And so vision for record has always been is like, okay, making softest fun is great. More people should do IT.

But um so for more people to do IT, IT needs to be an easier to do that needs to be in one place and need to be learned. It's easier learn. So that's the product today. IT is I think one of the easier ID is slash environment, slash deployment environment, uh on on the internet and and I think we make IT really easy for people to just jump and even without prior experience of coding, especially now with the new AI policy that we built.

This episode is brought to by work. O S, if you're building a set up at some point, your customers will start asking for enterprise features like saml authentication and skin provisioning. That's where worker west comes in, making a fast and pain list add enterprise features to your APP.

Their apps are easy to understand so that you can ship quickly and get back to building other features. Today, hundreds of companies are already powered by work, including ones you probably know like vir, sel, web flow and loom worker. Us also recently acquired warrant. The fine grain authorization service warrants product is based on a ground breaking authorization system called zenza, which was originally designed for google to power google dogs and youtube.

This enables fast author ization checks at enormous scale while maintaining a flexible model that can be adapted even the most complex use cases if you're currently looking to build role based access control or other enterprise features like single sign on scheme or user management, you should consider work of us. It's a drop in replacement for odd zero and supports up to one million monthly active users for free check IT out at work O S dot com. To learn more, let's work O S dot com.

This episode is brought to by persona, the adaptable identity platform that helps businesses fight fraud, me complaint requirements and build trust. While you are listening to this right now, how do you know that you're really listening to me leny these days, it's easier than ever for fraudsters to steal pii faces and identities. That's where persona comes in.

Persona helps leading companies like linked in ec and tWilly o security verify individuals and businesses across the world. What's its personal part is its configuration body. Every company has different needs depending on its industry use cases with toilet and user demographics.

That's why persona offers flexible building box that allow you to build tailor collection and verification, closed maxim conversion or minimizing race plus personas. Orchestration tools automate your identity process so that you can fight rapidly shifting fruit and meet new waves of regulation, whether you are start up or enterprise business. Persona has a plan for you learn more at with persona dr com slash.

Let me again, that's with P E R S O N A dark com slash. Leni, what's the scale of that at this point? How largest has gotten? I want people .

are using IT. We have thirty four million users globally. We have a large uh global presence.

There's um you people everywhere learning to code on a building startups. You building personal software, personal tools or international tools of companies. More recently we've been uh expanding to companies. We released our kind of B2B pac kage and in jul y and has bee n gro wing rea lly fas t. It's been really fun to see people bring rapper to work as well.

Damn, I knew was popular. I and realized with that large actually um as I was preparing for this podcast episodes is a tweet that kind of by all where there's a guy jovan, who actually know I know this guy from canada he's awesome, treated about how was eleven year old girl build an APP uh in replay SHE just like had an idea and SHE built IT and the best part of is someone in the like reply him and they're like like you have to like launch and happy have to like hosted summary. You have to build a data abase you have to like the ay there's no way to that is like, you know that's exactly report yeah .

that's what we do everything that that that a commenter was talking about IT and he's right, right. Like it's it's the surprising thing about an eleven year old building an APP is not so much even the coding IT is like all the nonsense around IT. And so we we just abstract all that way.

And that I struggle with that myself. And I was an engineer way back in the day.

how you are an engineer. I was an engineer .

for ten years. I was an engineer manager. And then I jump ship into product.

And wow, i'm happy I did, but I do. I was I was not an amazing engineer. I was like a good enough start of engineer.

So this is the kind of stuff that I would have left to use. So we're going to jump up to a demo of what this actually looks like. I thought maybe actually before we get into IT, there's other tools that people are aware of that help you build stuff. And so to kind of put a finer point on like what this does and how it's different from other things you may have heard of, say cursor comes up like these days talk about like a little bit about the competitive landscape of who also is out there that helps you build product.

Um again, we go go back this idea of light and to end platform for making software so that slight from writing code all the way of deploying and modernizing and all of that now every step in the process of the software involved in life cycle, there are a lot of different tools. So curr a fork obvious code that made um that has really awesome A I tools, but that's an editor. You still need a run time.

You still need a deployment environment. Actually quite a few users use use curse intended with rap because rappid just simplifies the untied and deployment to environment. And so you have products uh your AI products um in different places in the software development life cycle.

But really what differentiate rappid is that we we do everything but but also you know that makes IT harder to adopt for for certain people, like if you're at a big company is very easier to bring a new editor, uh, and start coding with that is quite hard to build to to bring you know something that's quite opinion about about everything a from the from how the code runs to how the code deploys. But that's a trade off we're willing to make is like, yeah, we're not going to get into into the enterprise you know main soft development pipeline, but we want to empower everyone to be able about the bells software. And that means product managers, designers, we have you know Operations people, sales ops, hr ops, we have lawyers uh using using rapid. And so IT is IT is democratizing the act of soft international.

amazing. And that's where you're here. Let's do a demo while you're putting IT up, you're going to share your screen and show us what this product can do. And the reason I am excited about doing a demo, this is an experiment, kind of a new type of pot test steps that i'm doing or diving in a specific products and what you they can do.

