GitHub's CEO says AI coding is ‘here to stay’
Decoder with Nilay Patel · 7 Aug 2025
Source: https://www.podtrac.com/pts/redirect.mp3/pdst.fm/e/pscrb.fm/rss/p/mgln.ai/e/257/traffic.megaphone.fm/VMP3212224333.mp3?updated=1754527494
Transcript: https://podcasts.opennash.com/shows/decoder-with-nilay-patel/githubs-ceo-says-ai-coding-is-here-to-stay-b71552/ (via OpenNash Podcasts, machine-generated)

As soon as autumn rolls around and you swap your iced lattes for pumpkin spice, that's your signal to start thinking about the holidays, the decor, the gifts, the shopping lists, but more importantly, the holiday treats, cheesecakes, cookies. Is there anything better? And when the recipe calls for cream cheese, there's only one option Philadelphia, the original cream cheese, made with fresh milk and real cream. This holiday season, whether you're going all out or keeping it simple, make it really philly good.

Find your favorite holiday recipe at creamcheese dot com.

Running a business shouldn't feel like surviving a software group project. One app for accounting, another for inventory, another for sales, and somehow none of them talk to each other. That's where Odoo comes in, an all-in-one business management software that brings every part of your business together. From sales and accounting to inventory and marketing, all in one powerful platform. No messy integrations, no bouncing between tabs, and best of all, no spreadsheets. Stop managing software and start managing your business with one unified system.

Try for free today at odoo dot com slash vox. That's o d o o dot com slash vox.

Support for the show comes from CrowdStrike. It's not a huge stretch to say that AI is the next major computing platform. Every platform shift changes the way we build software, and it changes security too. CrowdStrike is defining cybersecurity in the AI era with AI detection and response, or AIDR, a solution for seeing, monitoring, and securing AI across the enterprise. Learn more about how leading companies are turning to CrowdStrike to secure AI and secure their business at CrowdStrike dot com slash decoder.

Welcome to Decoder. This is Alex Heath, your Thursday episode guest host and deputy editor at The Verge. It's AI coding week here at Decoder. You just heard my friend Casey Newton's interview with the CEO behind Cursor, Michael Truell, and now I've got a conversation with GitHub CEO Thomas Dohmke. In many ways, GitHub Copilot set off the current AI coding boom we're all living in. But since Thomas was on the show a year ago, the rise of vibe coding has shifted the buzz to newer platforms like Cursor and Windsurf.

As you'll hear in our conversation, Thomas is thinking a lot about the competition and GitHub's role in the future of software development. It's a good time to be having this conversation. Thomas has a bird's eye view of where the industry is headed. I wanted to know how he thinks the role of a software engineer will keep changing with AI, and when I'll feel comfortable vibe coding myself. This is Decoder, after all, so I couldn't have him on the show without asking what life is like these days running GitHub as its own company within Microsoft.

Okay, GitHub CEO Thomas Dohmke, here we go. Thomas, thanks so much for being on the show. I have a lot to catch up on with you since a year ago when you were last on Decoder with Nilay. First, though, uh, I think since last year, the biggest shift has been this rise of vibe coding that we're talking about nonstop at the Verge and on Decoder. And when I think about vibe coding, as someone who has never engineered any software in my life, is the moment where I'll be able to create an app or a website without needing to understand coding at all.

That, to me, is very exciting, but it doesn't feel like we're quite there yet. Do you have a sense of when the industry will get to that point where someone like me can literally vibe code something from scratch?

It all depends what you're going to build, right? Like, I think we are there for small things. I challenge you to use something like GitHub Spark or Lovable, and you can probably build like a snake game or pong, or like little scoped apps. And I think we're already at that point. And then, as you get more complex, yeah, right, that requires you to have certain systems understanding, you know, architecture, design, databases, and all those kind of things.

I think we're getting closer and closer to a point where you get more of these, as we call them primitives, supported by the vibe coding platform. So if you take something like GitHub Spark, and when we launched it at GitHub Universe last October, it was only creating front-end applications, something that runs in your browser, um, with a language called JavaScript, um, so it could do things that work in a browser, and it had a little bit of backend storage.

Now the version of GitHub Spark actually generates a full-stack application, so it also has, you know, a back-end and a and a database, and it can connect to AI models and those kind of things. And so it got more and more complex over, like, what, nine months or so, and I think we will see exponential improvements to what these vibe coding platforms can generate. The challenge, I think, will remain that it's one thing to create something from scratch.

It's a totally different thing to take an existing software system and modify that and figure out where in the code base which functionality and which test cases. And as most software applications, you know, over the years become bigger, bigger and more complex, as sometimes hundreds, if not thousands, engineers have worked on that. I think we are really far away from AI being able to solve all these use cases that an expert in a system can solve.

That's a good segue to this blog post you recently published where you spoke to a bunch of developers. I'd be curious to know what kind of developers they were, but you interviewed a bunch of developers and kinda took their temperature on the state of AI coding and where they're at with it and what they see coming. What kind of developers did you talk to? These weren't all just people on your team, right? These were external developers.

Yeah, it's external developers, um, uh, folks that were like coding, you know, as when I joined GitHub seven years ago, was the first eye opening moment. Is that when you're at GitHub, given the GitHub brand and then how widely known GitHub is in the industry, it's really easy to find people that are willing to, um, you know, jump on a thirty minute interview and, you know, answer some questions. So it was, um, a range of developers that are certainly, you know, a affiliated with GitHub or have an interest to provide feedback to us.

But otherwise, it was a spectrum of of folks.

what was the top feedback you got? And I'd like to connect that to to the role of the software engineer and how it's changing, and what you basically heard from people about how they think the role of being an engineer is changing right now, and then how it's gonna change over the next year.

If I wanna summarize it really quickly, is AI is here to stay, and I think most developers that we interviewed and most developers in the in the industry, they have realized that the profession of a software developer is going to change. It's already changing, uh, through the use of AI, and as such, you know, the majority of them are in the process of adjusting, uh, to that new way of working. And I had another blog post earlier this year where I described this as the odyssey of developers.

