KubeCon + CloudNativeCon Europe 2024 | Máirín Duffy | Red Hat

Clips
News
More from KubeCon + CloudNativeCon Europe 2024

Máirín Duffy

Distinguished Engineer

Red Hat

Red Hat looks to improve the user experience for Kubernetes and AI

Kubernetes and artificial intelligence may be changing the industry, but having an intuitive user experience for developers interfacing with these technologies is key to their implementation.“As a [user experience] practitioner, the way that I look at it is … AI is just a technology. If you really want to make compelling software that changes the world, you have to think about, well what problems are the users trying to solve?” said Máirín Duffy (pictured), senior principal UX engineer at Red Hat Inc

play_circle_outline Exploring the Significance of User Feedback and Interacting with Open Source Communities in Kubernetes

play_circle_outline Overview of the Podman desktop tool and its role in developers' understanding and usage of Kubernetes

play_circle_outline The role of AI in improving user experience and practical applications in various fields

Info

Máirín Duffy | Red Hat

Máirín Duffy

Distinguished Engineer Red Hat

search

(soft music)

>> Good morning Cloud community and welcome back to Paris, France. We're here at Kubecon Cloudnativecon, CNCFs, flagship European event. I'm joined by my fabulous co-host, Rob Strechay, good morning, Rob.

>> Good morning again.

>> Yes, good morning. What a beautiful time to be in Paris.

>> It is, Paris in the spring.

>> Paris in the spring, and it's got a great user experience here in the spring. It smells good. It's fabulous. Speaking of user experience, our next guest, Mo, welcome to the show.

>> Thank you. Very happy to be here, bonjour

>> Ex boss at Red Hat, casually. How's the show going for you so far?

>> It's going great. I did a talk yesterday afternoon on the developer user experience in Kubernetes, and I was actually shocked, the room was full. I didn't know if anybody would be interested.

>> I think a lot of people are interested, and we've been talking about how Kubernetes is having its Linux moment. Certainly everyone's excited about it here. Congratulations on your talk.

>> Thanks.

>> Yeah. I think one of the things that a lot of people don't get into, because again, and you kind of mentioned this earlier when we were, you know, chatting beforehand, is the fact that how you actually interact with the open source and how you develop that interaction, it has to be different from an open source perspective. What kind of tips or tricks were you sharing with the folks yesterday?

>> So what we did is we basically went out and we talked to users like, it's not that-

>> Rocket science.

>> Yeah, exactly. So we interviewed 11 different developers and platform engineers that support developers. And we kind of, honestly, we booked them for half an hour and most of them, they were so lit up that it went to 45 minutes into an hour and some of them, and we did transcripts of all the interviews. Reviewed them, picked out all the interesting points and little nuggets of oh wow, actually that's insightful. And we did just a, you know, a minimal sticky note exercise to cluster the findings. And then we basically presented the findings to everybody at the conference. So there were eight major buckets, three of which involved YAML of just the feedback that these developers surfaced that, listen, if you want Kubernetes to be a platform that we're not intimidated by, that we feel really engaged and enabled on, these are the things you got to address.

>> Well, what was the findings on YAML? Because I always laugh when you get to a demo, every demo that they do from the stage, there's always YAML. You know, okay, we're going to, you know, edit the YAML file and go and what was the findings really around that?

>> So one of the first ones that we talked about was the fact that YAML indentation and spacing is a little fussy. And even if it passes the linting and it checks out, you can still have an indentation that's off that can change the configuration that goes out. We had suggested maybe having some tooling that would give you a preview of what would happen before you commit like a dry run and have that surface up to the developers.

>> I think that is one of those things that I know really is frustrating to people, especially. And I think from a user experience and getting started with any of these projects, it's intimidating to a lot of them. Was that a part of the topic that was happening as well?

>> That surfaced a lot. And it was really funny too 'cause when we said on stage, hey, you know, developers, like bare minimum baseline feedback is Kubernetes is intimidating. And everybody in the audience was like, yep. (light laughter)

>> I can imagine, I mean, decreasing complexity has been the mission of a lot of these projects here. Everyone wants an easy button when it comes to Kubernetes AI, the whole shebang. People also want an intuitive user experience. If it's not fun to use, they won't come.