I feel there are so much talk about A I and what's doing in people keep breathing about, oh, I can do this and ah I can do that and I feel like not many people actually like see the stuff and action, especially the most cutting edge stuff like I think people are unaware just how far things have gone and how much is actually possible, especially when someone that knows what they're doing is using the products. So meet IT to show people what is actually possible and especially because this is going to impact the future management and product teams. So already over to you. Give us a demo.

awesome. So this is replete a home page. You can create what's call the rapper, which is a project we have all sorts of languages you can, you can pick from, really, in the hundreds. But most recently, and this is how our life became, like a thousand times easier, is you can just describe what you want to make.

So go on the home page, we have this tax box and you can you can write something like, make me a cool APP uh, what have you but um you know a more descriptive prompted is Better. And so, uh, I asked A R P M at rapid amon mother, a fan of the show, to tell me what P M S like to build. And so he came up with a prom to kind of, uh, really crafted a great problem.

I just put IT here, uh, and basically what he what we're asking for is we want to build a web application. You can you can say what stack you want to use or you can leave IT up today. I to decide here we are saying you build that and no jazz for product measures to track feature requests on a public dashboard.

So I say, you know, I have a product of growing. I have a community. I want that community to engage with building the product.

I went them to admit feature, request, vote on them. I want to be able to manage that. So you were talking here about the features of voting system.

feature requirement of you of them, just for folks that I am watching a youtube given sense of some of stuff and .

prompt so so a feature request submission, so allowing the users to add features of voting system, so allowing users to up for these features, feature requests and sad stacking. Being able to, uh, is like a camping style board with columns like planned and progress. So that way the adman can can kind of share with the community, with their building. And we wanted to be user friendly design, so make IT modern and all that nice kind of from tea things and then uh, add controls for product manager. So you know, as a product measure, I want to be what kind of really managed this this community.

And that IT builds the internal tools too.

not just the front. Exactly, exactly. alright. So we're going to start building, uh, since this is not like a pretty big, big prompt, uh, the initial coding might take a while.

There's different styles of using them. Plate agents. A I often go with like a minimalist prompt. T that's also how I code as well.

Like I have like a big idea of what I want to build and enter IT from there. Other people product manager is like to like write prs and like more descriptive things, and you can do either of those things. The AI now responded.

None said, you know, i'll build all of that for you. I'm going to build up the initial prototype and you can tell me how IT feels and then we can make a Better from there. The I is also suggesting adding common threads and playing email notification tions and so I can select those.

And now it's being creative is selling me what else I could build. But for now, i'm just going to go with a prototype and then we can assess from there. So as you see, as the prototype is starting, you can see this progress pain where we can watch the I doing its thing.

So here is created a progress database. Obviously, when we are building a full stack application, you interview about save things ah. So this is one to cool things about report.

We have all these services storage database. So now it's coding. It's building the database schema. Uh now is building the um the home home page. And as actually you quite fun and itif ying to watch watch IT build us because you can uh you can really start to learn how to structure web apps.

And you know if a IT runs into a problem and you know as things get complicated, might run into into a problem and you want to be able to like help the bug and things like that, good to be to be able to have an idea of what's going on. But it's not necessary. I think a lot of people just don't care about the code and are still able to to build things. But we want to make the process transparent. We want to show people exactly what asian is like.

You're basically sitting there behind an engineer on a computer and just watching them code is what the exact feels like.

Yeah and and actually the way we built IT is like it's a multiple er system. So a rapid is you know has real time when we call a multiple are coding. And um we reused this the multiplayer system to build the agents of the agent in the code is structured as a another user of the profit.

So you basically were both coding together. So I can go into the files here and that that's the thing that makes rap really cool. I think people are familiar with some of the more like chat interfaces are like v zero and others where it's purely child. But this is like a full I D E where you can like go and look at the files and and add them yourself or ask ask the I for explanation.

What's kind of the limitation of what this can do today? Like what can you do say you're like a zero coding experience. What sorts of products can you not yet build with something like this that might be possible in the future? How far does this take? You know.

you know, you can build A M V P. I think you can also start to get some initial users. I think what is when you start iterating on the the product, like large iterations, you might run into problems, for example.

You know it's not very good at database migrations, and so we're trying to fix that. So you know a lot of when you're rating of the product, a lot of times you're actually um you're changing the the structure of the APP and that that pores into abase migrations. So now like IT might change the database in a way that create an error that's unrecoverable.

Uh at at that point, you might you might get stuck, especially if you don't know how to code. Some people will figure IT out by going to ChatGPT and claw and like asking questions and like I actually am really inspired about how persistent some other users are, which is really amazing. But I think, yeah, that's my Young to and an MVP pass the MVP where it's like a product that's working and he needs to change and generate on that.

