"The real work in Linux isn't being done by me," Linus reveals his recent situation: He hasn't been a programmer for nearly 20 years, hasn't touched AI programming, and all his stress comes from "people".

"For nearly the past 20 years, I haven't really been a programmer."

"As for Git, which I invented, I'm now just an onlooker."

"I used to say that my job was to reject (proposals), but now I have to say 'yes' to new things, even against the objections of some long - term maintainers."

"Vibe Coding enables people to do things that were previously impossible, but from the perspective of maintainers, maintaining the code it generates 'could be terrible'."

These words are neither a joke nor self - deprecation. They are the sober confessions of Linus Torvalds, the father of Linux and the inventor of Git, when facing the technological wave.

Earlier this month, Linus Torvalds had a conversation with Dirk Hohndel, the head of open - source at Verizon, at the Linux Foundation Open Source Summit held in Seoul, South Korea. He talked about the transformation of his role, how AI is reshaping software development, his thoughts when more and more hardware relies on Nvidia's proprietary GPUs and CUDA rather than open - source Linux, the conflicts caused by Rust in the kernel team, the real - world dilemma that kernel.org is severely disrupted by various AI crawler tools, as well as the daily stress he faces and how he relieves it.

In the current situation where the AI boom has almost rewritten the fate of developers, Torvalds admitted that he hasn't used AI to assist in writing code, and he hasn't even tried it. "But I'm sure someone is researching whether it's suitable for the kernel codebase." When asked if AI would make programmers' jobs disappear, he simply said, "AI is just another tool. Just as compilers freed people from writing assembly code and greatly improved productivity, but didn't make programmers disappear."

Of course, if someone disagrees with his views, they can send him an email. But he said, "I can almost guarantee that I'll read it, but I can also almost guarantee that I won't reply." He joked, "I rarely reply to emails. If you don't receive an email from me, it means I'm quite satisfied. I just don't let people know. I apologize for that."

The following is the full content of this conversation:

"I'm no longer a programmer. I'm not doing many things. I'm mostly watching Linux move forward."

Dirk Hohndel: My name is Dirk Hohndel, and I'm in charge of open - source work at Verizon. I've been involved since the founding of the Linux Foundation, and I've been in contact with Linux for almost as long as the person on stage here - because you are..."

Linus Torvalds: Yes, I'm Linus. We're doing this interview because I hate giving public speeches. In contrast, I have no idea what Dirk will ask me, but it's much more relaxing this way. Over the years, we've had this kind of conversation one or two times a year. This format isn't new. It makes me, who usually doesn't like the identity of a 'public figure', feel more comfortable than a traditional speech.

Dirk Hohndel: This is actually our 28th such conversation. It's quite interesting to think about it. The last time we were here was exactly ten years ago. I'm glad to be back in Seoul. For me, every time I come to Asia, it's very interesting. People here view open - source and software development in a completely different way. It's a whole different world, and I'm really fascinated by it.

Linus, ten years ago, you just released Linux 4.8. Can you briefly summarize the biggest changes over these years?

Linus Torvalds: A lot of work has been done in the past. But I want to emphasize something I often repeat because it's very important - I'm not the one doing the real work. For almost the past twenty years, I haven't really been a programmer. I'm more of a technical leader and maintainer of the system.

This is true for Linux, and even more so for Git. Now I'm almost just an onlooker.

I want to remind everyone that the real contributions are made by others, maybe by the people sitting in the audience. Many people attribute all the credit to me because I've been in the Linux project all the time. Actually, now I'm mostly 'watching' this kernel project move forward.

Dirk Hohndel: Looking back over the past ten years, what impressed you most during the evolution and development of Linux?

Linus Torvalds: What impressed me most is that - I used to say that one day this project would be 'finished'. But that was a long, long time ago. I've been working on Linux for almost 35 years, and I don't feel that there will ever be a point where I can say 'Okay, that's it'.

In fact, I've gradually realized that for all truly long - standing projects, the core work is actually maintenance and continuous support. Especially for the kernel, Greg (Greg Kroah - Hartman, a Linux kernel developer) and I discussed it yesterday. As long as new hardware keeps emerging, there will always be new work on the kernel side. But even without considering new hardware, I'm a bit surprised that after 35 years of the project, we're still modifying the core kernel code to make it cleaner, more maintainable, and more stable.