>> Exactly.

>> So how is UX different in the open source community than traditional models?

>> So the way that the open source community started out really is it's technical people with programming chops sort of scratching their own little itches. And a lot of open source communities have a long tail where it's like almost 50% of the community is just sort of drive by contributors or part-time contributors. The problem is that you can't practice UX in sort of a drive by fashion. You necessarily have to be very deep. So it's a little more difficult to get UX contributors to a project. You really have to nurture them and invest in them and try to keep them around for the long haul. So that's why we just have less practitioners doing it in the community. And I feel that, you know, our talk was not the only UX centric talk here at Kubecon this year. It's really good to see that it's getting picked up as a topic here. And it's that sort of nurturing of that discipline and practice in the community, that makes it a more appealing place for UX designers to come and do that deep dive and really get involved.

>> In fact, I saw a couple of talks during backstage con around some UX aspects of that and making it simpler to actually do things within there. And some of that was leveraging web assembly and to go and actually do some of the configurations and things of that nature. Do you see a lot more of that where people are leaning in and there's projects helping other projects from a UX perspective, 'cause that would seem like it would make too much sense, but.

>> It does happen. And I can say like I have a long history with the Fedora project and the Fedora design team. And it's really funny because other distros and other upstreams that we have associations with were coming to the Fedora design team for help with UX stuff. Just because we had an established team that people knew about. So we ended up kind of rebranding ourselves as the community design team. 'Cause we had set up sort of a nice little community where designers felt safe.

>> Yeah.

>> You know what I mean? 'Cause they can be overwhelmed by, you know, they're technical but they're not knee deep in the code. So there's sort of a little bit of a great confidence building, right?

>> Yeah, and fostering that community, how nice. I bet that was very well received.

>> Yeah.

>> Yeah. Let's talk about Podman desktop.

>> Sure. So Podman desktop was the reason that me and my colleague, Connor, did the UX research study on developers with Kubernetes. We've been slowly building out some tooling to understand Kubernetes right in the UI because the sort of mission of Podman desktop, you have developers who start with containers, they have their app on containers and maybe they're using a compose file or something like that. Then it got thrown over the wall to ops and they're rewriting the compose into YAML for Kubernetes. Then something happens in production and it's like the developer doesn't actually know, right, intimately how it's being deployed and how it's being run. So the thought is, let's shift that left. Let's get the developers comfortable with the cube YAML and have them working with that from the start. And even more have them using local versions of Kubernetes clusters on their machine. Like we have kind support, we have mini cube support. You can just click a button and it will set it up for you on your local machine. So you can start with your app in the pod, test it out, then push it to your local cluster, test it out and sort of do that inner dev loop and then when you're ready you can push it out to a dev cluster. And that's all from one tool.

>> Yeah, it would seem like that helps people get up and running that much faster and can help 'cause I think to your discussion earlier and what we've talked about, a lot of it is, you know, things have been solved in Kubernetes. Like from an infrastructure perspective, a lot of it, I know we're 10 years in, but still be able to do the cloud native design and do it efficiently is also a big thing. Does Podman help you understand, okay, this is how I built it and it's running efficiently and you understand the resources and things of that nature as well?

>> Yes. And we're building out that tooling. So we just released Podman desktop 1.8 this week and it has support for deployment services, ingress routes. So now you can look at those Kubernetes objects right from your local cluster and get more of a feeling of what's going on, rather than having to push to some remote cluster and then have it fall over and-

>> Right. And you're getting that real time feedback.

>> Exactly.

>> Which changes the whole experience.

>> Right, and again, it's also sort of tempting developers to do things the cloud native way.

>> So it's a win-win.

>> Yes.

>> In a lot of different ways. I love that. Let's talk about AI and UX, 'cause I feel like, I mean we're talking about AI a lot this week. Everyone definitely abuzz, first European Kubecon since chat GPT was announced in the talk tracks could parallel that. We were in Amsterdam just after that last year. How is UX guiding some of the development in AI?