It's self struggle now. But I expect over the next a few months, well will continue. It's like if you think about IT, like sort of we're building, we're building as you are building, so we're building out the agent so that I can continue getting Better as our users are also building their applications. Got IT.

So what i'm here is it's really good at building like the first version and helping you get to something that you can even have people use. It's not amazing yet IT evolving from there like using A I to help you make the product Better and Better Better iterate yes, but you can get in there if you now if you not a coat and .

take IT from there, right? yes. Or you can hire someone. Um we have a feature on the site called bounds where you can hire human quotas to kind of help your finisher.

That's gonna the our job for humans like that remain for a while.

You know what we want to do. We we want to get to a point where a the agent can go grab A A human when he runs into problem. I think I think that would be.

that would be said, oh my good. It's like everything's reversed. I love IT. Oh look, I think that might be done. Check that .

out yeah so so now um the asian is asking us, is the application running and .

showing the home page confirming .

like yeah almost asking us to do ka, i'll just say yes. So IT found an error. So there's an error here. It's like there's a dumb warning i'm going to fix IT. So in the meantime, as it's fixing IT, so you know I can be proactive, right, because if you know looks at all the errors and and things like that, but in the meantime, we can use that. I just crayon an account.

it's coating .

and just wait and how long .

would you say IT would take an engineer to build like a like .

a typical engineer a few days I would say to a week um I mean if you're really good at IT might be hours but um but I probably would take me a few days. I would say I am like a decent engineer. I'll take take a few days yes.

took much like five, ten minutes .

yeah and probably like cost of something in in the sense yeah IT was a .

computer yeah like probably .

know I would estimated like fifteen cents or something like that. Wow.

here IT is.

Here he is in the agent was like, okay, this is looking good to completed if you want to deploy deploy IT. But i'm like i'm going to ask that first.

And so currently it's living just locally on your local host.

Yeah yeah it's not local horse is on a reply. But but yes, it's it's the equivalent of local horses because it's really easier. I can even invite you this session.

I know I can. You can be here with me. And so it's it's all all mine.

So let's make a feature. So make the product pretty. That's what a typical user might say. So we have this here. You can you can up vote.

I guess I I can't afford IT because i'm the user created, but if great, another user you can, you can afford IT. But but now you know we need to be able to move things around right as the admin. So I don't know how to log into the admin panel. So i'm going to ask me and how do I log in to the adman panel so I might have already built the future. And it's it's not exposing the right way or what I love about just like watching .

you interact with this thing and just really like call throughout IT feels like an engineer like that is behind the scenes building this thing like on slack just talking to them and they built this thing. They're like, i'm draw, picked this out, i'm done and they're like, ah okay, but how do I log into the segment that like.

okay here you yeah so IT as uh uh it's going to would you like me to help you register account? So it's is creating an t an add account for me, so only um built things. It's also um IT also maintains things, right. So in this case it's it's actually doing single queries, not writing code to create a to create an add account for us. It's insane.

I want to talk about the implications of this on product development and product management and founders. But just like what would we just witness to somebody I know you do have technical abilities, but someone that didn't have to didn't have to have any technical skill build like a real product that people can use like in five minutes that looks good and works and you could keep making them Better. I target this agent.

I tell you from our experience, like we're seeing, uh, like you know, there is so many products that are are empowering developers. Like it's a very easy calculation to say we're going to make engineers twenty percent Better and we're going to like sell to companies and we take ten percent of that value, right? Like that's why there are so many startups sound that are just trying to make engineers little Better.

Our calculation is like, well, know what, if you made everyone developing, what does that? What does that looks like? And so when we released the agent and really made programme a lot easier, what we're seeing is that people like exactly like you said, people view this as a developer in their pocket essentially. Um what we're hearing from customers is that i'm doing things I would otherwise have to go higher developer but also because the activation energy is is lower than going to hire developer, whether or porker other places, i'm building a lot more ideas that otherwise I wouldn't have built.

I know IT I think he was so called the javelins paradox or something like that which is like when the cost of things go down, uh the total consumption of IT goes up which i'm not sure why they call IT paradox but like um you know the a cost of electricity goes down uh maybe you'd expect that know that span the tools span goes actually tall span goes up because people consume more of IT. And so I think that's going to be the case. Software like as a cost go down, people just like make a lot more software to improve their lives and to prove their work and and start more startups and all of that.

So to follow that, the read. What are you seeing inside of startups? Are you in big company? Is in terms of how folks are already using this, knowing this is this is like the worst that will be and IT will only become smarter and Better right these days. What if, what how people actually using this, say, that are sap product managers are just like non technical people within start up, certain bigger companies .

on the sb side of things, a lot of people are building kind of back office a tools, right? So we have real city agents that you know have A A lot of data, have a lot of um thin they they want to manage in their business that are building a lot of these these tools the day otherwise would have to buy. But typically when you buy, it's is actually not exactly what you need, and that's kind of the problem.

