 Voila! You can see it and it works. Yes, it works. So, welcome to this wonderful keynote for Powering Open Source Climate with Operate First. So, my guest speakers, and I'm the moderator, but I also have a small section, which is interesting. But Michael Thiemann, I'm super honored to speak with him. He's one of the few persons I know who really have a Wikipedia entry and if you're old enough and you've been typing G++ in your terminal window, he's the guy who created this addition to GCC. He then founded Cygnus, a company which was then acquired by Red Hat and he's been the CTO of Red Hat for quite some time and now he's leading or he's part of OS climate and he will talk about his role in OS climate during his section. So, welcome, Michael. We also have Eric Allenson. He's styling in from Phoenix, Arizona. So, kudos to him. It's now 6 a.m. in the morning for him and he's actually in this wicked place where they don't have daylight saving time. So, I have to think twice every, no, two times a year I have to think about what time zone is he now actually in. Is it seven hours minus eight hours minus? I don't know. So, Eric is a really early bird. He's been part of the AI Center of Excellence and AI is a small hint to his interests. He's super interested and working on all kinds of data science, machine learning in the cloud native world. So, he's working on how do we create Spark workloads on Kubernetes. He's contributing to Ray and he knows all the stuff about big data and AI. Welcome, Eric. And I'm a senior manager in the AI Center of Excellence, now Open Services at Red Hat and I've been focusing on AI ops and operations, hence operate first, which also hints a little bit at operations and my passion is also how can we teach robots and machine models to create better tools for operating the servers that we have at hand. With that, I would hand it over to Michael. Thank you. Thank you so much. As Vice President of Open Source Affairs at Red Hat, I am on loan to the OS climate project to be the OS climate project lead. And if we go to the next slide, you will see some of the members who have joined the OS climate project. Some of the largest banks in the world, stock exchanges, but also technology companies, are pooling together resources to address this a massive challenge that we face of getting our financial capital markets better aligned instead of misaligned with the current global challenges we face with respect to global climate change. So if we go to the next slide, I'm not going to bore you with a lot of details, but just as one small example, the Nobel Prize Committee actually recognized some scientists just a few months ago who predicted correctly how much the Earth's atmosphere would warm based on the introduction of carbon dioxide. And their 1960s computers working with a 1970s model correctly predicted a 30-year future of what the temperature would be in the year 2000. We continue to recognize that we have very little room to maneuver and maintain a 1.5 degree scenario. And yet we have $100 trillion of assets that are variously taxing and changing the profile of our carbon footprints. The goals that the net zero asset owner alliance have set is to use their financial capital to invest in a net zero future. And it sounds so easy. It sounds so easy to do things like I'm going to change my diet and exercise to become healthy. And it sounds easy because until you try and do it, until you try and ask the question, do I have access to healthy food? Do I have the ability, the freedom to move, and all these other things. So a grand challenge that we have is to figure out how to adapt a capital market system that has long ignored the realities of climate change into one that can be responsive. And to just recognize that given the current rate of change of our climate, the concept of a stable investment like a 10-year bond is something maybe a little bit out of a fantasy world. But let's go to the next slide and let's look at words from these members. This is a slide that was presented at the COP26 conference in Glasgow, Scotland by one of our founding members, Allianz. And they talked about the massive portfolio of assets under management that they handle as part of their global investments business and insurance and so forth and so on. It's almost a trillion dollars of assets under management. And the big change that has occurred in their mind is that it is no longer just a question of risk versus return but also a question of what is the real world impact. And adding this new dimension requires a complete retooling of the analytic tools and data sources that they use to make decisions across their asset classes. And when they looked at the scope of change and when they looked at the time frames available for making those changes, they realized they needed dramatic acceleration in terms of innovation and they needed a much broader base of participation. And they looked to the success of open source software as a model for bringing about this digital transformation. And no doubt, many people at Red Hat are very familiar with how open source not only brings technical innovation but also transformative thinking and potentiality to companies. And so this seemed like a very natural and exciting approach. And this is one of the reasons why I felt that I could bring my open source experience to this project and help make it successful. Next slide. So now that is