>> Right, so one of the frustrations I hear from just technologists about AI is, oh, it's another hype. You know, last time it was a blockchain, now it's AI. But as a UX practitioner, the way that I look at it is blockchain is just a technology. AI is just a technology. If you really want to make compelling software that changes the world, you have to think about, well what problems are the users trying to solve? And then you look at blockchain or you look at AI, it's just a tool in your toolkit and see does that tool actually help solve that problem? Okay, we'll apply it here. Does it solve this problem? No. Okay, well you're excited by AI, but if it doesn't actually solve the problem, don't do it, right.

>> Yeah. And in fact, it made it to the keynote a little discussion about UX and AI, which was the fact that, and this was yesterday where again, a chat interface may not be the right way to actually interact with the AI, depending on what the user's trying and what the use case you're trying to solve for. It may be something different than that, that you're trying to do with that. Are you seeing that a lot where people are saying, Hey, great chat GBT and prompts and prompt engineering, but that's not how people are going to actually interact with that.

>> Yeah, the approach that I prefer is sort of think of it as your application already has a set of functionality, it already has a user base. How can I infuse that preexisting app with AI to just supercharge it? So, you know, for example, just simple things like I have an application that has users from different countries. Is there a way that I could somehow automate, hey, if you're really stuck, can I use an AI tool to auto translate things into another language if I have a user base in that country? I mean, just simple ways and it should be transparent to the user. Like they shouldn't have to go to a prompt and ask for things.

>> It's a great point.

>> Just think of it, it's just like another API or a library you're calling. It's not like any kind of special thing, right.

>> Agreed. I think that to your point, it's another technology underneath the hood that has some really neat capabilities, but like being able to find logic and find things that are missing and when you're going in. So like alerting is one of those things. I think it's going to completely change observability, for instance, I think the observability-

>> Absolutely agree. Absolutely agree.

>> You have to go from here and tie that to this and tie that to this, that kind logic where you have to do, you know, mental math to kind of figure out vectoring would seem like a perfect, you know, place to go.

>> And this is an interesting parallel to like AI and say the user experience practitioners work, 'cause a lot of times, and this is something we said in the talk is Kubernetes is very chatty, gives you a fire hose of metadata and information about things. And when you're feeling intimidated and you haven't quite learned the system yet, when you have this fire hose, you have to manually for yourself sort out well what actually matters and what is just noise, right?

>> What data is relevant.

>> And that's the job of a UX designer is to work with the developers and like, okay, this is all the data we have on the system, what should we present to the users at what time? It's almost theatrical. Like at what stage of their process do we surface this data? So what is kind of promising about AI is does it have the potential, like I have this big log file. There's no way a UX designer can go through that and say, well, what should be surfaced? So could the AI process that in a way to surface just the lines of the log based on the errors that were detected? And again, none of this has to be with like a big blinking sign. This is AI, it's just the user's trying to do a thing just gently nudge what should be called to their attention. And that's it. Like, just infuse the app, don't make it all about AI.

>> Well, I love that point though 'cause it is, I mean, UX is all about making things feel intuitive and natural and to your point, I mean it's going to be like Intel inside. It's going to be the AI inside things eventually and we won't be talking about it. And I do-

>> Just another tech.

>> Yeah, well it's just going to be just another tool really that makes our lives a little bit easier. Speaking of that hype curve, and this is more taking off your red hat for a second, what excites you about where we're at as a person,

>> But where- >> In terms with AI, like are there any applications or any use cases where you are personally really excited about it?

>> Nothing to do with Kubernetes, nothing to do with anything. I am really into family history research and there are so many documents that have been scanned and they are in old school script. And I think that there's a lot of potential for AI to take that old school script, make it to searchable text so that people can find these family history records. I recently and it was weird, 'cause I knew, I used to live in Cambridge, Massachusetts and they use the standard old school English style of real estate. Like you have to have a volume that's the index. And then you find the index book and then that book has another. So it's like you're running around this library trying to find stuff. Because I knew that system, I was able to go through the, they have the equivalent online on this website family search where you find the PDFs and they mirror that book system. And I found like the license for, well it was a lease for my ancestors in Ireland from like the 18 hundreds using that system.