At three o'clock this morning, due to the time difference, I was still discussing with someone how to clean up some code.

For a system like Linux, the real work is continuous maintenance, keeping everything running smoothly, and dealing with new challenges - whether from hardware or from the ever - changing software ecosystem.

Dirk Hohndel: From a process perspective, the development model of the Linux kernel has been very stable in the past 15 years. But it's 'too boring' for the media. People often only focus on the moments when you raise your voice or when you reject a proposal. In your opinion, has the situation gotten better, worse, or remained the same? How often do you feel you have to stand up and say 'We're not doing this' now?

Linus Torvalds: One obvious change is that I used to say that my main job was to say 'no'. People came up with all kinds of radical new ideas, which might be interesting, but they clearly seemed like a maintenance nightmare. So I would say, 'No. Put it in your own sandbox, make it work, and prove me wrong with data, then come to me.' I think this was a large part of my job as a system maintainer.

But in the past few years, I've found that sometimes my job is actually to say 'yes'. Because... as you know, having been in this circle for so long, and with hundreds of maintainers having been here for decades, people have become a bit set in their ways. Sometimes you want to break the deadlock and say, 'Hey, we need to try this new thing,' and I'm the one who says 'Okay, let's do it'.

Take the adoption of Rust as an example. Although we've been working on Rust for five years now, it's not entirely new. But at the time, I thought that the kernel shouldn't stand still. We needed to try something new and attract new people.

This is one of the biggest changes for me: Now I need to encourage other maintainers to be more open to new ideas.

"Rust has become part of the Linux kernel, and it took longer than I expected."

Dirk Hohndel: Rust is exactly one of the examples I want to mention. I noticed that although Rust has been around for five years, it has only really entered the kernel code about three years ago. It has indeed sparked a lot of discussions and controversies.

Some people expressed their frustration, and there were also arguments about code formatting or inconsistent code review opinions for an unfamiliar language. Some maintainers even quit because of it. Do you think it's all worth it? Is it really worth disrupting our development process by introducing new technologies?

Linus Torvalds: I think it's worth it. But I also think that Rust has indeed attracted a lot of media attention, probably because it's quite prominent in the kernel. Of course, there is also obvious Rust code in other places. In fact, we have disagreements in almost every area of the kernel because this is part of new development and bug discovery. People can get very passionate when defending their views, but in this sense, Rust isn't fundamentally different from other areas. It just makes it into the news more easily.

I think we've reached a stage (Greg might elaborate on this as he pays more attention to it) - Rust is truly becoming part of the kernel, and it's no longer just an experimental thing.

Of course, it took longer than I expected, no doubt about it.

Dirk Hohndel: Actually, the more prominent 'intense controversy event' before wasn't entirely related to Rust. The first time a component was removed from the kernel also had nothing to do with Rust. It was actually completely due to interpersonal relationships.

Linus Torvalds: Yes, this year has been a bit turbulent. We had a lot of disagreements, and we even removed some kernel functions to reduce friction.

To be fair, this isn't the first time this has happened. Some modules that were no longer in use or had serious problems have been removed from the kernel before. In 35 years, this has happened very rarely, and it's not a pleasant thing, but I think we've handled it quite well. After all, this is a large project, with thousands of people involved in each release, and there's a release every two months. You'll have personal disagreements, professional disagreements, and frictions. It's all part of life. I think we're still a happy family on the whole.

Dirk Hohndel: I'd rather describe it as a group of very mature people who have found a way to get along with each other. But I'll go along with your 'happy family'. Usually, this is the first thing I ask you, but today, at the end of the first part: Do you have anything to say about version 6.18 RC4?

Linus Torvalds: No. That's the current kernel version. I like 'boring'. For me, 'boring' means there are no super - exciting new features, and it won't cause the machines of millions of people around the world to crash. Version 6.18 doesn't seem to be a problematic version. We had a series of test failures, but it turned out that to a large extent, it was actually a failure of the tests themselves, not the kernel. I was a bit worried a few weeks ago, but now it seems to be moving towards another incremental, boring - in the best way - release version.