Masses like it's it's like one size fits all. And so a lot of people are seeing at a sort of a casual placement for in house tools and and things I got. And then when you go to the like bigger companies um it's anywhere from from prototyping to to to to actually production apps are to tools as well. So um we've seen product managers build, like I said, like A V one of an APP.

And I should go out and tested with the users ah and I can name the company who put uh you know you there's a public company that that have used rapid to test A V one of of of an APP um and obviously after after that sort of work, they they take IT to the engineers and they're like, okay, we built this thing. We think it's a we think is a great thing. We test IT with some users.

H let's go actually put IT on the road map and add built built IT into the actual product. So you are a sort of unblocking product managers from having to need engineers for everything that they want to build, so and really built a py zero of one of the product ah and and and that's super empowering uh, for them. We thought also with like marketing departments, like a spot hero has a uh marketing a had a marketing that actually can code desipere well and but use rapper to to build build these apps.

And they built like a competitive analysis application that looks at um a competitive pricing and make sure that they are a beyond benchmark h correctly. And so it's a full stack APP use database and everything and IT runs on the contentious fashion. And we see cells engineers, uh you use rapid to spend up prototypes quickly.

Actually someone at x formally twitter um is on the side of partner engineering side of things. And he uses to spin up uh applications a protyle es for for customers to see how they can use the X A P. I.

I love this, all of these examples. By the way, the demo is anything else do you want to share about the demo?

And before we close, that was I created an add account. We can ask you with these name, password and kind of go into IT and manage IT. But basically that you know it's um the apps complete in terms of what we ask for. Uh, we can like send IT out. I can give you you all, we let this actually just deploy IT.

Really people link to the APP you could check .

IT out sounds good.

okay, cool. That's amazing. So this is displaying IT onto some like cloud provider. I don't know what is.

but we use google cloud, but we abstract, abstract all away. But but we use google cloud behind the scenes.

Imagine a place where you can find all your potential customers and get your message in front of them in a cost efficient way. If you're B2B bus iness, that place exists and it's called linked in, linked in ads allow you to build the right relationships, drive results and reach your customers in a respectful environment. Two mysore folio companies, web flow and census, are linked in success stories.

Census had a ten x increase in pipeline with a linton started team for web low. After rapping up on lincoln in q four, they had the highest marketing source revenue quarter to date. With linton ads, you will have direct access to and can build relationships with decision makers, including nine hundred fifty million members, one hundred eighty million senior exec in over ten million C U.

Level executives. You'll be able to drive results with targeting and measured tools built specifically for B2B int act lin coln and gen erated a t wo to fiv e x h ig her ret urn on at spe nd tha n any oth er soc ial med ia pla tforms. Audience is unlinked that they have two times the buying power of the average web audience.

And you work with a partner who respects the B2B wor ld you Ope rate in, make beef to be marketing everything you can be and get one hundred dollars credit on your next campaign. Just go to link in dcom slashed pod len, to claim your credit that's linked in dotcom slashed pod lenny, terms and conditions apply. What's good on this thread actually? Well, this is happening just like what what allows for this to be possible, technology wise? Like what is the kind of the stack, whatever you can share, that enables les, this exist that.

yes, 不是 first of all as it's the all the abstractions that we built。 So the way applets works, as you know, the very bottom later, it's um our run time. So this is the Operating system, this is the package manager, this is the language run times.

Uh, we build a system that uh is able to install packages in any language, including native packages to the AI. Anytime IT needs a IT needs a package. I I can go here and show one of those, by the way, day I can take screen shots as well so that I checks in the war. So here, uh you can see its a second screen shots, make sure the home page is rendering.

Here you can see um you know wanted to dragon and drop library and so IT installed and saw that and so IT has access to all all the packages across all languages and including a linux and all of that and the layer on top of that as the editor and the infrastructure that he runs the editor, including what I described as the multiple er editor and then we suppose all of that infrastructure to the AI and there's like a almost like a new discipline um called um A I computer interfaces. So sort of like H C I is now A C I and turns out like l ams need interfaces that there are actually quite different than than humans. Um they're trying to make them use humanity faces like anthropos computer use but those are really expensive and you need to kind of process all these images and video.

So instead we for the shell, for example, we give IT you know a sort of a tax representation of what the shall is doing at a certain increment um for for package installation, we give a certain to all for editing. We give IT a like like an editor tool that like when it's writing the code, it's getting feedback on whether their errors are not similar. What a human seas, but is actually like all taxes are just to make IT easier.

So that's a computer in her face. And obviously all of that sitting on on foundation foundation models 是 第 the improvement in in foundation models has has allowed allowed us to to to build this the uh the most important model that we use is the senate uh model from a laud uh from anthropic a and that is the best model at at coding。 So that's the model for coding.

But we use models from uh opening ee as well because it's a multi asian system. And so we have models that are a you are critical. We have manager editor model.