funny. That does not look correct. We've got a rendering error. There we go. Ah, you know what it is? It might be a build that shouldn't happen, but in any event, implied temperature rise is related to that first slide I showed about how additional carbon raises temperature. And so the OS Climate Project has built an open source software tool that basically takes data about projections of what kind of emission intensity, what kind of emissions targets companies have set, and also the relative requirements that global policy people believe are necessary based on industry allocations to come up with trajectories, budgets, and targets to become aligned with the global carbon budget. They calculate these overshoot or undershoot ratios and temperature scores and come up with a final score that can then be weighted according to various portfolio parameters that they could be weighted based on market cap, they could be weighted based on revenues, they could be weighted based on other factors. And the goal of this tool is to help asset managers understand whether their investments are aligned with global portfolio goals or not. And one of the things about this is that it requires use of data, which is typically highly proprietary, but organizations, data providers, such as the London Stock Exchange Group and Standard & Poor's have committed in principle to make this data available for the purposes of doing climate aligned financial analysis with open source software tools. Next slide. So what we are building at OS Climate is a way to reach and unify all the different types of data that are necessary to make these climate aligned finance decisions. And this relates to data from the natural environment, it relates to data from the economic world, from the physical assets that are placed in latitudes and longitudes, it requires understanding corporate ownership, corporate identity, relationship, hierarchy, it reaches into the stock markets, it reaches into production and emissions data. And the size and the scope of this is really quite large and building these relationships is one of the great challenges because most of this data has never really been linked together. And one of the premises that we have for OS Climate is that by making this data, by encouraging participation and connection between these various communities with open source tooling to promote open data sharing, we can provide the information necessary to reprice and reinvest those hundred trillion dollars worth of assets that are presently pushing the climate in a particular direction and need to start pushing it in a different direction. Next slide. One of the most exciting developments of the project came at the end of last year when Airbus Industries joined as a member and brought with them an open source contribution named WITNESS. WITNESS is an acronym for this world environmental impact and economic scenarios integrated assessment modeling platform and I just want to quickly take you through the concept. Airbus, as everybody knows, manufactures airplanes which are large energy users and also very expensive devices and what Airbus has recognized is that when they think about building their next trillion dollars of inventory, that trillion dollars of inventory is going to be landing in a world that needs to be a lot closer to net zero emissions than we are today. And the transformation necessary in how they design aircraft, how they build aircraft, how they service and support those aircraft, everything needs to change and not just for Airbus. They can't make the change themselves. They need the entire supply chain to walk with them into that new future and so by building an open source modeling framework, they can inform not only themselves about the changes that they need to make, but they can communicate to the rest of the world the other changes that others need to make to align their activities. So what WITNESS does is it takes data from the policy world, from the economic world, from the energy world, from materials world, from the world of natural resources and the environment and allows for the evaluation of various modeling assumptions, various production assumptions, various economic assumptions and basically helps answer the question whether a feasible solution is possible under a set of assumptions in a given scenario. And by doing all of this open source, the transparency builds the trust that influences the decisions that gives people the courage to make those decisions work. And so that's just one example. In our parlance, we talk about transition risk. This integrated assessment modeling system aimed at understanding and acting on transition risk is one piece of a larger landscape of OS climate. Next slide, please. So generally speaking, as is probably no big surprise, there's a set of open source tooling which is necessary. For me, that's the fun stuff because code is creativity. There's also a global data commons that we're building. This is the hard stuff because data are facts. And facts don't always tell you what you want to hear. Sometimes it's very difficult to integrate facts because they contradict each other based on implicit world views that were not understood when the data was gathered. But in any event, without the data, the decisions cannot be made. And then finally, a set of specific analytic tools which have methodologies that could be validated or improved by scientific and