>> That's so cool.

>> But the thing is, is I had that special knowledge of understanding how to use the different things. And I had to manually read the script. How many people, like their ancestors are just there waiting for them in these, you know, these scanned documents and they have no clue. So that's what I'm really excited about. And that's a real human connection, right.

>> A hundred percent.

>> That the technology can make.

>> Yeah and it's a UX problem too.

>> It really is.

>> It's like how many, I think that is a great way. Great.

>> A perfect example. Well, and that's just it. I mean we try and dispel the doomer hype on the show a bit and when people are worried about AI or things like that, they forget that it could do something like help them remember their legacy.

>> Exactly.

>> I mean, it's so powerful. I'm so glad I asked you that question. That was a wonderfully brilliant answer. That was awesome. Putting back your red hat on, so obviously big event, we love asking this question. What do you hope you can say at the next Kubecon, that you can't quite say yet?

>> I personally have concerns about AI, the way that is structured right now with large models, you have to have a lot of resources. You have to be at a certain size, you have to have a lot of data to be able to make a model. And to me that seems a little like the power relationship there. You know, it feels like the old days of like when we all had to use certain operating systems that maybe weren't as nice as Linux. I'm sorry, I'm Red Hat. I have to say that, where people almost felt like I'm at the mercy of this company, I don't have any way to give back. Whereas with Linux it was like, you know what? It's made by a community of people. If something really bothers you and you want to make a change, you're empowered to do that. And I would love to see something like that for AI where the end users who it is being foisted upon, have some say in some amount of control and power in that situation, 'cause right now it seems very top down.

>> I think that's beautifully stated and hopefully we can say that in Salt Lake City or in London at the next Kubecon.

>> I hope so.

>> Mo, you've been an absolute gem. Thank you so much for joining us.

>> Sure. It's been my pleasure chatting with you guys.

>> Yeah, I feel smarter already. It's a great start of the day. Rob, I hope you feel-

>> I do. I love it. The user. It's about the user and I love it.

>> It is about the user.

>> I love it.

>> It's all about the community. It's all about you watching live at home to our three days of coverage here at Kubecon Cloudnativecon, CNCFs, flagship European event. My name's Savannah Peterson. You're watching The Cube, the leading source of enterprise tech news. (soft music)

Máirín Duffy | Red Hat

search

(soft music)

>> Good morning Cloud community and welcome back to Paris, France. We're here at Kubecon Cloudnativecon, CNCFs, flagship European event. I'm joined by my fabulous co-host, Rob Strechay, good morning, Rob.

>> Good morning again.

>> Yes, good morning. What a beautiful time to be in Paris.

>> It is, Paris in the spring.

>> Paris in the spring, and it's got a great user experience here in the spring. It smells good. It's fabulous. Speaking of user experience, our next guest, Mo, welcome to the show.

>> Thank you. Very happy to be here, bonjour

>> Ex boss at Red Hat, casually. How's the show going for you so far?

>> It's going great. I did a talk yesterday afternoon on the developer user experience in Kubernetes, and I was actually shocked, the room was full. I didn't know if anybody would be interested.

>> I think a lot of people are interested, and we've been talking about how Kubernetes is having its Linux moment. Certainly everyone's excited about it here. Congratulations on your talk.

>> Thanks.

>> Yeah. I think one of the things that a lot of people don't get into, because again, and you kind of mentioned this earlier when we were, you know, chatting beforehand, is the fact that how you actually interact with the open source and how you develop that interaction, it has to be different from an open source perspective. What kind of tips or tricks were you sharing with the folks yesterday?

>> So what we did is we basically went out and we talked to users like, it's not that-

>> Rocket science.