We have like critic model and different models will have a different powers, will also train some of our models like the embedding model for for search, uh, something we train internally. So you I actually wrote about IT back in like twenty two. Um I said it's going to be society models like products will be made of a lot of different models. And you know it's a quite a heavy engineering engineering project.

to say the least. We were talking off line and you said you've been working on this since two thousand and nine when you first built a first idea of replay. yes. Oh my god.

here's the deployed APP. Um I can I can send IT to you and you can use and you can you can see my my, my request even on the log out page so I can register unfolded, log in as admit and move things around. We can see what's in progress, what's completed.

Like this looks like a product. I can see designer spending like days, you know designing, passing IT to engineering pms, you know, having feedback, engineers taking a few days to build IT. Yes, and here is just a prompt. Here's what I want.

That's that's right. And we can interact on IT very easily. We can also interact the U. I. We can we can say, you know, I don't like this for that and i'll do a good jobs so we can go here. We can start a new session or like a new session to create an entirely new feature here. And i'll just do the right thing.

And IT built from that code base, IT understands. Here's what you've built. I wants to add this thing.

Yeah okay. And that becomes that becomes your sort of your history, right? Like this was the v one and now on more working on this new future.

And you know and and you know it's almost like what engineers do and get commit messages, by the way, and generates get commit messages for every for everything that IT does so you can roll back up as well. And and so we're china like make IT so that yes, it's is for everyone. But we're trying not to abstract too much away.

We want to build tools right for you to learn to use. And so we want power users to be able to understands the full power of of replant and is really deep product. I think you can spend you can spend a couple of years to to kind of master IT.

I want to talk about implications, but I want to come back to something you mention that, uh, is incredible that people may have missed. You basically built a computer specifically designed for the A I agent to use. That is a different version of a computer specifically optimized for .

how A I wants to use a computer yeah yeah yeah so you know there's an entire discipline are call .

like H E R N yeah .

so so now there are papers about AI computer h interfaces and interactions um and and until you know large length models are trained on large tax corpus from the internet. But they're still like kind of alien creatures. So they're not like humans.

So they they have different behavior. S and like it's unclear like what's the best way to give IT an editor? So there is so many experimental about like what's about sway to give IT of you on what's editing? How many files can you show IT like you know, before IT starts to illustrate ates and and right now it's like more than an art than science, but it's becoming more more like a .

science saying, so simple way to think about IT. There's this foundation model. Here's what I want you to build, and here's a computer to use to build IT. Yes, yeah.

Here's a computer with a set of tools. Here's a tool to install a package. Here's a tool to to edit the the code.

Here's a tool to run a equal query and also services. Here's a bunch of services you can grab from. Here's a database services. An object store service is an off service. Um so you think about a bunch of extra services, the computer with a with a bunch of tools and and they're all interfacing with the with the foundation model.

It's funny listening to this how IT starts to feel like, uh the fact that we might be living in a simulation is not as far fetched as IT may feel like this is feels like the beginnings of what a .

simulation computer would be. Yes, yes, I I you know it's pretty like, you know you you can t go reified on this and is like where where is IT had IT right? Like, you know, if we if we give IT enough tools, like, let's say, uh, you can I can drop IT and slack.

And instead of interfacing weather in this fashion, I went interface with a in a total autonomous way. So we actually have this feature coming up where instead of me testing IT, we give IT at another agent. So here, you know, set of me interfacing with that and saying, you know, the this is running or not running, we can give another agent that is actually testing the application.

And so and then, let's say, interface with out entirely throws lack and i'll say something like. Give me you know give me tail shrift tickets the moment they land. And so it'll build an APP uh that continuously monitors the web for for when tells lift a to get sand.

And there's like an agent that's using the APP uh to be able to get that. And then and you can imagine that has some kind of wallet or credit card. And then the moment IT lands, a kind of guess that I mean, what i'm trying to say is that software like agents being able to do software is how A I gets more general because software runs our lives, runs the internet, runs our businesses. And so the more competent A I becomes as software, the more general they are in terms of what they can do.

Okay, let's can go in so many directions and we didn't bring us back to the implications for people building products, the product managers, founders. How does this change that function, that skills set? Like what skills do you see will matter more, matter less which functions, or maybe in some danger? And they should start thinking about a different career path.

What one interesting persona that we're seeing is the C E O the C E O of uh start up the C E O adria Wilson son from from tiny is is a big user. Um and so these people are are typically new crayons ves, right? They built a company, they hired people. A lot of them like can't code. A lot of them are our designers of product managers or something else.

And and the you can imagine a bottom, like you can imagine the bunch of ideas in their head, uh and the ideas have to translate through them talking and then someone else listening to them, and like assuming that someone else actually understand what they say and that someone else going and trying to build what they want to, what they want build. And also assume me that person has has time, right? Because a lot of times your engineers are kind of stuck building the current thing, do not talking about the future thing.