academic research bodies, as well as improved and optimized by the coding community. Next slide. So, I told you that there's a lot of pieces to the puzzle. And here's just one example. If we look at this as sort of a clock face, if we're somewhere between one and two o'clock where we talk about physical and transition risk. So I shared with you the witness platform, which is a tool for assessing transition risks. Transition risks are the decisions that we can proactively make as we go from one set of decisions to another. Physical risk is what nature brings to our doorstep every day. When we have a physical risk event, we don't get to make a decision about what happens or where. It happens. And financial investors need to understand whether it's a sudden event like an earthquake or a gradual event like sea level rise, nature bats last, and nature has a strong say in what happens. And physical risk modeling and physical risk analysis is part of how we understand what are our vulnerabilities, what are our exposures to the climate hazards that we are facing. As we go around this circle, we see that everything is connected, understanding climate-related risk drivers, being able to collect everything from satellite sensor data to market data to economic and census data. There's a tremendous amount of data that touches all this together. There's a great challenge of greenwashing. We in the open source community are familiar with the problem of open source washing where people say something is open and it's not really. But in the world of greenwashing, we face a continual challenge of people saying wonderful things and then you look at what's really happening and it's not all they make it up to be. There's an important aspect of reputation that relates to scientific methodology that relates to the fidelity with which accepted science is implemented in code and then understood and communicated through analysis. There's the challenge of getting public support and we are hopeful that by making this software and this data open, permissive, inclusive, and inviting, that we can garner the public support because people will see and trust what they themselves can participate in. And then finally, this may also lead to better public policy that again gives us a leg up when it comes to dealing with these risky problems. Next slide. So I don't need to preach a whole lot about this open source approach but I will just say from a high level, the topic of governance is a very important one and again our governance is not just based on Linux Foundation principles and bylaws but it also is highly respectful of the importance of including the scientific community. As far as our collaboration structure is concerned, we are in the process of we have been using GitHub for a while but we're in the process of opening these repositories. We want there to be a set of repositories which when people come to visit them, have enough content, have enough documentation, have enough CI CD support, etc. that people can find their way around. And then finally, our default license is Apache of public license version 2.0. Our data license is CDLA permissive 2.0 and we are of course accepting and using data under various other open source licenses but OS climate is a Linux foundation and a proper open source project. And last slide for me, I believe, or maybe the last slide we'll see. The last thing I do want to talk about is a sort of a way of teeing up this data challenge. The central banks have written a wonderful paper that talks about the challenges that they see from a regulatory perspective, from a fiduciary and a financial security perspective, and they've identified in these three broad categories, the challenge of data availability, do we have the data we need? Is that data comparable? Can we make proper decisions based on the way that the data links together? And finally, reliability, can we really ensure that the data is correct and can be corrected as we need to do stuff? I have walked through enormous amounts of data since taking on this role and there almost isn't a data set where I don't find at least one glaring error. And to give you just an example, somebody can accidentally report something as trillions of something because they reported millions in a unit of millions by mistake. And that kind of thing happens largely because this is the first time people are making these reports. They don't have the process. They don't have the iterations completed to know, oh yeah, you got to check this, you got to check that. So it's a big challenge. Next slide. So I'm going to pass now back to Marcel Hild because one of the super exciting things about working with Red Hat is not just its great depth and knowledge about open source tools and technology but really innovative thinking about making this manageable and deployable. Thanks Michael. So yes, OS climate meets operate first. I was at a hard emoji there because I think it's really, it's a good marriage. It's a good teaming. So if you remember last year we had that keynote about operate first, open services. How do we apply the open source principles to operations, to services? And it's great to see now with OS climate how the same principles are applied to climate change. How do we gather data in the financial industries? And if you think about it, that's just a little bit more up the stack from services and from operations. So operate first was