 I'm happy to be here today, and I'm happy that you all got up so early this morning, given that most people probably started having dinner at about 10 p.m. last night. What a wonderful town, you know, such a beautiful location here in Bilbao, and throughout the whole Basque country in Spain, it's really good to be here. Today I want to talk about the state of open source and the state of the Linux Foundation. What I can tell you about the Linux Foundation, I'll kind of get that part over first, things are going amazingly well. Our biggest challenge is hiring smart people to come in and help facilitate all of the collective development that happens across the now thousands of open source projects that we steward. And, you know, at the Linux Foundation, we have a phrase that everybody talks about all the time, sort of our cultural phrase, humble, helpful, hopeful. The humility has to come from the fact that every day in our communities, roughly 800,000 developers work on all the different projects that are housed at the Foundation. And we depend on facilitating that work for the outcomes and impact of the organization and of these open source projects. And almost all of those people are not Linux Foundation employees. So if you don't have humility, you really can't work to influence outcomes because, you know, if ego gets in the way, you can't sort of lead through influence. That's something important to us. Being helpful is just, hey, you know, the rock stars of open source or the developers and technology is just that create this great code and standards, and we're just here to facilitate. So, you know, what can we do to be helpful, to create good developer experiences, maybe to help with documentation, to bring new people into the community, and so forth. But the final part, the hopeful part, is what I think is most important. Because you know, in the open source community, we tend to like a good crisis, right? Oh, no, you know, like there's, you know, there's a licensing crisis. What are we going to do about AI and open source and, you know, the sky is kind of always falling, falling and, you know, oh, you're never going to be able to do that. This open source thing is not going to work. If you're not hopeful, you know, it's just really hard to kind of stick with it and have the grit and determination to get these amazing outcomes. And so today I want to talk about some things that, you know, we've all been talking about in the open source community where we're a little worried that something terrible is happening. You know, I think one of these things are around licensing, one of these things are around AI, and the first one I want to talk about is around licensing. How many people here have heard about the recent license change of Terraform? There you go. See, this is my crab. You're all my peeps. How many people here know this sort of niche open source licensing conversation? So HashiCorp, a commercial company, recently changed the license of a very popular open source project, Terraform, from an open source license, the MPL-2, to a non-open source license, essentially a business license. And there's been a lot of chatter in the community that, oh my goodness, the sky is falling. And I would like to offer a counterpoint to this, which is I don't really think the sky is falling because HashiCorp changed their license to a business license. You know, I think the community and industry are just getting more mature, and you're going to see today how actually the sky isn't falling, it's just getting sunnier. You know, HashiCorp, it's their business decision to change the license for the copyright that they own, and they're a reasonable company. I think some people might argue with that, given this change. But you know, at the end of the day, it is essentially their business decision. And when I saw all this chatter in my news feeds and from our staff, like, oh my gosh, there's this huge shift from open to these business licenses out there, we've got to do something, I kind of looked into it. And I asked the team, you know, find me every single example that you can find of a commercial company that owned the copyright in a popular open source project that changed it from an open source license to a commercial license. And this was the list they came up with. What do we got here, nine? Total. Now for all of you open source wonks out there, I truly challenge you to add to this list, please. But what this is telling us is out of the thousands of critical open source projects out there, out of the hundreds of thousands of long tail projects out there that people use every day, is this really that big of a thing? And the answer is probably not so much. And the even better answer is that the community and the license agreements in open source provide a good way to adapt to this change. But I do think there's one thing that you should all know as to why does this seem to be happening recently? You know, a lot of these are 2021, 2023, why are we all of a sudden kind of seeing this thing? The likely answer is about five, six years ago, seven years ago, there was a lot of private equity and venture investing in open source. Venture capitalists looked for popular open source projects and went to the maintainer. Let's start a company off of this. And a lot of capital was raised and people went to market. Using open source is a way to go to market. It's a great way to go to market, right? We all know that, right? More usage begets more interest and that enables you to go and sell commercial products in conjunction with the open source project and that's a good go to market strategy. Well, now most of those companies are maturing. They might have hundreds of millions of dollars in revenue. The company has grown significantly. They're maybe looking at going public and they're starting to look at their business and try and make decisions about, hey, was that open source thing good or not on an ongoing basis for us? And that's kind of it. And it's their prerogative to do so. I think from the Linux Foundation's perspective, it's a further evolution of how users of open source software and co-developers of open source software need to better educate themselves on the risk of a century-long copyright regime from a single company that gives them the ability to change a license later on. And I think Dirk Rale, a European researcher, did a good job of kind of describing this, right? You know, if you look in the upper corner there, this sort of vendor-owned open source, you have open license, but you don't really have open governance, right? And that allows for this sort of switch to a business license later because you don't have