>> Yeah, exactly. So we interviewed 11 different developers and platform engineers that support developers. And we kind of, honestly, we booked them for half an hour and most of them, they were so lit up that it went to 45 minutes into an hour and some of them, and we did transcripts of all the interviews. Reviewed them, picked out all the interesting points and little nuggets of oh wow, actually that's insightful. And we did just a, you know, a minimal sticky note exercise to cluster the findings. And then we basically presented the findings to everybody at the conference. So there were eight major buckets, three of which involved YAML of just the feedback that these developers surfaced that, listen, if you want Kubernetes to be a platform that we're not intimidated by, that we feel really engaged and enabled on, these are the things you got to address.

>> Well, what was the findings on YAML? Because I always laugh when you get to a demo, every demo that they do from the stage, there's always YAML. You know, okay, we're going to, you know, edit the YAML file and go and what was the findings really around that?

>> So one of the first ones that we talked about was the fact that YAML indentation and spacing is a little fussy. And even if it passes the linting and it checks out, you can still have an indentation that's off that can change the configuration that goes out. We had suggested maybe having some tooling that would give you a preview of what would happen before you commit like a dry run and have that surface up to the developers.

>> I think that is one of those things that I know really is frustrating to people, especially. And I think from a user experience and getting started with any of these projects, it's intimidating to a lot of them. Was that a part of the topic that was happening as well?

>> That surfaced a lot. And it was really funny too 'cause when we said on stage, hey, you know, developers, like bare minimum baseline feedback is Kubernetes is intimidating. And everybody in the audience was like, yep. (light laughter)

>> I can imagine, I mean, decreasing complexity has been the mission of a lot of these projects here. Everyone wants an easy button when it comes to Kubernetes AI, the whole shebang. People also want an intuitive user experience. If it's not fun to use, they won't come.

>> Exactly.

>> So how is UX different in the open source community than traditional models?

>> So the way that the open source community started out really is it's technical people with programming chops sort of scratching their own little itches. And a lot of open source communities have a long tail where it's like almost 50% of the community is just sort of drive by contributors or part-time contributors. The problem is that you can't practice UX in sort of a drive by fashion. You necessarily have to be very deep. So it's a little more difficult to get UX contributors to a project. You really have to nurture them and invest in them and try to keep them around for the long haul. So that's why we just have less practitioners doing it in the community. And I feel that, you know, our talk was not the only UX centric talk here at Kubecon this year. It's really good to see that it's getting picked up as a topic here. And it's that sort of nurturing of that discipline and practice in the community, that makes it a more appealing place for UX designers to come and do that deep dive and really get involved.

>> In fact, I saw a couple of talks during backstage con around some UX aspects of that and making it simpler to actually do things within there. And some of that was leveraging web assembly and to go and actually do some of the configurations and things of that nature. Do you see a lot more of that where people are leaning in and there's projects helping other projects from a UX perspective, 'cause that would seem like it would make too much sense, but.

>> It does happen. And I can say like I have a long history with the Fedora project and the Fedora design team. And it's really funny because other distros and other upstreams that we have associations with were coming to the Fedora design team for help with UX stuff. Just because we had an established team that people knew about. So we ended up kind of rebranding ourselves as the community design team. 'Cause we had set up sort of a nice little community where designers felt safe.

>> Yeah.

>> You know what I mean? 'Cause they can be overwhelmed by, you know, they're technical but they're not knee deep in the code. So there's sort of a little bit of a great confidence building, right?

>> Yeah, and fostering that community, how nice. I bet that was very well received.

>> Yeah.

>> Yeah. Let's talk about Podman desktop.

>> Sure. So Podman desktop was the reason that me and my colleague, Connor, did the UX research study on developers with Kubernetes. We've been slowly building out some tooling to understand Kubernetes right in the UI because the sort of mission of Podman desktop, you have developers who start with containers, they have their app on containers and maybe they're using a compose file or something like that. Then it got thrown over the wall to ops and they're rewriting the compose into YAML for Kubernetes. Then something happens in production and it's like the developer doesn't actually know, right, intimately how it's being deployed and how it's being run. So the thought is, let's shift that left. Let's get the developers comfortable with the cube YAML and have them working with that from the start. And even more have them using local versions of Kubernetes clusters on their machine. Like we have kind support, we have mini cube support. You can just click a button and it will set it up for you on your local machine. So you can start with your app in the pod, test it out, then push it to your local cluster, test it out and sort of do that inner dev loop and then when you're ready you can push it out to a dev cluster. And that's all from one tool.

