 Well, Charlotte, you know, thank you so much for joining us in our little car interview or a little tour of Amsterdam in our fancy Audi, which we are really thankful we were able to borrow. It's a lot of fun. Do you want to kind of introduce yourself to our guests? Yeah, sure. Thanks for having me. This is really fun. I'm Charlotte Dunlap. I am a research director at Global Data and I've been there for some time now actually covering application platforms for for about a decade. And so I'm really following a lot of developer technology over these years, which, you know, started out being about middleware and then with the cloud, advent of the cloud that moved into Paz and yeah, and then all kinds of fun, cool emerging technologies that have come off of that, of course. And so, of course, well, so Kubernetes eventually a much more kind of sophisticated concept than Paz in some ways. Yes, that's right, particularly that would bring in the operational side of things. So, you know, developers figured out pretty early how to pull a credit card out and order up AWS services initially, because they were under the gun to create applications much faster than they used to. They used to have, say, nine months to procure hardware and software, you know, in the old days, yeah, with monolithic apps. And now suddenly they had to do this, you know, with digital transformations in a matter of weeks or even days. So, so that switched all that up. And that went along fine for a while until operations teams wanted to get their arms around what they were doing, what they were running, what they were running. Yeah. And so, so you're right. So, you know, enter Kubernetes and right around, well, and then, of course, I don't need to mention the last few years have been really tough between the pandemic and geopolitical issues and shaky economy. And that's kind of exacerbated what enterprises are dealing with and what operations teams are up against. And so, again, everybody's been looking to Kubernetes to solve all of this. And it has addressed a lot of the pain points by these teams in, you know, pain points such as more demands on IT and explosion of data and increased endpoints by offering a cloud infrastructure that as a scalable and elastic platform and also offers cognitive services and other disruptive services. And that's all good. But, but Kubernetes definitely has its problems as well, which, which is has to do with enterprises, not always having the kind of solutions that abstract the complexities around configuration of these new application architectures. And, and that's, that's tricky. And so, you know, especially in distributed apps, right. So that's kind of laying a little bit of the groundwork for you, LinkedIn, on, you know, the pros and cons of Kubernetes, which brings us to where we are now, which is unfortunately we've got this big, huge technology skills gap globally. And so we're, we're at this point where vendors are looking to how they can help their customers, because I think vendors are seeing that enterprises recognize that it's, it's better for them if they can tap existing talent, then companies can leverage employees, institutional knowledge, and that's very valuable to them to be able to do that. So it just feels like every other day, like to be honest, right? That's actually one of my pet peeves about the industry is that, you know, a lot of the times with recruiting for software engineers, you know, people are looking for, you know, a Python person with, you know, seven years of experience and, you know, two years with this particular module and stuff. And the problem with that is that I think it makes the gap worse, because like, you know, if somebody has, you know, an obstetrics, you know, software development design background, you know, it doesn't really kind of matter what language it was that they, you know, kind of grew up in, if you can give them a bit of time to learn whatever this new thing is, you know, you can get, you know, somebody who has, you know, someone who has 10 years of programming, programming experience, it doesn't actually really matter that much as far as the language is concerned, as long as you can give them some time to kind of, you know, you need to learn that new toolbox, actually, with my students when I, when I talk about like Python and what Python modules are and stuff, you know, I'm like, you know, imagine that you're like a carpenter, right? Or like another kind of construction work or whatever. And you have a whole set of tools in your garage, but you don't go to every job with all the tools, right? You take a couple of toolboxes because you know this job is going to need it. And that's what we do with Python modules, right? Is that you need, you know, you need to have this set of tools so you don't build everything from scratch. But they're, you know, they're, they're kind of similar. Like if I borrow your hammer, it's still a hammer, right? You know, it takes me a little while maybe to know that the details of the weight or whatever of your hammer, but you know, they're, they're all kind of at the end of the day. It's like, I know how to program. I can figure out this particular language in this particular set of libraries. I just might need a little bit of time to come up to speed. And so what I'm, what I'm hoping is that some of the things that you're talking about are going to actually lead us to allow for kind of software engineers to kind of migrate more into, you know, kind of coming from, you know, their job and developers and going over to Python or, you know, vice versa. And can we actually address some of that skills gap, right? Or the, the people gap really by allowing more latitude in movement. For sure. And then, okay, I know the big joke always is how, how many minutes does it take someone here at KubeCon to bring up chat GPT? But with what you just were talking about, then, then even if you're not familiar with a language, you know, leveraging that to, for converting the scripts, right? Especially into languages that you're, you're less familiar with. So yeah, I agree with you. You've got that kind of basic information that certainly does apply in various ways. Right, right. And chat GPT in some ways is, is really just kind of, you know, broader stack overflow, right? I mean, like, you know, it's funny, I used to, and I talk about this a bunch on the show, but it's like, you know, when I was consulting, I'd have grand plans of what I was going to build on the plane, right? And then I realized when I'm on the plane, I can't actually code without the internet. Like, I, you know, I need to be able to