We move from, you know, punch cards and mainframes, uh, to the personal computer, which drastically changed the inner loop of how fast I can try out something. Because before that, I had to book time on a mainframe. And I remember going to my mom's office, she had punch cards on her desk, and then all of a sudden everybody had a PC on their desk. At least, you know, the the nerds had them and they could just start coding.

And then we went from, you know, assembly language and and basic and Pascal and to Python, go and and Ruby, right? We went from no internet to millions of open source libraries that I can just pull into my application from my own servers to the cloud. So we've always moved up the abstraction ladder, and today, a developer that builds, you know, a full stack application, likely does not actually know what CPU is running that application and RAM and and gigahertz and and what have you?

The technical specification of that CPU is. And unless they're a game developer, they're likely not doing assembly either. So we we are moving naturally up that stack because the amount of code that we are managing has grown exponentially over the last fifty years and will keep growing exponentially, and humans are really, really bad on exponential curves. And so we need these tools to, you know, at least for us, flatten the curve, not make it flat, but, like, make it a little bit less steep, so we can actually still handle, uh, these complex applications with, you know, now billions of lines of code.

Half of the developers you spoke to for this post said that they believe that within two years, 90% of all code will be written by AI, which is a thing I've heard from others. Uh, Dario Amadei, the CEO of Anthropic, he recently had an even more aggressive timeline, he said he thinks we'll be there in three to six months. I'm curious, where do you fall on that debate? Where do you see the timelines for a world where pretty much all code in production was at least generated in part or fully by AI?

The first thing that always comes to my mind when we have this conversation is that we need to acknowledge that 90% of our stack today is already written by somebody else, aka millions of open source developers around the world. You know, if you look at the Verge web page a and the back end and all that, I'm sure it's 90% open source, like if you count the lines of code, 90% of those are from open source libraries, um, open source operating system programming languages, dependencies, all that.

And the team, the engineering team, is writing 10%, same same for us at GitHub and and even, you know, the same for our friends at Microsoft, right? So we're already in a world where focusing on the thin layer on top, uh, that defines what makes our business, while we're pulling in the work of, you know, m millions of developers from around the world to accelerate and to innovate and and what have you to secure, you know, uh, the ecosystem.

And the same will happen with AI, that AI will write, you know, nine pieces of code. So I can still focus on my eight hour workday and have my one piece of code. Um, so it doesn't mean that I stop writing code, it just means that now I have ten times as much code and ten times as much functionality and features, uh, as I could produce on my own, right? It's an amplifier in the same way that open source is an amplifier.

On timeline, I think we're going to see a huge spectrum of teams and and companies. And depending on how they adopt AI, not only because they're willing, you know, to use AI, but also because they have designed their company, they have designed their software, their their infrastructure in such a way that AI agents can leverage that infrastructure, right? Like, for example, they have a design system, so whenever you wanna ask the agent to add a feature, it doesn't invent new cascading style sheets and and and what have you?

Instead, it uses a component, so all the features the agent builds actually look, you know, uh, uh, aligned with the with the design language of your product. And so you can imagine there's a whole new way of architecting applications that make it much easier for an agent to use those Lego blocks, you know, to compose new features. And I think that will heavily, uh, influence how fast a company can get to the point that 90% of code is written by AI, and then vice versa.

There's obviously lots of companies still that have COBOL code running on mainframes, uh, and that have, you know, PHP and Pearl and and, and, you know, all the legacy spaghetti code, uh, that sits around with here at GitHub, we have a uh, Ruby on Rails monolith, uh, that is now as old as GitHub is. And uh, it's not that I can just, you know, make the decision as CEO to say, split on the monolith and uh, and we will have, uh, a a nice modern architecture, uh, let alone that, you know, that often is a philosophical discussion, as much as it is an engineering discussion.

But, you know, getting out of this legacy state, you know, going through the cloud transformation, the digital transformation, now the AI transformation, that will take organizations way longer than those that can be on the cutting edge, where ninety or even more percent of the code is written by AI.

So it sounds like you don't have a timeline.

Like today, lots of, uh, folks already out there that are, you know, um, documenting that, uh, on on their blog posts. Like, if you know what you're doing and you start it from scratch with something like Claude Code, you can have it write 100% of the code. And the only thing the author was doing was modifying these in human language just written files to then instruct Claude Code to build their applications. And so that application is, by definition, 100% of the code is written by AI, and all the author is still writing is instruction specification.

And whenever, you know, the source code that was generated was not functional, you, of course, can just go in and modify that. Or you can say, no, no, that I don't wanna do that because then I'm breaking my system. Instead, I'm trying to figure out how do I change my instruction files to then, you know, keep going with cloud code, writing all my code.

Yeah, I'm kind of mad at myself, honestly, for even asking you about this, because this trend that's happening of CEOs saying, you know, random percent of our code is made by AI, actually really frustrates me. Uh, your boss, Satya Nadella, has said actually that twenty to 30% of the of your code at at Microsoft is written by by AI. And you know, Sundar Pichai has put out a stat, Mark Benioff's put out a stat, Zuck has put out a stat.

But I've been wondering, is this actually good code? Is this code that is something you want over time, uh, to increase? Because I was looking at this recent Stack Overflow survey where they interviewed a bunch of developers, I'm sure you've seen it, and this it was really interesting because the vast majority of them were saying that they either used or planned to use AI tools over the next year. It was like over 80%. But I thought something really interesting was that about half of them said they really distrust the accuracy of these AI coding tools, uh, and that their biggest frustration, which was cited by, I think, 66% of the survey, was that the code that it

spits out is just not quite right. Which often leads to them having to spend a lot of time debugging, and that AI coding is actually more time consuming than productivity gaining. And I'm wondering, do you see that across, you know, everything that GitHub does and also inside GitHub, or do you think this is more more isolated? It seems like a big problem that, yes, AI code output is on the rise, but maybe it's not actually good code.

yeah, and now you met that. I didn't give you a percentage,

Well, I was mad that I was actually just asking you for percentages when I'm, like, looking at my notes here and going well, actually, like, people take issue with these percentage questions. Is this even the right way to be talking about this? Because, yes, AI code output is going up across companies and CEOs are using it to to to brag about their efforts, but are we focusing on good code? I'm curious how what you make of that debate.

those two, from my perspective, actually go together, and I think the percentage obviously is a bit of a marketing instrument, and what really does it matter whether it's 90%, 80% or 95%, whatever? I think what matters is that it's clear that the profession of software development is changing, and that, um, we're moving up the stack, and we're moving, you know, from understanding every single line of code to switching back and forth between a specification.