born out of the idea that we need to make our software more manageable, that it's easier to operate, operable. And we have that concept of operators where you do your manual ops as an operator, as a person. You do that obviously in the cloud context and then you create an operator out of it. So that's the notion, the concept of this operate first idea, why we came up in the first place. And the good thing about working in the office of the CTO is that we can imagine how the world would look like in five years and how we can evolve the company or the communities, the principles and apply them to the current world. And I think we see a little bit of history repeating here like what we had in back in the days when I was just a kid with mainframes where you could only buy the computers from large vendors like IBM and they would upon a phone call give you access to that CPU, which is already inside the mainframe. And it's similar with the cloud age nowadays, where the only power that you have to the hyperscaler to the cloud provider is filing a support ticket to get you more access to the API. And if you think more about this paradigm shift of applying the open source principles not just to the source code, but also to everything that has stakes in the source code in the thing that you're actually running. That means operations, that means services, but that also means the data and how we work together. So operate first is really in its essence, I think, re-envisioning open source in the current age of cloud and age of scale. So I'm just going to read it out. Operate first is an initiative to operate software in a production-grade environment, bringing users, developers and operators closer together. And I think it's more than just users and operators and users. It's really a community that is aimed at all the different personas that work on a certain use case, like the use case that we just saw with OS Climate. So we have people that are doing operations that really maintain this stack. We have the developers that work on parts of the platform or developers that deploy their workloads on the platform. And we have end users coming to use the platform and to use the workloads that are deployed there. You have end user support. And you have architects that are using those building blocks, those LEGO bricks, and build new reference architectures and maybe apply the learnings from OS Climate to another industry. And as you saw, Airbus is not really in the financial business, but they are coming from the airplane manufacturing business, but they have a shared interest in this. So I think what we're building there can be always reapplied to different use cases. That was the beauty of open source where you just fork a repository or you just download something and then you make changes and you tweak it to your own use case. That's not possible with hyperscaders with the large clouds, but it's possible in a really open environment, in a community environment, where we come together and build something. So that was the philosophical underpinning, but if you really want to do something, where can I touch it? So another way to look at it is we are building in one implementation a community cloud, a cloud, hybrid cloud environment with full visibility into the operations center. So everything that I will touch now is out there and you can look at it from a user perspective, from an ops perspective, from a deployment perspective, it's all out there. Pretty much like you can go to a software repository of an open source project and go as deep as you like. We have a bunch of hard bare metal deployments at data centers. One is in the Massachusetts open cloud in Boston. One is at a German data rec space, kind of like provider. We have stuff running in AWS as a matter of fact, OS climate is running most of their workloads in AWS. And we're working with other universities or clouds to also onboard them into this community cloud. So it really should be a lot of multi-geo, hybrid, compute resources available to the community so that we can deploy workloads on that environment. Open data hub has one of the largest or larger use cases and larger projects running there and which also is heavily used by the OS climate community because it's all about data. So why not use that open data hub which is running there already. But we also have stuff from the middleware, from the Java community and Project TOF with all their tooling to do a lot of management like build pipelines, CICD. That's all available for the community to use. We have IGO CD to manage those workloads. So it's really a GitOps driven approach for running these things and for managing these things so that you have that peak into the future. How would I do it on a green field deployment? So we can really experiment with stuff without the constraints of carrying an SLA, without the constraints of also supporting any legacy systems. And we treat everything as a service because I think that's the true essence of the cloud age. Treat stuff as a service so that your users don't really worry about whether it's up or running but they just use it. So we embrace the concept of operators. People can bring beta versions of their operators to that environment and provide PostgreSQL as a service because it's deployed via an operator and actually the PostgreSQL community could support that service running. They're getting immediate feedback from