The rise of Nvidia and AMD hardware and its impact on Linux

Dirk Hohndel: If we look at the major changes in the industry, I think one of the biggest changes is in hardware. For decades, everything revolved around the CPU. Everyone was talking about CPUs. Who had the fastest CPU and the best architecture. In the past few years, with the rise of companies like Nvidia and AMD, accelerated processing units (APUs) have become the focus.

Interestingly, although these processors are related to Linux machines, Linux doesn't actually run on these processors. What do you think of this trend of the hardware focus gradually moving away from Linux?

Linus Torvalds: I don't see it that way. I still think the most interesting part is the general - purpose CPU. It may not make it into the news as often because it's been around for a long time and people are used to it. What Linux does is maintain the system, start the system, handle the UI, and all the things you expect a system to do. The AI part is the new darling of the industry, and that's okay. However, it's not completely independent. It's a different environment that Linux helps to nurture and realize, and I don't think the kernel has to be an extremely indispensable part of it.

For me as a kernel maintainer, this is actually not fundamentally different from the user space. Although I personally love open - source and don't want to be involved in non - open - source projects, open - source has never been a religious belief for me. I do open - source, and Linux is also open - source, but people have always run commercial applications on Linux, such as large databases, cloud services, etc. This is very normal.

For me, the GPU is just another form of the same thing. You run your AI workloads on top of the kernel. The fact that it has its own system to maintain the GPU hardware usually isn't something that Linux needs to worry too much about. We're actually involved in it to some extent. There are a lot of things like resource management and virtual memory processing that the kernel is deeply involved in.

This is actually one of the benefits brought by AI. It has made Nvidia a good participant in the Linux kernel field. As we all know, it wasn't like this 20 years ago. Now, when Linux is so important for AI clouds, Nvidia suddenly cares a lot about Linux, and we have a lot of kernel maintainers in that field. So this is one of the positive aspects of the AI boom.

"The application of AI in the Linux kernel is at best experimental. I've never played with AI - assisted code."

Dirk Hohndel: I think whenever a vendor embraces what we do and gets involved, it's a very positive thing. It's great. Since you've mentioned AI so many times, I have to talk about it.

Last year, we talked about the potential of AI or generative AI in code review and code explanation. The Linux kernel community has done a lot of work around this. How's the progress now?

Linus Torvalds: Well, it's not quite there yet. Some people are really doing a lot of work. Some are trying to use AI to help maintainers handle patch streams, backport patches to stable versions, etc. To be honest, most of it is still experimental. The biggest problem we've encountered is that AI has a great impact on the infrastructure. For example, AI crawlers are scraping the source code on kernel.org everywhere, which has caused a huge headache and isn't always pleasant.

But there are also some good aspects. I'm looking forward to the day when AI is no longer over - hyped and becomes more of an everyday reality, and no one will keep talking about it. Obviously, it will be a few years before that day. I think exciting new technologies are always topics that people want to talk about. Of course, with trillions of dollars being invested, people are becoming more and more curious.

Dirk Hohndel: One thing that impressed me was that at the Amsterdam Open Source Summit, Daniel Stenberg of Libcurl mentioned that the low - quality security reports generated by AI were almost like a 'denial - of - service attack' on his project. Have you encountered a similar situation in the kernel?

Linus Torvalds: We've also encountered it in the kernel, but it's not as serious. But we've indeed seen some bug reports and security notices that are obviously fabricated by someone misusing AI. This will consume the resources of maintainers. In some projects, this problem is more serious than in the kernel.

Dirk Hohndel: Another topic that everyone really wants to talk about is AI - generated code. I often compare it to 'enhanced auto - correction' because AI is really good at code completion, syntax checking, and using standard libraries. On the other hand, the Agentic AI that people talk about a lot now - basically, you tell the AI, 'Hey, Claude, I want you to develop this function,' and some people even say, 'With the help of AI, I made a complete product in a week.' Have you played with these things yourself?

Linus Torvalds: I haven't played with them at all. But I'm sure someone is researching them and even wants to apply them to the kernel codebase. However, I think the kernel is complex and special enough. Although we've open - sourced a lot of code for AI to learn from, it's very difficult to directly apply it to the kernel. I estimate that few people will write kernel code in the Vibe Coding way. It's more likely to be used in their own small projects.

Actually, I think most of this is a good thing. When I was a kid, the way I got in touch with computers was very simple. I typed out programs line by line from magazines. That's how I fell in love