And you know, human language is inherently non-deterministic. You, we can both say the same sentence and and mean different things, let alone that different languages have different constructs of of describing these things. Programming language is an abstraction of the processor, right of the CPU or GPU, or what have you. Uh, as such, as it's, uh, nothing else than describing, you know, the how the transistors ultimately are are flipping, uh, from zero to one, right? And we, we're going to have both because the machines are running on on CPUs, that's not going to change, uh, but we're thinking in in human language.

Okay, so then the question really is, do we wanna write more human language or do we wanna write more code? And that's not, uh, ai determines that for you, that's a decision that you make as a developer, and I think this is going to be the creative freedom for me as a developer to say I wanna write code because I know what I'm doing in this moment, um, or I wanna use an AI agent. And in fact, if you think then about agents, the first question really is is, is it faster to use an agent, or is it faster to do it myself?

And if I know I can do it myself in, like, three seconds, I'm, I'm wasting time, energy, you know, resources. If I, if I ask the agent to do that, it's like, you know, I hope when we're texting, you know, using ChatGPT to write the response to my text message.

I have many reactions like this when I'm using consumer agents, where it's like, yes, you can go off and book me the hotel or whatever, but it takes you an hour and it's not really right. At the end, I should just go do it myself. It's more productive for me to do it myself.

And even, you know, the notification summaries, uh, in mail or outlook or Apple iMessage, are often for me, at least, less helpful than just reading the first three lines of text, right? Then the second question is okay. So now AI is writing all this code, and if AI is writing 90% of the code, as in nine times as much code as I am writing, am I reviewing all that code for, you know, quality, security, coding standards?

Or do I find ways for AI to help me with that as well? Because ultimately, we are going to DDoS, uh, the human developer population with the agent population, as those agents, uh, d sorry, DDoS, uh, the um, distributed, uh, denial of service attack. Um, uh, as these agents don't sleep, they don't take time off over the weekend, and you can run them in parallel. If you look at OpenAI Codex, they are already generating four variants of every task that you're giving, and then you're, and you're effectively helping them with their, uh, uh, reinforcement learning by picking the one that you like most.

But you could not only generate four, you could generate forty or or 400, right? And then you could have another agent that looks at the quality of that code and, uh, rates that quality, and then you have a feedback loop between those agents. So it can actually, I think we're getting to the point that agents will generate always better quality than a human can generate, because you can just run this at infinite scale, and as such, you know, find all the bugs that a human wouldn't find because they don't have the time to do that.

But the challenge will remain is, do we trust that code? And there, you know, eh, we're actually coming back to GitHub. Um, GitHub was built exactly for that reason. It was built for human to human collaboration, right? GitHub is all about me having an open source project, you forking that project, you like something, or you wanna add something, you send me back what we call a pull request. And then it doesn't just go into my code base, but instead I review that code and maybe give you feedback and say, hey, Alex, this is cool, but like, how about you rewrite this?

And sometimes the tone is a little bit harsher on that as well. Uh, and then at some point, you know, we decide, okay, this is cool, let's merge it, right? That is actually quality control of human to human collaboration, and of course, you can use the exact same process for human to agent collaboration. Except if you now have thousands of agents, you have to rethink that approach, and I think that's going to be the biggest differentiator for developer tools.

Those who figure out how we can have agents generate so much more code than humans and have the humans still be in control and and make sure that what goes into production and handles customer data and and billing, uh, uh, uh, processes and and and what have you, that that code is actually functional and and and has quality and and doesn't have any vulnerabilities, right? That is the challenge of of the years to come. Is not how much code is agents writing, is how much code from agents can I actually accept into my production system.

it sounds like you're saying the future of a developer is someone who is just looking at how all these agents are behaving all day long and uh, accepting pull requests or rejecting them, or reviewing code, but not really writing it. Uh, maybe giving very high level instructions, but really stepping back from the actual process of coding. Is that what you're saying is gonna happen broadly?

I think I think to some degree this will happen. We started this with vibe coding, and vibe coding is already happening. Um, all you do is, uh, you know, write a prompt, um, it's actually also happening with image models. If you, you know, use ChatGPT, you create, like, a whatever studio Ghibli image, for example. Uh, you're not writing the code to do that, and you're just giving the instructions, and then there's something in the model that does that for you.

And in fact, you know, one of the hacks is that you can upload an existing image and have it decompose the image into, like, a JSON file, like a data structure, and then you modify that data structure and use that as the prompt to generate an image that has something removed or something modified on that image, right? So you're kind of like acting like a developer, but you're not writing any code, um, you're just like having the model do that.

So I think there is going to be a class of developers that will use models and agents to build and verify systems. And I think there's going to be a class of developers that also still love coding and that I'm mixing up their workday with offloading some work, uh, to AI agents, like writing test cases, documentation, you know, offloading the stuff that they don't wanna do, so they have time for the stuff they actually love doing.

Because the question then becomes, how do you spend your eight hours a day? Um, what is it actually that lets you, you know, explore your creativity and and ship innovation and so on, versus just getting stuff done? And I think it needs to be a mix of both, eh? Most developers today, in small and large companies, code, you know, maybe four hours a day, sometimes even less. The rest of the day, they do all kinds of other things, you know, they are in meetings and they write emails and review others people's code and update servers.

There's lots of other things that developers also have to do. So I think most of them will preserve a certain amount of time during the day where they write code or instructions to generate code, which will feel like coding. In fact, you know, if you think about it, I brought this example earlier, is we move from assembly language to higher programming languages and and and open source libraries, right? Like, I haven't seen many people that are sad that they no longer have to write assembly.

They can, it's not like assembly isn't there anymore, it's just that we have a compiler that takes my programming language and compiles it down into assembly. Now you could see the AI model effectively like a compiler, it takes my human language instruction and compiles it down into programming language, and then the real compiler takes that into assembly. Your code actually would be the human language, right? Like, you wouldn't ever have to look at the programming language.