And so what gets me excited, a lot of these CEO is, are building the future concept. The next, the next product that are going about, the next they see company, they're gone to build. And so uh, IT all locks uh, the creativity and again, sort of all blocks them from that and look at you to A V one of the product.

But you can push things forward. You can touch IT. You can feel that. You can say, okay, this is this really has legs and and we should we should work on that. You give up to our engineers and they can they can improve on IT from there.

So that's that's one personal um but i'm really excited about IT, the CEO slash slash founder um in uh in the uh uh companies um what what are the things that that I think is sort of hard about c tech companies is sort of these silos between designers, product manager, engineers and you know everyone feels that pain of kind of we have low band with communication, which is which is language which then taxed on on slack and zone calls and IT delete S A lot of frustration because it's really easy to misinterpret people. And I get leads to sort of silvering where people working on on something and then you pass IT on to the to the next team. And it's not really what they what they expect.

That happens a lot between designers and engineers. But but like the common language that h that um that everyone shares this code right? Like ultimately and soft attack companies, everything that we're talking about need to eventually flush out in terms of code.

And so so what if the language becomes actually working prototypes and working applications? Um for example, we have a figura extension that translate a uh figure a Marks into into react that runs on on rapid. So instead of instead of you giving giving the engineers, you just just mox or screen shots, whatever you just say.

So here's here's a bunch of ref code, you know just make sure that runs on our infrastructure but like don't mess with don't move the pixel right. Um and and so I think is just like opens opens up a asylum of the companies make communication around around product a lot more concrete because I can I can give you a working a prototype and that'll change how how people work. Like if you if you can imagine that everyone can can makes software, it's really kind of a radical reimagining of not just what tech companies are, but really what what most companies are. Because because you know, everyone can be more, more general.

So say, P, M, listening to this, an engineer, designer, what skills do you think if you you are one of these folks, if you are in building replay right now, what kind of skills would you suggest focus, focus on more and which you think you're just like a, this can be less valuable in the future. Don't worry about these sorts things that you can either pick one of those three functions or all three.

I think a very important scale that that's like perhaps harder to develop, but it's worth working on is being generated, being more general um being able to generate a new ideas and quickly because you know you can think about IT as as like a factory line, right? Like so so you have ideas. You have the the production of of these of these ideas or like the initial kind of production of of these ideas and then you have your other people that want to consume these ideas or work with you on on these ideas. And so typically you're bottle act by by the middle part, where your ideas are kind of like there a lot of them, and though not fitting in because like they need to be made and they need be made quickly.

And so now you open up that button like so now like actually making things a is a lot easier actually you become limited by how fast you can generate ideas um and so and and I find that true of of myself as well like you know I I concern myself a quite narrative but but now I have this tool and I can look, build build a lot more and explore a lot more and i'm finding that well actually i'm running out of ideas sometimes and so so so um you know training that a that muscle I think is a is a good thing um. I think like learning a little bit of coding and like not the traditional way of learning coding. Like when you go if like if you go to like a coding boot cap, they're going to start with like what is get actually my cofounder, how designer first building a raft let together, uh, he went to to a well assembly a to do like according course and the first day they were like spend this whole time on't get ah she's like what is that? Like what does he do? Like like I solder know get exactly that, but it's like um you're you're inverting the process, like you are giving the tool before the actual problem.

And so I think all of that stuff you don't have to worry about. So things that you don't have to worry about, I think a lot of the you know as A P M, as a designer and someone has not like in your connect every day, don't worry about all the tooling. And if you learn A A little bit coating just by, you know talking to any I doing a lot of bit deep bugging, building something with rap, you running into a problem and trying to fix that, uh, just using, you learn a bit of coding.

And you know, I have this, I have that's been called not by me dub as i'm jed's law, which is the return investment for learn to code is doubling every six months um and really just learning a little bit of that scale, learning a bit of scale about how to your prompt uh A I how to read code and be able to debug IT every every six months. That's netty you more more power because you're going to be able to create a lot more. You're going to be able easier to create, be able to create a lot h, you know a lot more complete a things. Um so so that's that's another scale that I think I could be could be necessary.

This is super interesting. okay. So the last point you made, i'm judge law. Uh it's interesting as when people like as something listening to this, I could see them being like engineers in trouble.

Why do you need engineers at this point? These agents are building the code. Your point is specific engineering skills are gonna incredibly valuable. And more and more about how often the doubling things you say every year, he said.

no, I every six months.

every six months, that these specific engineering skills are becoming more valuable. And the ideas this you don't need to like, know everything you don't need to know, the foundations would like to build the APP as much as more to unblock the agent and understand the mental model of how this stuff is built so you can move for fast.

That's right. This right, understanding the basic components of that.