users and from the ops team to build operability back into their software. So in essence it's about sharing knowledge. It's about sharing operational data to build AI. It's about sharing best practices, blueprints, architectural decision records for making decisions that you would face if you were setting up such an environment. But as we saw with OS Climate, it's also really a great opportunity to support other communities with 80% of the stack that they would have to provide otherwise so that they can focus on their use case. So OS Climate wants to focus on how do I massage that financial data. They don't want to focus on how do I set up a CI CD pipeline. They don't want to focus on how do I do proper GitOps for deploying my software. They want to have an opinionated environment which is future proof because it represents the current state of the art thinking in operations and in services. And to be way more specific and way more applicable, I will hand it over to Eric giving us a talk or a walk through what has actually be done hands-on in that environment. Eric, we're getting bad sales at the moment. If you stop sharing your video maybe, maybe it's a bandwidth problem. I've been on weekly calls with Eric now for months and we've never had this problem but hopefully it will resolve. Hey, Eric, we still can't hear you. Maybe can you change your audio input and also are you using Chrome? Well, you just hang up and then join back. In the meantime, do you mind if I ask some question, Michael? Please go ahead. Michael, Tomas was asking about the ESG scoring and whether that is somehow standardized and transparent across the companies and how that makes it up. That's a great question. Let me give you just a little bit of texture on that. So I mentioned earlier the Net Zero Asset Owner Alliance which has over $10 trillion of assets under management and they are only investing in equities, bonds, etc. Companies who have a pledge to be Net Zero by 2050 at the latest and in many cases much earlier. It is not possible for them to invest in a company that doesn't produce the data required to evaluate the Net Zero pledge. So the European Commission passed something called the NFRD which stands for Non-Financial Regulatory Disclosures and that is the sort of the framework in which these environmental non-financial but still very impactful disclosures must be made. However, although there is this mandatory reporting requirement, there has not been a great standardization yet on how that should show up in terms of a report one can download as data ready for analysis. So the GRI, SASB, and many others have sort of CDP have all published ways to say you can make your disclosure with our format. Unfortunately, those formats are all different and it's a real nightmare and OS climate is working with both the European and the US-based financial authorities to try to establish what could be a common reporting framework that will make our lives all a lot easier. But TLDR, there is a well-accepted greenhouse gas protocol that talks about scope one, scope two, and scope three emissions and everything kind of points in that direction as it comes to those factors. All right, Eric, you are back. Let's see if your audio is better. We still can't hear you unfortunately. If you unplug your headphones maybe. Yeah, try the over-the-air microphone. In worst case, I can present the slides. I have another question that I can still ask you, Michael. It just related from Harish, individual financial institutions or even governments can join OS climate and whether they would get any benefits from joining the community. Sure. So as many of you know, the Linux Foundation has enormous reach in terms of people who are members of various projects and the OS climate project actually has a very open membership for NGOs. There are many ways in which governmental organizations are not themselves permitted to join. However, we have working groups that they can participate in and we have regular conference calls with these regulatory authorities on an informational basis. So I'm hearing something from Eric. Maybe it's better now. Nope. Nope. No dice. Should we continue on the slides then? Yeah, in the interest of time, let me do my best and let us see if, okay. So one of the things that Red Hat has brought to the table, which has been absolutely fantastic, is Red Hat has an outstanding discipline in design thinking and doing user-centric design. And so what we did is we brought this design process to the OS climate project to get away from the trap of just grabbing technologies and just grabbing supposed requirements and instead building these stories. And so we have a set of user stories. I'll let you read the slides because we have to catch up in time a little bit. The slides will be available after the presentation and we're recording this. But fundamentally, by starting with these user stories, we end up with implementations that then have a greater through line in terms of meeting the diverse set of requirements that we're aiming for. Next slide. So we have this wide range of data sources, structured and unstructured data, data which is wrapped up in databases, data that's sitting on websites and CSV files. People want to use APIs. And so we need to have a sort of a universal source system protocol to just figure out how do we connect to the data that we want to use. And then within the platform, what Vincent