Except that all the models today still have hallucinations, and they still write code that doesn't exactly do what I wanted. Or maybe, you know, I do it ask it ten times, and each of the ten times, it writes different version of that code. Um, and you can actually see that when you watch people vibe coding live on stage. The snake game, uh, that I have, you know, used as a demo, looks different every single time. And so we don't trust the human language conversion into programming language enough to actually cut out that layer, but we were able to cut out the assembly layer and we're good about this.

And it's, I think, as long as we have those two layers, we are always going to also write code, uh, and and learn coding, and and go back and forth between the deterministic and the non-deterministic layer.

Because LLMs are inherently non-deterministic and and do make things up, it's actually a function, not a bug, it's it's how they're architected, they're designed to hallucinate. Is there ever gonna be a point where I can get to what I was starting this conversation with, where someone like me, who has no understanding of of coding, who can't tell you the difference between languages, can reliably trust that? Something I'm building that requires access to different APIs, a database, web, all that can be built through natural language.

Cause if we ever get there, that seems big. That seems like it will have profound impacts on the world. But like you said, maybe it's just inherently not in this technology to be able to get there.

We'll get there as long as the scope is, you know, limited to a certain degree, and that scope will increase. I think I think you're going to be able to do certain things, um, where you're, and if you will, um, if you talk today to an AI agent and let it do, um, you know, render a chart for you. What that actually does is writes a Python script and then use that as a Python script to to render that chart.

And you can, you know, ChatGPT's go and say, show me that Python script. And I think most users never need to understand that Python code to to look at the chart, right? So, and those use cases, I think, will increase at at a rapid pace. But then also what we consider as professional software systems, the products that professional software developers write, will become so much more complex, um, right? Because it is now then easier for a professional software developer to build more functionality in the same amount of time where in the past they would have built that Python script to render that chart.

If that makes sense, what we consider now today kind of like, you know, state-of-the-art software will s be so much more complex than, uh, uh, what we can imagine a human developer can write. And in fact, you know, that has always been true, also with what my my iPhone can do today is unbelievably more powerful than what a Commodore sixty four could do in the early nineteen nineties, right? And that developer that was building for the Commodore sixty four would look at the way of how iPhone apps are written and and and think about that as magic, in the same way that we we see AI as magic.

And I think if you think about it that way, we will get to the point, but we will also get at the same time. Have the professional developer feel still like a magician because they can write, uh, uh, such complex applications.

We'll be right back after this short break. We're back. Let's talk about competition. Uh, you've actually already mentioned several, uh, competitors to GitHub Copilot in this conversation, so that tells me you're open to talking about it. Uh, coding, I think, is safe to say, the fiercest competitive corner of the AI world right now, at least that's my read of it. You've got a lot of deals happening, startups being sold, the some of the fastest growing companies of all time, in fact, are AI coding companies.

Cursor, Lovable, the vibe coding app we've talked about already in this conversation, GitHub Copilot was really the first to this space in terms of being an an AI coding assistant, and still huge, obviously. But I'm curious, like, especially in the year that's lapsed since you've been on Decoder last, I'm wondering if you agree with the sentiment that GitHub Copilot seems to have lost mind share, at least in the Silicon Valley niche, uh, that that is obsessed with this stuff and and talking about it every day.

I'm sure you still have a ton of users, but I'm curious if you feel that, and if you feel like that's a pressure on you to evolve the product faster. Uh, and yeah, how you look at how competition has changed, uh, this year.

I think the overall market, um, of artificial intelligence, large language models, you know, what is possible with these models, is moving at a really rapid pace. I've never seen anything like that, uh, in in tech, and I've been, you know, around for a while. And you're right, AI code generation is at the forefront of of that innovation, you know, logically, if you think about it, because all these companies are also employing developers, and so there's an in inherent, uh, motivation for everybody to make coding easier, because that means I can innovate faster.

And if I have, you know, uh, a coding tool that lets my developers move 5% faster than my competitor, I I will use that. Since I was on, uh, last time at Decoder, that market has accelerated, and, um, I want to be really open to this. This is amazing. Like, I've been a developer for so long, and I've never seen so much innovation in software development tools. And as GitHub, we've always understood, uh, us as a part of an ecosystem that is both competing and partnering with, you know, companies in our space.

Um, whether it's in AI code generation, or whether it's in, uh, CI CD, like how applications are built, um, we are not making decisions of what, uh, open source, uh, uh, library you prefer, right? Like, you wouldn't, most people wouldn't use GitHub if it would only offer you the JavaScript ecosystem, but would tell you to go somewhere else to use Python. And, um, nobody wants to watch sports if there's only one team, uh, winning it all.

Um, where I don't agree with you is that, uh, uh, we lost mindshare. I think we, you know, have, um, won some and lost others, and, uh, you know, sometimes that's on a week by week basis that you see folks, uh, posting on on their X that, uh, now they're back on on versus code, um, because the latest release and and we're doing nightly builds and and monthly releases, ha, has something they want. That's that's something we should, as an industry, be really proud of.

Uh, uh, Copilot was late on moving from a single model to multi-model choice, but we got there, uh, uh, last year in October. Uh, you know, we added agentic support, um, we p built an MCP a model context protocol integration, uh, and now all these, uh, uh, coding agents are using the GitHub MCP server, um, that we have built, um, and and partnership with Atlassian. So it's a race, uh, it's a race where, you know, we just announced in Microsoft earnings that we now have 20 million, uh, users on on GitHub Copilot, um, 75% quarter-over-quarter growth in in enterprise usage, and, you know, 90% of the Fortune one hundred.