this neutral open governance. What I think we're seeing is a good lesson for all of us that this upper right corner, the idea of both open governance and an open source license is really the sweet spot that enables commercial organizations to invest in the code base because they know it's not going to be switched, it's not going to be owned by a single entity, and that in particular, if it's housed at a neutral organization, like the Linux Foundation or the many open source foundations that are out there, we're not the only game in town, but if it's in the nonprofit foundation, that's really the sweet spot. They're not going to change the license, you know you can trust it, you know you can co-invest, and that's what gets the best outcomes. And so that sweet spot, once HashiCorp changed their license, almost immediately flung into action via a well-known method in open source, which is the right to fork. And that right to fork coincided with that open governance, and that is what we want to talk to you all about today. We have an announcement for a new project called Open Tofu. This is... That's right. This is a drop-in replacement for Terraform. It's open source. It is neutrally housed at the Linux Foundation and hits that sweet spot of open source and open governance. And here to tell you more about this is an Open Tofu core contributor, Sebastian Sadle. Come on out, Sebastian. All right, exciting times. Hi, my name is Sebastian Sadle. I'm one of the core contributors on the Open Tofu project, and I'll talk a little bit about that. First of all, it's just so hard to believe that all of this started just five weeks ago. It feels like it's been an eternity. Just five weeks ago, a bait and switch or rug pull was made, and so that got the community into action. And in the space of five weeks, we've had hundreds and hundreds of developers pledge to contribute to the project, 150-plus companies. We've got 41,000 stars on GitHub between the manifesto that really resonated with folks, as well as the fork itself we just opened just a week and a half ago. And it's been trending on GitHub as the between number one and number three project for weeks now. It's one of the fastest-growing open source projects in the cloud space. And so we're incredibly excited about the momentum and getting the community going on the project. And it's not just individuals and developers like myself that are excited about this. It's also enterprises. And so enterprises make large investments into open source, and we're continuing to see that. So here to talk a little bit more about their usage of open source and their excitement for open tofu is David Berhal. Please welcome him. So David Berhal is the head of software engineering at Alliance. And could you introduce yourself a little bit? I'm heading software engineering in Alliance Indonesia. For those of you that don't know, Alliance is an insurance multinational company. We are the third-largest insurance in the world, according to Forbes. And in particular in Indonesia, we are the number one in the market in life and health insurance business. And you were telling me backstage that you guys use a lot of open source. Can you tell us a little bit about how important open source is to the company? Yeah. So Alliance's purpose is we secure your future. When it comes to taking ideas, the decisions, we always ask ourselves, how is this going to secure our customers' future? Open source tends to be most of the times the right answer to that question. That's why many of our business critical applications run entirely on open source. And you were telling me as well that you guys are huge Terraform users, right? You've got to deploy it across many applications, many business units, et cetera. So Terraform and infrastructure is coded in general. It's super important to you guys. Can you tell us a little bit about that? Yeah. When it comes to securing the future of our customers, IT operations is business critical. It's critical for them. Doing infrastructure as code provides us with resilience and agility that is much needed, particularly for us in Indonesia, living in the ring of fire, you can imagine. IT operations is critical for securing our customers. And we had invested millions of euros in developing Terraform code. So when there was this change of license, we were wondering, is this still the right way to secure the future of our customers? But now with the announcement of Open Tofu, we are so happy to join and to pivot to Open Tofu. It's great to have you in the community. Thank you so much. Let's give him... Thank you. Quick show of hands. Who here has heard of or uses ExpressVPN? It's amazing to see. Here to talk a little bit more about Open Tofu as well, let's welcome Carlos Clemente. He's a staff platform engineer at ExpressVPN. And he'll talk a little bit more about some of the work they do. Just give him a warm welcome. Hello, everyone. Thank you. So yeah, I'm Carlos Clemente. I work in ExpressVPN as a platform engineer. And in ExpressVPN, we value our customers' privacy and their right to... Well, to be... Whatever they want... To private, right? Sorry. So we use Open Source a lot to meet that goal. And we're very excited to be here now with this new excitement on Open TF. We use Terraform a lot. As a platform engineer, we try to reduce complexity using Terraform, not just for infrastructure, but for any kind of cloud resource we managed. And we believe that Open Tofu is a game change in the next step on the IAC ecosystem. So yeah, we believe that Open Tofu is going to bring power. It's going to empower not only individuals, but also enterprises. And yeah, apart from that, ExpressVPN also believes a lot on Open Source. That's why we open our core. We call it Lightware Core. Everybody can use it. It's a different protocol. And with that in mind, we believe that we can contribute a lot, not just use Open TF, but also contribute with maybe ideas. Maybe we can also contribute some code. But yeah, game on. Awesome. Thank you so much. Yeah, we're looking forward to the ideas and the contributions and all of that. Thank you. So the Linux Foundation is an amazing place for any software project. They provided Open Tofu with an amazing governance framework to make sure that across organizations, across sometimes competing interests that everyone can come together and make a great project. And so they provided just a fantastic framework for that. There's the Technical Steering Committee, if you're not familiar with that, that's kind of like a Supreme Court for Technical Arguments for any disagreements on the vision or the direction for the project. The Linux Foundation has also provided us