I yeah yes. So it's like we need new engineering schools to teach you these very specific skills, yes. Versus spending years on like algorithm.

algorithms. And and I think I think no one has done that yet, right? And I hate this is this is like, you know big business probably really like to get built us like A A I native coding is totally different than than like traditional coding. No, that's why on hacker news, there's still much scepticism about like A I native coding tools because I can it's it's a glorified auto complete.

And I understand like if you're you know writing Operating system kernels, you know it's not really doing that much for you, but if you're building products, it's building that for you this point, right? And so you know if if you if you're starting a school to to teach I native coding, you would keep so much of computer science and that in the basic tools and um you would you would teach the basic idea of how to structure and up and then you would teach prompting uh and then you would teach I think a little bit debugging. I think debugging is quite a quite a good scale a right now to learn.

And interesting, if you want to be a good at debugging, you like there's a lot you need to understand, which is basically what you're saying is like that's the subset of things understand this, things that break. And to do that, you've understand our all works. What are servers or aps all he thinks.

So how far? So we've been talking about how this is very good right now, building a protoColing. Everyone, M, B, P, people can use IT. You can deploy, you deploy with that people can serve using IT. And there's them like up scale IT can reach. Do you see a future where you can build like a sales for size business, fully replay or other tools that can scale to hundreds of billions of dollars of value? Or is there just going to always be some limited like you need like actual engineers and designer sitting on the thing building and .

think you awesome if like my lie is like, you know directionally correct even even if the months or not, uh not exactly right that the duration is correct, you're going to see a compounding effect of of the power. Like it's actually quite hard to convince yourself, but if you really convince yourself that we are on a massive scale of improvement and N A I then then the answer is yes.

IT is like epsom to my engineer mind that i'm saying IT is but you know rake as well is like your future you know talks about how expansions are really hard for humans to to grasp. And so actually when we started building the agent, um I told the team it's easy and we fall in the strap before it's easy to build and optimize for today. You know in twenty two we built copied like thing in all to complete betraying our own models.

We optimize the hell out of them. But at some point like that modality was kind of you not the right madam, which is like the auto complete modality. And that and the right madam is actually this, I think, for now, as being able to chat inside the program environment and for the agent to create things for you. But in order for us to make that bad, you know, a year ago, the models were actually not there, like the model could not do this, but we were like, okay, we're gna build for the models that are landing in six months.

And truly like six months later, the model start to land that are capable of this of the reasoning that we need and and whatever and I was like your song if you want which is oh wow like we spirit ate and the reasoning improved so much and six months later you have thought of you too um and it's really almost like a six months kids. And so if if we're really on on the strategy ory, then you know I would say, you know, next year you're able to scale and maybe maybe you get your thousands of users paying you a the A I can do maintenance and already showed the A I doing like equal quiz and doing migrations um so I will never do maintenance debugging things like that. Um I think where IT gets really tough is that you know um when you hinting scale and you were in in a architecture system that resilient.

And so that means you know ah you start you are starting database and we start like using different queue systems and components and and things like that. And I think um you the AI needs to have access to the entire sweet of of tools to be able to do this. And and and I think that that's going to be the next bottle ic, and I think that I need to be a lot more reliable a doing that.

But I could imagine like whatever five years from now, someone running you know a billion doll company with zero employees where it's like the support is handled by A I, the development handled by A I. Um and and and and you're you're just you're just a building creating this thing that is you know that people are finding valuable on our paying you for IT. That being said, it's worth like thinking about the economics of IT.

Like if the you know if the cost of software goes down a lot like a, then what is, uh, was that the Price that you can charge on on software rate? So can you actually build the next cells force? If anyone can generate us force and then and then the question is like, what is the year? And this is why I emphasize being generous because I think then the the thing that will make you Better is like by being able to iterate and improve the thing really quickly and enjoy .

the new ideas instead head of all the other of people building these tools so quickly. Yeah oh my god. An interesting other kind of mental model, seeing as you talk about the sort of thing, uh, is not to offend religious folks, but there is a conservative god of the gaps.

I imagine you heard that, yes, where it's like god explains all the things that we don't yet understand, and over time that kind of space shrink. And god's like all the things we don't get yet those gaps. That was god that that proved that the CBA god, and he feels like right now, humans are like the gaps in these tools for these agencies. You talk about that you can hire within repleted, like fixing these little gaps. And over time, A I will fix these things themselves, this right.

and his gaps, or shrink. I mean, unless unless we had some fundamental limit and the current regime of of of A I, which i'm not i'm not an expert about like how far transformers could scale, but I I feel like IT is here. We found the thing that could that could scale pretty far, but maybe there are limitations in data or or other things like that, that we could be we could be surprised by. But if there isn't, then we are on a massive trajectory of removing these gaps are quickly.

Yeah, very true. We have no idea. We keep thinking it's just gona keep going, but maybe maybe it'll stop at some point. I could keep going and going, uh, but I think we should also let people go play with these things and process all the things we've been talking. Uh, there there anything else you think might be helpful for folks to think about or learn study?