I can give you all the stats, um,

hold on on the user number, I gotta ask you about this user number. Tom is, there's something funny going on with this user number. This is lifetime users, this is users that have ever used it, or is this monthly, is this daily? Can you give me a little more granular detail here?

it's 20 million users that have, um, g 20 million GitHub users that have Copilot enabled, Enabled so it it matches the 150 million accounts number, um, same, uh, same way of measuring it. And we have 50 million users on, uh, fifty s five zero million users on Visual Studio, Visual Studio Code. So that's gives you an idea of, you know, how far we are in, um, in activating, you know, the the users on our IDs.

this post of yours where you announced that, uh, that new user number stat, you wrote. The true measure of a company is never drawn during its hype wave, but by its resilience when pressure tested. Uh, and then you wrote something even more interesting, I thought, where you said, even with all the constraints we faced, we've proved that a little grit wins the game. What constraints were you referring to there in that post? Cause you're you're the CEO of GitHub, I would think that you guys are not really constrained.

You're also model agnostic, whereas the rest of Microsoft seems much more still connected to OpenAI, at least for now. But it seems like you're able to work with whatever model you want. You're still as a a separate part of Microsoft relative to the other parts. So what constraints were you talking about?

Any size of company is constrained, right? That's the definition of, uh, of budgets and, uh, uh, uh, employees or headcount, uh, what have you? Um, and you can just always add more people, and then you run into the mythical person months that uh, uh, adding more people to a team actually slows you down and doesn't accelerate you. As such, you know, uh, whether you're a ten-person startup or a three-thousand-person GitHub, or, you know, 200,000 or so person Microsoft, uh, you have to make decisions.

And Apple famously said, you know, uh, uh, a thousand no's for every yes. And so one constraint is that, of course, uh, we have endless, you know, uh, number of items in our backlog, um, and uh, many of the backlog items that that GitHub has stem from a time before AI. Um, uh, you know, customer feedback, people wanting a feature in GitHub issues or GitHub projects, um, uh, y you name it, um, uh, in fact, you know, there is a running joke that for every customer feedback item we get, there's already one there from months or months or years ago, where somebody else had that same idea, right?

Like, our naturally, because our customer base are developers and we are developers, uh, there's there's always going to be, uh, somebody having an idea that was already there, or or telling us how to do something better. And then, uh, you know, you have, uh, a a group on on X that says, you know, well, that was a stupid decision from GitHub to make it default, and and now they're saying this is the best decision ever. And so we are living in that world where we have way too many input signals and we got the the constraints.

The grit is to pick the right ones where you're really f moving fast and where nobody in industry is actually knowing what AI code generation or AI in general looks like in in two or three years. I think that if we're all honest to ourselves, that is true. We're like, if ChatGPT was the Netscape moment of the twenty twenties, then we are like in 1995 now, where we have an Amazon and a Google. But we have not seen a Facebook, not seen a Shopify, and not seen the iPhone, right?

Like in 1995, it would have been preposterous to predict that those things were happening, and you actually have the confidence that that is true. And I think that's the world that GitHub has been living in, uh, ever since its founding is, um, it was ahead of its time, leveraging Git, I mean, to to build what nobody called DevOps back then, uh, but what is ultimately now one of the the largest developer platform.

Um, and then it was swinging back and forth of what is the next big innovation and what do others in the in the industry are doing and where do they actually prefer to partner instead of building it up themselves? And I think with Copilot, we have navigated that well, and we had our moments where this wasn't, you know, working, and and strategic decision that ultimately ISCO made wasn't the one that pulled us in the right direction, um, to win in this space.

Um, and now I'd say, uh, in August, um, 2025, we are still ahead of the curve, and and we are still the leader of the market, and as such, we're really proud and and happy about, you know, the the fruits of the hard work that we're earning.

Speaking of partnering or not, you know, coding has become a thing that affects huge transactions that people read about as it relates to AI. And I'm thinking about Google and Windsurf. Windsurf was this, you know, still is very ascendant AI coding tool, and they were gonna sell to OpenAI. That deal got caught off, and that's where, you know, Microsoft comes in here, and I know you're GitHub, but it's all kind of related. And I'm curious, you know, a big thing with that and why Google ultimately ended up hiring that team from Windsurf, is that the company wasn't able to sell to OpenAI.

Because there was a concern that the IP from Windsurf was not going to be shielded from Microsoft, that there were competitive dynamics at play given the Microsoft OpenAI connection. So I'm wondering, why was that such an important issue that it actually ended up killing a huge deal? That would have been a huge deal for not just the AI coding space, but Silicon Valley in general. This would have been a multi-billion dollar sale to OpenAI, and I'm curious why.

From where you said, I know you, you probably weren't driving that, but w why? Why? Why was that such an important issue?

You have to ask the question, and I, uh, can refuse to answer the question. Um, no, I look, you know, uh, obviously I have a conflict of interest, um, uh of even involved in in in these conversations as such, I I don't have any background knowledge I can I can share with you on this. What that actually shows you, though, is something else, which is the fear that developers are soon going to be replaced by AI, is actually not justified.

In fact, the opposite is true, and teams, um, like the founders and and and core team behind Windsurf, or, you know, other other parts of the industry, you know, meta meta comes to mind, uh, uh, get offers that match professional basketball players. Some of them even have agents helping them to land a new contract, and that should be something that we should be excited about. Uh, you know, how we get to that point is is a good question to explore for somebody who has access to all the documents.

I don't, but I think, you know, as a developer, uh, that have been in this in this industry, building developer tools and using them myself, that's, you know, freaking exciting. Like, because it shows you that the market value of those at the top of their game keeps growing, and it hopefully also motivates, you know, kids, teenagers, those that are often first into gaming and then wanna build their own stuff, uh, still get into a profession. Because I do believe that we will need more developers than ever if this all plays out the way, um, uh, uh, uh, I hope it does.

let's turn to Cursor specifically, which I think has really taken the mantle, as at least in the Silicon Valley, really early adopter set, like the hottest AI coding tool right now. What have they cracked? What was it that they saw, the insight they had that you think got them such a quick lead in terms of just mindshare awareness? I know GitHub Copilot is still bigger, but they're growing very fast. And, uh, you know, the CEO is actually on Decoder recently with Casey Newton, and he he cited you all as as inspiration for, you know, them starting the company.

But w what have, what did they crack that you didn't see, and and how are you responding?