with other things for security and legal governance. And it's just an amazing place for hosting and governing a project as important and critical as this one. There's a number of things that... I'm a developer. I'm sure you guys are developers too. So the next big thing here is an official release that will come very soon. But right now the repository is open. And so if that's a first raise, just getting a release out, that's a drop-in replacement. The next phase is starting to pull via an RFC process ideas from the community and features improvements and start drawing upon the brains of folks all over the world to make a fantastic project. And all the while maintaining that backwards compatibility to make sure that you can use and migrate to OpenTofu at your own speed. And with that said, a little call to action here. There's a talk tomorrow at 3 p.m. Please do not miss it. There's going to be two developers that will answer whatever questions you have as well as provide a little bit more information on the vision for the project, how we govern it in more detail, etc. And there's a QR code as well. If you use that, it'll take you to the Slack community where we're all talking about the project and what to do, how to self-organize and all of that. So please consider joining that community as well. Thank you very much. Seeing amazing outcomes from this project and we talked about that sweet spot of open governance and open code. And last year, Meta contributed PyTorch to the Linux Foundation as well. So TensorFlow and PyTorch were somewhat similar in that they were open source and they're both great open source projects. Google is an incredibly good steward of open source, but in this case, TensorFlow is owned and managed and controlled by Google. Meta decided to contribute to the Linux Foundation, their technology. And just in one year, we've already seen what happens when you hit that sweet spot of open source and open governance. I expect we'll see the same thing from the open-toe-foo project. So what we've seen is that what we thought was this big, terrible trend in open source around license changes has a pretty straightforward reaction to it and the sky is not necessarily falling. So the next big thing that I've been hearing a lot about is what are we going to do about open source and AI? What is the role of open source and large language models? You know, should we even have open source large language models? Everything is terrible. And what I can tell you is large language models, things like all of the great technology we're seeing out there around generative AI are largely built on open source components. And while the sky is not falling, I think we're seeing more and more open code, I do think there are some things that we should all think about around large language models and we should resist some calls that we've been seeing out there to regulate or restrict open source large language models. Recently, we've heard from different people around the world, largely folks that already have a lot of capital, a lot of GPUs and good foundation models that we need to take a six month pause on all this stuff until we figure it out. We're even hearing calls for folks who are saying, hey, this large language model technology, this advanced AI technology is so powerful that in 20 years in the hands of individual actors, people could do terrible things, create bio weapons, create massive cyber attacks and so forth. And what I'm telling you folks today is that kind of fear and that kind of concern that the availability of open source large language models would create some terrible outcome simply isn't true. That open source always creates sunshine and that fear has a counterbalance around hope because it's not just bad things people do with large language models, it's good things too. Like discovering advanced drugs, helping to manufacturing to become more efficient, using large language models to create more environmentally friendly building and construction. Like for every action there can be a reaction and we're already seeing open source immediately start to tackle some of these things people are concerned about when it comes to AI because there are things we should legitimately be concerned about when it comes to large language models, job displacement, bias in these models, hallucinations, security implications, these are real concerns. But those concerns aren't going to be addressed well by simply trying to ban open source large language models. What's more likely is that the open source community and you're going to hear a little more about this later is going to create tools that root out bias that can answer some of these ethical complaints that genuine concerns and the open source community can also work with regulators to address those things. You see we've seen this movie before around trying to ban open source in certain areas of technology and we know how it ends and that was cryptography in the 1990s. How many people here remember when we were, the US government treated cryptography as a munition and the concern was, if we have open source cryptography it's going to be terrible, we need to lock this down, we need to have security through obscurity and Phil Zimmerman, an open source developer who published some cryptography code actually was arrested and a series of court battles went underway and the way it actually ended is Phil was exonerated. Open source was seen in the United States as free speech and what we learned was if we try to lock everything down people who don't care about ethics are just going to go work on cryptography anyway. Open source is not something as a digital instantly distributable good something that can be easily contained and that it's better to have this stuff in the open and to work collectively for our collective good than to try and control it and lock it down. Let's not repeat that mistake. Open source and AI allows for better examination of how these large language models work it allows for better collaboration and knowledge sharing about how we can work technologically to solve some of these big challenges around it and democratization of access to these tools largely gets better positive outcomes and the concerns around the negative outcomes and the same holds for open data. Data and code are essential in building large language models and we need to make sure that we have both of those things open this is why the Linux Foundation created our community data license this is a common open source style license for data that allows people to easily share large data sets for the creations of LLFs. Wow we got some CDLA fans in the audience I love it. You're my