Caldera has been helping to define is a set of rules for allowing data infrastructure to function as a platform. And also this new concept that or at least new to me is concept of federated governance. Now we have done a lot of things, very ad hoc in the open source community for a long time. And many of these design patterns of federated governance are things that we would recognize in things like the Fedora community governance guidelines. But the good news is that this is really sort of stretching into best practices of management discussion and IT management. And so we're able to point to some senior references that illuminate these ideas and make people comfortable that they can participate because their voice will work the way it needs to work in a highly regulated, very compliance oriented system. And then finally on the data consumer side, there's just lots of ways that people need to consume this data, including ways that sort of cycle back and sort of re-integrate. But next slide now. So one of the... Oh Eric, you want to try one more time? Did you get your audio sorted? Maybe. Can you hear me? Yes. There you are. Awesome. Welcome. Welcome back. It wasn't working before the keynote, I brought it. I love it. It's just completely on brand for me. Okay. So yeah, did you get to the part where you were talking about open data hop then? I just did not actually utter the word. So you're fresh on this slide. Okay. So I'm not entirely sure what Michael teed up, but we had the need for obviously providing self-service data environments for people so that we could scale, both in terms of simply compute scale, but more importantly, actually people scale for having to service individual people all the time to stand up environments for them. We're never going to be able to actually, you know, grow a community. And so fortunately, we already have tooling for this, which we call open data hub. And open data hub is an open source downstream of the community coop flow project. And so it consists of a bunch of, you know, common data science tools, you can see a few up there that are somewhat loosely federated in the sense that, you know, when you deploy open data hub and you can do this using operator hub with open shift, you can mix and match what you actually want to deploy. And it's not just federated in terms of like the tooling that you can use from that operator. But of course, this is all running on a open shift cluster. And so it means that of course, if you want to integrate with other tooling, it's all just pods running on a cluster. And so you can, you know, easily install other tools to integrate with what you want. And so we've again, calling back to operate first briefly, you know, operate first has a lot of embedded experience now, standing up things like ODH. And so it was like extremely easy to onboard the cluster we had deployed with operate first when they're now managing this deployment. So if you look at this two axis view here, which I kind of like talking about like, what does ODH give you for data science? You know, on like the Y axis, you have personas. And so like it covers like the persona needs everywhere from like business stakeholders through that engineering data science all the way to actual IT operations and deployment. And then on the X axis, you know, it covers all of the stages of deploying a machine learning workflow or data science workflow in a modern, you know, cluster environment, which we sometimes say model to microservice. And, you know, I called out like, there's a lot of different tools here you can see that come with open data hub, the big ones in green are kind of like the centerpieces of what we've been doing on the data commons platform so far. So sort of like the Lynch pin is Jupiter and the Lyra for doing data science and then translating Jupiter notebooks to actual repeatable pipelines. The core of the data itself is Trino, which, you know, gives you basically massive ability to federate different data sets into a data mesh and then expose all of it through a single SQL interface. So it's extremely powerful for federating data. Michael team and himself has done the enormous number of use cases already taking these data sets and extracting value by just doing SQL joins. But under the hood, it's completely heterogeneous data sets. And lastly, we've doing a bunch of, you know, visualization. So like, you know, your business stakeholders, what do they want? They want to see a dashboard, you know, and visualizations and so superset, which has already stood up. It's also a very key component. The rest of these are in the wings, waiting to be deployed as we as we need them. So next slide. Yes. So how, you know, we want to sort of leverage the theme, which we may have seen bubbling underneath the surface in the previous content here is we're trying to manage the data as a product. And so like this touches, you know, the entire spectrum of personas and use cases, you know, all the way from just people who say, like, okay, I'm a member, I'm a prospective member, I have some data I'd like to contribute to the community. And then of course, the data engineering, they to exactly trying to make, you know, sort of logistical sense of all these different contributors and data sets. And of course, the point of all doing all this is to enable community data science in a truly scalable way. And again, scalable really in the human sense, even