You give advice to to founders or or leaders at at companies. The way we work is, is going to change rapidly, and it's important to sort of resilient to that change. One thing that I think is really difficult now is having your old maps, especially if you're doing anything in A I, but really any anything that eye could affect, you want to be able to react to IT really quickly.

And so you know when they um anthropic drop the computer use um uh set of capability, you know we slaughter roadmap because we don't really have IT explicit robot like immediate jumped on IT and set of building things. And we've launched some things around IT. We going be doing more with IT. But like there is going to be capabilities that that are gonna drop and you you want to really uh in some cases, if you really affect your business, you want to be able to jump on IT really, really quickly.

So you are being argile, not being sort of stuck with with road maps, be able to just just say h, which is going to switch priorities right away, uh, as it's going to be a super important not being um you know like I said with silos um and rap, there's so many people that are on the scale of like designer to engineer, designer product manager. Actually I mentioned on on earlier, he started as a designer at rapper and there was a product manager. We have people who start as designers become engineers and we have people in the middle we're comfortable with that, like design engineers um and IT that fit a different parts of the scale.

And the design engineers go to the design correct meetings and and some desires go the injury meetings and you gotta flew ID, right? Because you know again, when when designers can code and and engineers can design, I mean, it's it's really becomes you can't have a lot of structure on that. You want to build a culture and you want to build, uh, environment or mill. You that is like really, really flexible, which is comfortable for a lot of people.

Man, the future is wild. Everyone's everyone's a hybrid person. Now, uh, let me just actually doubled down on what you just said, which I think is really interesting. It's almost like if you're an engineer, where your skill set will become most valuable is unblocking these A I tools and knowing debugging and figure out IT on blow allow you to go further, further, further within P M design land, if based on which you're describing, where the skills will become more valuables, uh, generating ideas almost like finding opportunities, discovery, finding what prompts need to be solved and then articulating that is clearly as possible to the A I tooling.

That's right. So this is a very crisp sort of advice that that people can fall today.

What the world okay. Uh, i'm judge, this is incredible. Um my mind is racing. I've gotta build some apps immediately.

Yeah, too back.

And I will do that. So just to leave listeners with couple things. One is just what do they know should?

Where do they find you? Where do they? How do they try rappid IT anything else other just going to replay 到 com。

Yeah to go after that com. It's an open beta right now or kind of quickly improving and going to exit beta, I think, in a few weeks. But if if you're comfortable testing that, something that's like not not perfect, go to update a com.

If you submit to our a core plan, you should be able to access the agent and start using IT. Um and we are you know I think the place where we're most active with twitter. So twitter like x uh the handle rapid R A P L I T, or my handle a emo de.

Oh, one other thing, I wanted to make sure we had a chant to touch on, as you can, and something new, something that's coming in the very near future, maybe the day the episode drops, talk about that.

So depending on when when the episode is is coming out, this could be the first, first time people hear hear about IT. But we have this product called agent IT is sort of high agency does everything from setting up the project and all of that, right? And so now we have uh, we are working on assistant.

So assistant is a like i'll say that the cousin of agent IT is a little less powerful but a lot more controllable. So you can like focus on a features or areas of the code that you you want to change. And uh, you you still don't have to know how to code, but there is a lot more manageable and IT is a lot faster.

So you saw how you like I took some time to kind of create the project and code some of things assistance is in the order of milliseconds and seconds to Better respond to you. Um and so again, as you know, as I talk about the idea of tools we want people have as much of power and autonomy as possible. And so there are certain instances where where asian is the best is gna do the debugging for you is going to create the database is for you, but if you want more control assistance is is gone to give you that just .

so folks totally understand what this is going to do for them. What's like at the mental model for what this is like if it's like a person helping you out.

agent a is like, um you're having a developer, uh, where you you give them the P R D, right and they're going to go built the thing um assistance is like you're sitting next to them. So so they built the thing and now you walk over to their to the desk and you say, let me move this spot on three three pixel to the left. Let me you change this thing. So like small increase of changes that we won't happen really quickly and you want to reliably ah that will give you that. So um it's just like much faster iteration on U I and and things like that.

incredible. The future is wild. Final question I was asked everybody, how can listers be useful to you?

Come working, replant. We have A P M. All, I think up with you if your product measure, we're hiding engineers and product measures.

So h so come work a of light of some of us, especially if you like our tools and you want them to to get Better. The best way to to do that is to get us great people. We can. I well.

you're ready to get a flat of product managers applying. amazing. I look, i'm job. Thank you so much for being here. This was credible.

Thank you. Thank you for for your podcast in the community. You built a news letter and everything is awesome to watch.

Thanks, man. Appreciate that by everyone. Thank you so much for listening. If you done this valuable, you can subscribe to the show on apple podcast spotify or your favorite podcast APP.

Also, please consider giving us a rating or leaving our view as that really helps other listeners find the podcast. You can find all past episodes or learn more about the show at lenny's podcast dot com. See you in the next episode.