What Cursor cracked was to realize that it's not just about adding AI into the IDE, but about changing the IDE itself and designing what you might call AI native workflows. Um, but it's really like thinking about how will developers work if AI is the default and and not an add-on? They were the first, you know, with Claude Sonnet 3.5, so they were also, you know, ahead of the curve of saying, okay, multiple models, um, play a role, and let's give developers the choice and and pick the model that that works best for them, instead of us making the choice, which was our philosophy, um, you know, a year ago when I talked with Nilay, um, that's two different

philosophies, right? Like, we ran lots of eval suites, we looked at these models, and then we picked the one that we believe was best. But that's not a binary choice because each programming language and each test case and, uh, you're generating f functional code or testing code and so on, has a different score. And so then you have different scores for different rows, and then you make a judgment call. Okay, this is overall a better model, right?

Well, but we know from the history of developer tools, in fact, we at GitHub know this very well, is that the best thing you can always offer to a developer is choice and have them pick the model that they think is right. And then they will know, you know, from the experience, or you may call it craft, which to pick, and, um, and I think that has, you know, changed the market. Um, today, you cannot compete in AI coding if you're not offering multiple models, if you're not having the best models, quote unquote, the best perceived by the majority of developers.

If you don't o offer them, you know, um, to bring their own models. If you don't have an agent mode that runs within the IDE, if you can't, you know, take that agent mode and offload it into the cloud. And what we call the Copilot coding agent, um, that is actually integrated into into GitHub platform. And so, you know, I'm a big Formula One fan, so I like to always think about the world like that. Um, you know, there's teams that win whole seasons and then the next season, uh, they're behind, and they're not taking that as in, well, shit, we lost, and, uh, we'll never win again.

But they're taking that as an inspiration to say, okay, we need to rethink of how we're doing these things. And I think that's what happened, uh, to Copilot, you know, in in mid-late 2024. And, you know, I think now, where we are in August 2025, we can win races again. Not every race, uh, and neither does the competition win every race, but we certainly have, um, innovated in AI code generation ourselves, while also, you know, caught up, um, to what has made others faster.

This space does seem so dynamic because of that point you brought up, which is that it is so dependent on the models. Whatever tool is currently leading, the models they're using, their way that they price for those models, and the models themselves, who is leading changes a lot. It's a almost impossible, even being in it, to keep up with all of these model releases. You know, Llama three was a good model, you know, Llama four not really as much like Anthropic right now seems to be the king of coding.

Maybe that changes, maybe it becomes OpenAI in a few months. How does that change the app layer, right? So it's I I appreciate how, you know, quickly the space is changing. So I think to say that you all are for sure lost, or Cursor is for sure won, is way too early. But as I think about Anthropic, um, and you've done a lot of work with them, you mentioned the MCP server, and and obviously you can access Claude through GitHub Copilot.

A thing I've been asking people is, why is Anthropic so good at coding? What is this secret sauce they have? You know, Dario was on, uh, the Big Technology podcast recently, and he he did not want to talk about why they're so good at coding. Um, but it's the billion, I mean, god, it's maybe the 1 trillion dollar question is is, why is it such a good model at that? Uh, I'd be curious, as someone who who works with them closely, why?

Why do you think they've gotten such a lead in coding specifically?

Let me, you know, for a second, go back to the first part of your question, and I come to anthropic why anthropic is good, um, in a minute. I think a in tech, we have this notion that for somebody to win, somebody else has to lose. Um, and that has not been true, you know, for Windows and Mac OS, that have not been true for iPhone and Android, and for many other technologies. In fact, in developer tools, that never was has been true, because if if so, we wouldn't have, you know, the dozens and dozens of programming languages today, and everybody would have moved to what is perceived as the best one.

And if you look at, you know, the competitors of of GitHub Copilot, like Cursor, Lovable, Windsurf, Old Vercel, etc. Well, you know, where do they those people that use these tools actually store their source code on GitHub? You know, where do they manage, you know, their issues and projects on GitHub, where they do they run their CI CD on GitHub? You know, where many of those competitors actually run their model inference on Azure, right? And so as such, you know, we are part of an ecosystem as Microsoft.

We have always been a platform company. GitHub has always been a platform company, and so we can be both be competing and we can benefiting, uh, from these companies driving the size of the overall software ecosystem. Because let's face it, software development has only ever grown. It hasn't really detracted, it is not getting smaller, the number of developers are getting bigger, the value generated from look at the tech companies, uh, you know, on on the stock exchange, um, on any given day, they're often outperforming the market.

If our competitors are winning, we are winning too. And I think that's the mindset that is really important to keep in mind when we talk about these these battles. Uh, I, in in this,

That's the benefit of being a hyperscaler, Thomas, is that when the competitors are winning, you're winning too. Not everyone can say that.

sure, that's called, you know, differentiation, um, and then on Anthropic, you know, specifically, Claude Sonnet, where Claude Sonnet, you know, if you believe YouTubers and bloggers and and experts like Simon Willison, Sfix and others, where really it outcompetes other models, is tool use, like the ability of the model to pick the right tool for the next step of the agent mode, and so, and that shows you it's no longer just about, you know, the training in in programming languages, code and whatnot, it's also training in what are the tools that the agent should use in a similar fashion as a developer should use.

And if that fails in a step, and you see that with some other models where the tool usage is just not as good as Claude Sonnet, then your agent mode just falls apart, right? Because it can go through the chain of thought and actually, when, like, imagine it needs to install, uh, an NPM package, like a JavaScript dependency, and that doesn't work, then the agent can have the best training set in the world. If it cannot achieve that, if one thing, the whole thing, uh, uh, no longer works.

And I think that's where, uh, Anthropic had an early insight, or had better, better eval suites, better testing, um, that that let them to have this head start, and and others are trying to catch up. But, you know, the funny thing is, we are recording this, uh, uh, uh, and then, you know, by the time it actually gets published, um, maybe that world has already changed again. Um, uh, that's that's the fun part of that industry,

I think you're maybe alluding to GPT five, which, uh, may be coming when this episode comes out. you Audience's hands are in the air. We need to take another short break. Welcome back. Right before the break, Thomas and I were talking about GitHub's competition. That gave me the perfect window to ask about how things are going with Microsoft and OpenAI. We've touched on pretty much all the main competitors. I do want to end just with I I know you can't say much, cause this is all being discussed, but Microsoft is currently talking with OpenAI about what the next kind of chapter of the corporate relationship looks like.