more than the computation sense. And because these are, you know, real, real people with real needs, it has to be, you know, more than just lip service to data quality engineering. I think Michael mentioned earlier, all of these data sets have non trivial quality issues. And much, much of the reporting is done with things like PDF documents, it's, you know, highly non standardized, it's just basically people, you know, typing out reports, you know, into PDFs and Excel spreadsheets. And you're just hoping that like, you know, it's actually correct. And so we have to have processes in place and repeatable processes for detecting errors and actually, you know, being able to not just not just fix stuff algorithmically, but have people actually be able to edit databases in a way that's trackable and in the open. And so what's a key tool here, I see, yes, ourselves teed up our sort of platform here. So I just mentioned the word repeatable. And of course, that's, that's a serious issue, especially if you're trying to do data science with things like Jupiter. So we have been leveraging the Elyra environment with Jupiter hub to a high degree. So it allows people to do exploratory iterative development data science. Of course, classic Jupiter is great for that. You can actually just, you know, iterate on things and watch, which are getting back until you get what you want. But then at some point it's working and you want to transition that into something more like a production workflow. It's like, well, Elyra is a great tool for this because it allows you to take these notebooks and build Elyra pipelines like down a little right there and save them off as objects. They're open shift objects that run in tecton. And so you can take your notebooks, check them into GitHub. You know, the images themselves can be built from the same GitHub repos using the open services CI bots that other people in the open services group have been creating for quite some time now. And of the images are all managed via Quay. We've set up an open source climate GitHub org in a corresponding OS climate Quay repository. Of course, all this comes together in an actual tecton pipeline run that people use. And so you can see there's a lot of moving parts here. We've built out a lot of documents that are trying to help people on board of this process. The people who have joined the community so far have been like extremely flexible and just generally game for taking on all these new tools. I've been super happy with the participation and how flexibility the community members. But just just deploying all of this again, you know, we've leveraged open data hub extensively and leveraged operate first extensively and actually running these things and configuring them. And you can actually go out on the operate first GitHub org and see all the pull requests we've submitted to, you know, make the configurations. And so it's all done in the open. Next slide. And so this last this last diagram is useful for like talking about where we are and where we're going. The at the bottom here, you know, the physical sort of storage layer or using a lot of just S3 stuff buckets for physical landing of incoming data. And we're mapping all this into parquet and wrapping it with a layer of Apache iceberg. And again, the thing that's actually exposing interface to all these things, the physical layer is Trino. And the tooling for actually manipulating the data at the user level, the data science level is, of course, the tooling I already talked about open data hub. And so that's about where we are climbing Maslow's hierarchy of, you know, data towards data science, you know, future work, we hope to get to in the coming business year is increasing our tooling around tracking provenance and open federated one stop shop authentication and security. We've done a lot of work with this, driving all of our authentication through GitHub, and it's working really well. But of course, sooner or later, we're going to need to sort of like federate authentication across multiple tools. And so we need better tooling around that. All the way up to things like, you know, standardized APIs for the future. I'm hoping most of this gets done ultimately through standard things like SQL, but people already are talking about writing, you know, their own microservices to support some of these data flows. And so now we're talking about, you know, open, open APIs. And so that's that's work for the future. Next slide. So here's the call to action. If you got people are interested in learning more or even better yet participating, you can see up at the top left, we have a link to the OS climate website, you can find more about, you know, what's actually going on is the very high level in the use case level. And you can learn more about open data hub itself at open data hubs, excellent website. And if you want to talk about the data commons platform or building or participate in it, please feel free to reach out to myself or Vincent Caldera on email. We are building out sort of developer onboarding guide. The link is at the bottom of that slide. And trying to teach people about all these tools I just talked about and how we're actually using them on our platform. So please feel free to engage with that. You know, become a member of the GitHub org, we can onboard you with credentials. And you can play with the tooling. Again, we've already, you know, assembled