people. I love it. But I do want to ask y'all to help me out with one challenge when it comes to generative AI in large language models so I'm going to give especially the developers in the room some homework today. This is sort of the left hand side of the software supply chain where you know developers create code they build and then package that code and then it goes to an end user this is sort of the writing of that code side and my question to developers is really simple does any of this stuff work how good is it can we use this stuff to make even more code right there's all this concern around code completion tools and that'll put developers out of jobs and it'll be terrible well I can tell you that that's very unlikely to happen these tools are really going to be a force multiplier for the creative minds we see in the open source community because just code alone is not the same as context like writing is not the same as thinking and anybody who's worked in the open source community for a long time knows that it's the creative minds that make for great software not just generative AI style source code development so I'm asking all of you today to go check out some of these things our code documentation tools that are out there good let us know do these code completion tools work effectively are there better ways that we can do testing and debugging and fuzzing of code and let us know the Linux foundation is dedicated to providing the best developer experience possible for our communities and we really want to know how these work and we understand that there are concerns about these tools risk management around intellectual property that's already being addressed by features like similar code suppression using S-bombs in order to understand the provenance of where source code and the license that it derived from came from and so on and so forth so each generation of these LLMs around software development are getting better the tools are addressing concerns around intellectual property regulators are working on making this more clear in terms of what's a derivative work and so forth but the simple thing I'd like y'all to help me with is does this stuff work and how could we apply it to our open source communities in a way that would create better outcomes and if we do that I think that the world of open source isn't going to be catastrophic but it's actually going to be pretty amazing now as we talk about AI there's one other big trend happening in the world of large language models and that is the shift from sort of general purpose compute to accelerated computing you know it's ironic how many people here love video games play video games it's thanks to all of you gamers that we have large language models in the first place right? Nvidia back in the 90s was this you know GPU accelerator for games and it's you know they their business was essentially made off of gaming and this was in an era where general purpose computing just Intel was the king of general purpose computing and accelerated computing was a big big niche well today that's kind of flipped because of large language models accelerated GPUs are you know there's a two year backlog to get this stuff and it's becoming one of the main streams of how computing is done and what this is telling us is that we need more standards that can meet the challenges of a world where instead of general purpose hardware we have this accelerated hardware and that's why today we're announcing the UXL foundation this is a multi architecture multi vendor software ecosystem for accelerated computing it's a set of standards that unify different heterogeneous compute ecosystems around a set of open standards technically we announced this yesterday so I stand corrected on that this is something that's going to be really impactful it is a derivative of the one API initiative which focuses on a set of development specifications for open source projects and I would like to introduce the folks from the UXL foundation Rod Burns, Sanjeev and Vivek who are going to talk about this new initiative so why don't you come on out okay thanks Jim it's great to be here and we're really excited to announce this unifies acceleration foundation or UXL foundation for short I'm also joined here by a couple of the founding members for this foundation which is great Jim talked a little bit about the landscape of computing is changing the use of accelerators is increasing rapidly one of the challenges that developers have is that whilst they can write software for these accelerators there's no unified way to actually do that so what we're doing is we're getting together these organizations from across the industry to define an open standards an open source way to develop software for all accelerators and so together we want to build the largest open ecosystem for accelerator computing and we're not starting from a you know not starting from a clean slate a blank slate you know we have a solid foundation to build on Jim mentioned the one API specification the one API projects that are open source and being contributed so we're starting from a really good place I'm joined with me here with Sanjeev and Vivek you're going to say a few words about their involvement and why they're here so Sanjeev thank you Rod you know we're frequently told by our customers that it's really really hard to write software for accelerators without building a dependence on closed source proprietary technology right so one of the things the charter of UXL is really to build a fully open source fully open spec set of software that can run anywhere on any accelerator that can be ported to anywhere and so on so I'm Sanjeev from Intel proudly representing the team that built one API Intel is contributing with the one API software stack to UXL and we cannot wait to see what UXL will build with it and can't wait for all of you to join us so thank you thanks Sanjeev and we're also joined by Vivek from Fujitsu good morning I'm Vivek here so as Sanjeev mentioned for us with our long history in HPC we believe UXL is going to be critical for us going forward as you know the amount of cost it takes as he mentioned on the software for accelerator I mean it is just mind boggling and that impedes the progress so we are absolutely very proud to be part of this and I think it's a very key initiative for us thanks thanks Vivek and as the chairperson for the Steering Committee for this new foundation you know I encourage you to come and give your comments contributions and even join us as members you can find us on the website it's uxlfoundation.org take a look and please get involved in some of the special interest groups thanks thank you