>> Yeah, it would seem like that helps people get up and running that much faster and can help 'cause I think to your discussion earlier and what we've talked about, a lot of it is, you know, things have been solved in Kubernetes. Like from an infrastructure perspective, a lot of it, I know we're 10 years in, but still be able to do the cloud native design and do it efficiently is also a big thing. Does Podman help you understand, okay, this is how I built it and it's running efficiently and you understand the resources and things of that nature as well?

>> Yes. And we're building out that tooling. So we just released Podman desktop 1.8 this week and it has support for deployment services, ingress routes. So now you can look at those Kubernetes objects right from your local cluster and get more of a feeling of what's going on, rather than having to push to some remote cluster and then have it fall over and-

>> Right. And you're getting that real time feedback.

>> Exactly.

>> Which changes the whole experience.

>> Right, and again, it's also sort of tempting developers to do things the cloud native way.

>> So it's a win-win.

>> Yes.

>> In a lot of different ways. I love that. Let's talk about AI and UX, 'cause I feel like, I mean we're talking about AI a lot this week. Everyone definitely abuzz, first European Kubecon since chat GPT was announced in the talk tracks could parallel that. We were in Amsterdam just after that last year. How is UX guiding some of the development in AI?

>> Right, so one of the frustrations I hear from just technologists about AI is, oh, it's another hype. You know, last time it was a blockchain, now it's AI. But as a UX practitioner, the way that I look at it is blockchain is just a technology. AI is just a technology. If you really want to make compelling software that changes the world, you have to think about, well what problems are the users trying to solve? And then you look at blockchain or you look at AI, it's just a tool in your toolkit and see does that tool actually help solve that problem? Okay, we'll apply it here. Does it solve this problem? No. Okay, well you're excited by AI, but if it doesn't actually solve the problem, don't do it, right.

>> Yeah. And in fact, it made it to the keynote a little discussion about UX and AI, which was the fact that, and this was yesterday where again, a chat interface may not be the right way to actually interact with the AI, depending on what the user's trying and what the use case you're trying to solve for. It may be something different than that, that you're trying to do with that. Are you seeing that a lot where people are saying, Hey, great chat GBT and prompts and prompt engineering, but that's not how people are going to actually interact with that.

>> Yeah, the approach that I prefer is sort of think of it as your application already has a set of functionality, it already has a user base. How can I infuse that preexisting app with AI to just supercharge it? So, you know, for example, just simple things like I have an application that has users from different countries. Is there a way that I could somehow automate, hey, if you're really stuck, can I use an AI tool to auto translate things into another language if I have a user base in that country? I mean, just simple ways and it should be transparent to the user. Like they shouldn't have to go to a prompt and ask for things.

>> It's a great point.

>> Just think of it, it's just like another API or a library you're calling. It's not like any kind of special thing, right.

>> Agreed. I think that to your point, it's another technology underneath the hood that has some really neat capabilities, but like being able to find logic and find things that are missing and when you're going in. So like alerting is one of those things. I think it's going to completely change observability, for instance, I think the observability-

>> Absolutely agree. Absolutely agree.

>> You have to go from here and tie that to this and tie that to this, that kind logic where you have to do, you know, mental math to kind of figure out vectoring would seem like a perfect, you know, place to go.

>> And this is an interesting parallel to like AI and say the user experience practitioners work, 'cause a lot of times, and this is something we said in the talk is Kubernetes is very chatty, gives you a fire hose of metadata and information about things. And when you're feeling intimidated and you haven't quite learned the system yet, when you have this fire hose, you have to manually for yourself sort out well what actually matters and what is just noise, right?

>> What data is relevant.