quite a large variety of actual fully open repositories there. People are creating notebooks, creating pipelines. And in the coming year, we're hoping to take those and actually run them in our upcoming brand new production cluster to really get some truly usable datasets. And you can be a part of that. And to add on that call to action, if you rather fancy operating or the tooling that Eric just showed or operating open Kubernetes classes or storage or whatever, go to operate-first.cloud where you find another active community that supports communities like OS climate doing all the fun stuff here. Oh, thank you guys. That's a lot of interesting stuff. And unfortunately, we don't have that much time for questions as I was hoping for. So a really quick one for each one of you. First thing, Michael, Eric mentioned that you're actually doing some Jupyter notebooks combining the social climate financial data or together. Have you actually found something really interesting and something the OS climate community really found interesting? Yeah, what I'll just say briefly about that is that most of this data science is about trying to figure out where the center of things are so you can get your data to all be comparable. But in the process, you find these outliers and every outlier tells an interesting story. And so I'm always telling my wife about these completely random connections that are absolutely fascinating when you touch the outlying data. That's amazing. Quick question for Marcel. Last year, we've talked about operate-first as some new community that we're trying to build. And it's been a year out. How did operate-first change since then? How did the community around operate-first grew? I think we just grew. So it's more people. It's more compute. It's more resources. The new clusters, I think it's almost doubled. It's more than double in size. And if you're interested in more tomorrow at 10 or 11, I am having a 20 minutes talk about operate-first with some numbers in it. The great thing with OS Climate was to see that we actually can support such a large community. And the challenges were more in like how do we talk to each other? How do we find a governance together? How do we organize things? So sometimes the technical bits are easy, but getting together in a sustainable way in a way that we can actually leave something behind for the people that are coming next after us, I think that was the larger challenge. But I think we're a good setup now. Perfect. Perfect. And Eric, let's talk about some technical challenges, can we? Not your camera, but about onboarding the pipeline that you presented on operate-first. How was that different where you had this community and people actually watching what you were trying to do? And how is that helping you with setting up the pipeline that you presented? Yeah, it's been fascinating because again, a lot of these people, of course, not only are they not really familiar with our culture and tooling, but they're simply unfamiliar with open source at all. You know, there was a great practically full day set of talks that Michael and Vincent and a bunch of other people delivered at the COP26 conference this fall where a lot of it was they devoted an entire talk to like, why should you want to use open source? And I think that the people, you have to understand because a lot of these people are coming out of finance. And of course, with finance, there's a lot of very real reasons why you don't want to have open source or at least not open data. And so you find yourself having to not just get stuff working, but you really have to explain to people while you're doing it, what it is you're doing, why you're using these tools, and why they should be comfortable writing their code in a public repository. And it's, you know, again, a lot of the there's a lot of work to do documentation. It's quite, it's quite a lift is, you know, just like I spend a lot of time just like getting figuring out how to do stuff with Jupiter and sequel alchemy and Trino, just so I can then teach a bunch of other people how to do the same thing. So I'm building out a bunch of example notebooks, which are of course also public get repo people can use. But we've, we hold office hours twice a week in the mornings. That's been a big thing. So people can just show up and share their screens and say, Hey, this is what I'm seeing. And that's like, I can figure out, you know, it's everything from that to like, this week, some people couldn't get their Jupiter up and running. And like, you know, I had to I had to like get on the cluster and reset their, you know, persistent volume claim. The entire the entire level is like, you know, nuts and bolts all the way up to all the way up to data science. We're trying to help people with. Yeah, thanks. Thanks, Eric. Thanks, everyone. Guys, we are out of time. It's really amazing how we can turn this huge, huge topic that is slightly overhyped these days into something tangible for us and the open source community to actually do something about it. So amazing stuff for the folks here in the room. We're going to stay here on the chat for a while. So if you still have some additional questions, you can ask them in the chat. Otherwise, we're ending the recording here and we can even talk off record in the chat now. Again, thanks a lot. It was great to have you here and everyone have a great rest of the day. Thanks. Bye.