Because OpenAI is moving from a nonprofit to a for-profit, and Microsoft, you know, is obviously a huge shareholder and and currently has exclusive access to the IP. Um, I'm curious, you know, as a CEO of GitHub, how would you like to see the OpenAI Microsoft relationship continue? Like, what do you think is important for where you sit for that relationship, and what would you like to see?

for GitHub and for all developers, it is always going to be important that it's a healthy relationship, one where, uh, both sides are partnering with each other. You know, that's how the original Copilot came to be, uh, through a partnership between OpenAI, um, who had developed GPT three, and then the Codex model, Microsoft, who had not only the hyperscaler infrastructure, but also years of experience in responsible AI, scaling these processes. Because, look, you know, if you think about the original Copilot, it was a masterpiece on its own, because on every single keystroke, we're running inference against a large language model.

That was unheard of at that point in time, and at that scale, you know, of of of millions of developers in in a very short amount of time. And so that, you know, innovation between GitHub, Microsoft and OpenAI created this market, uh, that we we are now in, and that has all these competitors and and probably another, you know, dozen, uh, to come in the next month. And I think that partnership going forward will create, you know, new such innovations through new models and through new ways of running these models more efficiently.

I mean, a lot has been written about, uh, energy consumption of AI and uh, cost and gross margins and and those kind of things. And and through GitHub, which always has put developer first, that's what we are, you know, breathing. It has always been about human to human collaboration, and we are moving into human to agent collaboration, and more likely into agent to agent collaboration, and all these three layers will stack on top of each other.

And, uh, OpenAI has, you know, the largest, uh, uh, with ChatGPT, the largest AI platform today, and so partnering with them, making sure the content that, uh, developers generate, uh, in ChatGPT with the OpenAI codex complies with the same we call this the control system to the same security compliance, you know, um, enterprise company standards is going to be crucial for developers, you know, to adopt these tools and integrate them into their workflows. And I think the ultimate battle in AI is going to be fought at the platform layer, because if all agents are insanely powerful, um, then what you really care about is how can you integrate those into your other business processes.

And for us, obviously, that's, um, uh, the developer lifecycle.

Makes sense. Well, let's end on some GitHub Microsoft specific questions. Um, it's been a year since your last, uh, annual revenue run rate stat, I believe it was 2 billion. Uh, why no update? What's what's the current number?

Investor relations team has made the decision what's being disclosed in earnings and what's not. So, as per usual, I can only tell you what's in the earnings script, and, uh, this time we made the decision to not publish the number.

Okay, I'm assuming it's higher than that. I'm assuming it's higher than 2 billion.

We published other, we have published other impressive numbers, and we're really happy about the q four results, uh, of of over Microsoft.

Nilay, I tried this one last year, and I'm expecting to get a similar response. But, um, a big question that we have is is, does Copilot make money for GitHub? Does that even matter? Is that even a consideration? Or because you're running effectively your own P and L to a degree, as a as a separate org within Microsoft, like, is this actually making money, or is it a is it a cost center right now to to gain market share?

you're right, GitHub is, uh, has its own P and L, um, and as such, we are measured against the the goals of that P and L. You know, I think in general, if you look at Microsoft as a business, there's always going to be the desire for, uh, these units to to make money. The timeframe is different, um, and often, you know, the bets are really big. Uh, and as such, I think the better question is, is it going to make money, uh, over over the lifetime of of of the customer contract, so-called LTV?

And on that, I'm very confident that, uh, Microsoft shareholders and and Microsoft's, um, uh, uh, leadership team is going to be happy about GitHub's business results. Look, you know, we bought GitHub in 2018, so now, you know, a little bit over seven years ago, for $7.5 billion at a time, uh, when the last published revenue number was 200 million, uh, year before that, and and you mentioned last year we published te 2 billion, so the revenue has ten times increased, uh, in in in seven years, and you, you know, it's safe to assume that, you know, that number keeps growing.

Uh, you, you mentioned, uh, uh, the the business numbers from Cursor, uh, given that, you know, we are very open about that, we believe Copilot is, uh, bigger than Cursor. Uh, you can, you know, forecast what our number might be. And so, as such, you know, we we are really, really happy as Microsoft about this acquisition. I I d I personally put it into the top three of deals that Microsoft has ever made. I think, you know, uh, Satya Nadella and Amy Hood would agree with me on that one.

And we are really happy about that. We have grown the plat platform in a sustainable way because we haven't left open source behind. You know, I may say we haven't fucked it up. Look, GitHub is still GitHub. GitHub is still like the most beloved brand in the developer ecosystem. It is adopted by almost any enterprise company in the world. Not all of them have it for all their developers, but most of them have it for some of their developers.

And it has, in that age of AI, reinvented itself. We have reinvented itself. Because just as we u to used Git to build GitHub, we used large language models, GPT three, together with OpenAI, to build the first ever Copilot and to lead that that market.

yeah, that's a good response. I'm sorry, there was a lot for me to unpack there. Um, so I is there a is there a big goal where you all see this all flipping and this becomes super profitable? If it's not right now, like the the Copilot stuff, is it is there? Is it the market gets to a certain size, or there's bundling that happens with other Microsoft products, or what is the if this, then that that you're seeing that makes this all worth it?

We have a version of, you know, Moore's law, um, maybe it's Jensen's law now, or whatever, uh, where the GPUs get faster and cheaper. Um, there's like, you know, you mentioned the valley a bunch of times. There's like, what, a dozen startups that are trying to invent, uh, uh, uh, processing units for transformers, um, that do that more efficiently. But I believe, uh, I in in scale, uh, you could have asked the same question to Jeff Bezos, like, what fifteen years ago, and so and look at where, uh, the hyperscalers are today in terms of profitability.

And so I think we will have the necessary patience, we will drive efficiencies. Um, caching prompt caching plays already a huge role, uh, for agents, um, we will see an evolution of business models, for sure, and we will see, you know, our own developers being more productive with AI. And as such, you know, that's not only, uh, that you're optimizing, you know, gross margins, whatever those may be, but you're also optimizing, you know, the way software is built in in our own company.