>> And that's the job of a UX designer is to work with the developers and like, okay, this is all the data we have on the system, what should we present to the users at what time? It's almost theatrical. Like at what stage of their process do we surface this data? So what is kind of promising about AI is does it have the potential, like I have this big log file. There's no way a UX designer can go through that and say, well, what should be surfaced? So could the AI process that in a way to surface just the lines of the log based on the errors that were detected? And again, none of this has to be with like a big blinking sign. This is AI, it's just the user's trying to do a thing just gently nudge what should be called to their attention. And that's it. Like, just infuse the app, don't make it all about AI.

>> Well, I love that point though 'cause it is, I mean, UX is all about making things feel intuitive and natural and to your point, I mean it's going to be like Intel inside. It's going to be the AI inside things eventually and we won't be talking about it. And I do-

>> Just another tech.

>> Yeah, well it's just going to be just another tool really that makes our lives a little bit easier. Speaking of that hype curve, and this is more taking off your red hat for a second, what excites you about where we're at as a person,

>> But where- >> In terms with AI, like are there any applications or any use cases where you are personally really excited about it?

>> Nothing to do with Kubernetes, nothing to do with anything. I am really into family history research and there are so many documents that have been scanned and they are in old school script. And I think that there's a lot of potential for AI to take that old school script, make it to searchable text so that people can find these family history records. I recently and it was weird, 'cause I knew, I used to live in Cambridge, Massachusetts and they use the standard old school English style of real estate. Like you have to have a volume that's the index. And then you find the index book and then that book has another. So it's like you're running around this library trying to find stuff. Because I knew that system, I was able to go through the, they have the equivalent online on this website family search where you find the PDFs and they mirror that book system. And I found like the license for, well it was a lease for my ancestors in Ireland from like the 18 hundreds using that system.

>> That's so cool.

>> But the thing is, is I had that special knowledge of understanding how to use the different things. And I had to manually read the script. How many people, like their ancestors are just there waiting for them in these, you know, these scanned documents and they have no clue. So that's what I'm really excited about. And that's a real human connection, right.

>> A hundred percent.

>> That the technology can make.

>> Yeah and it's a UX problem too.

>> It really is.

>> It's like how many, I think that is a great way. Great.

>> A perfect example. Well, and that's just it. I mean we try and dispel the doomer hype on the show a bit and when people are worried about AI or things like that, they forget that it could do something like help them remember their legacy.

>> Exactly.

>> I mean, it's so powerful. I'm so glad I asked you that question. That was a wonderfully brilliant answer. That was awesome. Putting back your red hat on, so obviously big event, we love asking this question. What do you hope you can say at the next Kubecon, that you can't quite say yet?

>> I personally have concerns about AI, the way that is structured right now with large models, you have to have a lot of resources. You have to be at a certain size, you have to have a lot of data to be able to make a model. And to me that seems a little like the power relationship there. You know, it feels like the old days of like when we all had to use certain operating systems that maybe weren't as nice as Linux. I'm sorry, I'm Red Hat. I have to say that, where people almost felt like I'm at the mercy of this company, I don't have any way to give back. Whereas with Linux it was like, you know what? It's made by a community of people. If something really bothers you and you want to make a change, you're empowered to do that. And I would love to see something like that for AI where the end users who it is being foisted upon, have some say in some amount of control and power in that situation, 'cause right now it seems very top down.

>> I think that's beautifully stated and hopefully we can say that in Salt Lake City or in London at the next Kubecon.

>> I hope so.

>> Mo, you've been an absolute gem. Thank you so much for joining us.

>> Sure. It's been my pleasure chatting with you guys.

>> Yeah, I feel smarter already. It's a great start of the day. Rob, I hope you feel-

>> I do. I love it. The user. It's about the user and I love it.

>> It is about the user.

>> I love it.

>> It's all about the community. It's all about you watching live at home to our three days of coverage here at Kubecon Cloudnativecon, CNCFs, flagship European event. My name's Savannah Peterson. You're watching The Cube, the leading source of enterprise tech news. (soft music)