I actually believe strongly believe that those companies that are still doubting the use of AI, they're going to put themselves in a comp, uh, position that ultimately lets them fall behind in innovation. Whoever is not using AI should have massive FOMO, fear of missed opportunity, that everybody else in their in their in their space, in their industry, uh, is already on AI and have their developers 10% more productive, 20%. And look, you know, the these numbers are always like, uh, why is it only 20%?

couldn't it be 30%? 20% is a massive productivity gain compared to, you know, what a software developer costs, how long it costs them to go through, uh, uh, uh, you know, higher education to ramp up in a company, what have you, to move from one project to another, right? And as such, whoever is not using internally AI to make their own processes more efficient, not only for developers, for everything else. Support is another great example, sales, you know, lead generation, all these things are going to ultimately fall behind.

And whether they have positive gross margins or not won't matter to them, what will matter is that they're no longer being able to compete in the in the market.

That brings me to a memo one of your colleagues recently wrote internally at Microsoft, where it was instructing managers to evaluate employee performance based on the use of internal AI tools, um, saying it's quote no longer optional, it's core to every role and every level. I'm sure you can appreciate the angst that people feel about this shift, um, people who are farther along in their careers and feel like they're set in their ways, or even younger people who are just coming in and are overwhelmed by all the change.

I feel everyone's overwhelmed. And if you're a mid-level engineer at Microsoft, you're probably wondering, uh, am I going to get left behind in this shift? And is the company actually going to be able to even evaluate this correctly? Do you feel, as the CEO of of GitHub, do you feel that the management team at Microsoft has a actual measurable way to determine productivity gains internally with AI? And if that's actually going to be something that employees can can know, uh, in a in a real way.

The memo was a bit more nuanced, uh, the memo talked about AI learning, AI usage in what we call the connect, which is a conversation between the employee and the manager. And at GitHub, we have a similar process, um, and it's about the employee, you know, writing up what they have achieved in the last year, you know, what could they have done better to have more impact, what are they going to do in the next year, and how they're growing.

That's kinda like the framework, and I think in 2025, it's totally fair game to say you should reflect on your AI usage, and you should reflect, what did you learn about AI? Did you use, um, uh, GitHub Copilot or Microsoft Copilot, Teams Copilot to summarize a meeting, and if not, why not? And then the manager, you know, provides their feedback on that and and what they have learned, which often also is a good learning experience for the manager to say, oh, the employee is actually ahead of me in in AI usage.

Like, I certainly have had those moments where somebody showed me how they did something, I'm like, oh, um, you know, didn't realize that that actually works. And I think that process is key to Microsoft's culture, and, you know, the term that was coined around this is growth mindset, having the mindset of saying, okay, you know, I did something, and there's a way of doing this better. And I wasn't born, you know, with all the capabilities I have, I was born with the capability to improve myself.

And I think in that, if you see that memo in that context, it actually perfectly aligns uh with Microsoft's philosophy the other thing I would say you know is like specifically to GitHub is there is no GitHub employee that cannot use GitHub no matter what function you have in the company so it's not only developers and product managers also HR and finance and legal and all g and a all the GNA functions all the technical functions all the sales functions they're all using GitHub there is no world where I would allow for somebody to say well sorry I I don't wanna use GitHub um and I think you know that's fair game if the employee doesn't want that and then

there's tens of thousands of other tech companies out there uh where they can have that but it's part of our company culture that everybody at GitHub uses GitHub and it should be part of our company culture that everybody uses Copilot uh and AI that doesn't mean that we're looking at how many lines of code you have written uh uh today with AI in the same way that we never looked at an employee and evaluated them based on how many pull requests have they filed or how many Git comments have they done cause these metrics are easily you know uh gamified right but it shows you know a mindset that aligns with our culture if you're using our tools to build

our tools

Okay that's that's a good answer I I wanna end on it's kind of a heavy topic uh you can take it whatever direction you want but this debate about AGI super intelligence how we get there I think I've heard Zuckerberg has mentioned this others have mentioned this is that AI coding is actually the path because it gets you to self supervised AIS building and maintaining other AIS and that the company that wins in AI coding will probably get to super intelligence whatever you wanna call it if you have a su definition of super intelligence I'd actually love to hear it uh but that AI coding will actually be the way we get there I'm curious if you agree with that

or if you don't and if you actually have a definition of super intelligence please share Thomas

I don't have a definition of super intelligence myself because I think ultimately those definitions uh uh don't matter much um other than either they are encoded in in contracts or or like a nice marketing instrument then you can go on stage and say we are the first to have super intelligence and then next week somebody comes and says well no no no we are the first that have actual super intelligence right these terms are like um s uh stretch goals uh and they're constantly changing I think you know the the crucial moment where we you may say call it a AGI or ASI is that the AI can improve itself and uh and make itself better the the

model you're basically jumping from GPT four four to five without you know humans in the loop I think when we get to that stage where all of a sudden AI behaves like a kiddo where your four year old moves from laughing about a joke to telling a joke I think that's when we will talk about AGI that's the moment when we will have say it has happened and there is now a a thing that is able to improve itself and keep going that's what we as humans really consider intelligence it's not what you know like just because you've been in Trivial Pursuit uh it doesn't mean you're like the smartest person i in the world just means that you have

you know good mapping between the questions and the answers but you know the the ability to constantly evolve and improve yourself I would I think that's that's the definition that we should use

When it acts like a kiddo I like that I haven't heard that one yet Thomas thanks so much for your time

Thank you Alex really enjoyed the conversation

Thanks again to Thomas for joining the show and thank you for tuning in if you'd like to let us know what you thought about this episode or what else you'd like us to cover drop us a line you can email us at decoder at the verge dot com we also have a TikTok and an Instagram check those out at decoder pod if you like decoder please share it with your friends and subscribe wherever you get your podcasts and if you haven't already don't forget to subscribe to The Verge which gets you access to all of our stories and newsletters including the one I author called Command Line Decoder is a production of The Verge and is part of the Vox

Media Podcast Network our producers are Kate Cox and Nick Statt this episode was edited by Xander Adams the decoder music is by Breakmaster Cylinder see you next time