like, so I'm like, oh wait, how do I do this in this particular language? I can't remember, so I have to go look it up. You know, and it's super weird, you know, that, you know, I just, I have a really hard time programming without the internet. And I think, you know, that's not necessarily a bad thing. Right. No, and this is what I understand too. I have been talking with a lot of developers or wannabe developers, meaning, especially in particular, guys that are cis admins that want to reinvent themselves. And frankly, I'm underplaying them. They're really moving into coding a lot, but you're right. They tell me, they Google a lot of how to build these apps and they spend hours doing it. And so I, I tell people, I don't really care about chat GPT. And as far as kids that are using it to write their papers, you probably do because you're a professor. But what I'm hearing from these, these folks that, you know, new coders is that it saves them so much time with providing this baseline coding that they would have spent hours on the weekend working on. And then like I said, and helping them converse script. Well, and this is one of the big challenges, a lot of the times with like new software projects or whatever, is like the setup time where well before you're like actually adding any value is high effort. And just in, even for like professional developers, just infrequent enough that you have to like relearn it every time. You know, so it's like, you know, if I set up a web server or something, I have to go and dig up the docs on Apache to find out exactly how I set that up every time because the gap between when I do it, you know, one time and the next project is just long enough for me to forget it all. Sure. And, you know, because even though I've done it, I don't know, at least hundreds, if not close to a thousand times, you know, I still kind of have to go like go re-up on the docs and just be like, okay, and plus the fact that it may have changed. It might have gotten better in some way or maybe they changed it a little bit in another way. So I'd have to relearn it. And so I think tools like where you're kind of, you know, and we've been chasing this grail for a long time, you know, I've personally written three production level code generator applications, right? You know, one way back in the day based on UML, you know, where you drew out the architecture you wanted and it would generate you a baseline application and then you could actually turn it into what you needed. So I think it's a huge benefit, the thing that as teachers and, you know, other kinds of roles like that, what we need to figure out is like, okay, now how do we incorporate it? Actually Boston University, the part I'm in that's called the Faculty of Computing and Data Sciences and we just released our policy on it, which is that you can use chat GPT, but you have to cite it. And so if you're doing that and taking that approach, okay, you know. And I swear one of the things that I try, so we do these classes that are, we call them practicum or practica classes, where we're actually having student teams work on projects for third parties. And they walk into the project and they get assigned to the project and it's no-jazz project. And the student team will say, we don't have anybody on the team that knows no-jazz. And I'm like, okay, Google it. You know, like this, what you have to learn to be an industry professional is how to learn about these things quickly and identify, okay, you know, the things like, I know these things about Python. So what I need to do is kind of know those same things in no-jazz or JavaScript or whatever. And then what I remember, what's funny is that, because I've done so many different languages, the things I remember about a language are the things that are weird. And so that way, because those are the things that are very difficult to look up, right? Oh, sure. You know, but me looking up how to do a for loop in Python is, you know, half a second. But knowing, you know, how do you do, you know, threading models in Python that involve, you know, long-time connections, that's using Twisted, which is an unusual framework, but has a very important use case. But that I can tell you off the top of my head. Yeah, yeah. Interesting. Yeah. So that's great. No, that's real-life skills that they're learning there, including, as you say, how to just solve a problem that, you know, you can't just tell the client. Yeah, no, no, we're not doing this one. Yeah, that's interesting. Yeah. Yeah, yeah. So when I'm, you know, back to these vendor training programs again, it just feels like I'm hearing about a new one now every other day, because like I said, how relevant they are that these cloud and platform providers are realizing how much their customers are needed in. And, you know, Red Hat has such a cool example with Ford Motor Company. You're right, right. And basically what Red Hat's done for them is just gone in, you know, on site for them offering a free self-service training. So again, they can do that upskilling and reskilling. And we see training in those two areas. Upskilling is pretty obvious. It's expanding your knowledge of new technology. Reskilling is about looking at employees that have the aptitude for moving into new areas. So say you're an accountant, just an example, you know, your accountant, maybe you could become... Data scientist? Well, at Data Science maybe. There's a lot of overlap there. Okay, actually I was going to say maybe like a security admin. Oh, right, yeah. And then say you're... Here, let's see what's another example. A graphic artist, maybe you might become a UX designer. Just kind of leveraging some of your core talents. But anyway, that's kind of what we see is reskilling. And so anyway, just again, to use Red Hat as an example, and I'll mention other vendors, but I mean, I know it's got 17, they call them training paths, I believe. Oh, the learning paths? I'm sorry, the learning paths, very good. I spoke with a Red Hat guy yesterday who said they're adding another one. They just decided a cube vert. Does that sound familiar to you? Yeah, I can't remember when it launches exactly, but it's VEM workloads on Kubernetes. Right, so one of the things I always like to mention about that is why do you want to do that? Well, because you maybe aren't prepared or aren't interested in moving a particular application to being quote, unquote, like cloud native, but you can get that kind of single pane of glass where everything is being managed by the same environment by kind of just bringing that VM over. Now, there's also use cases where you really do need a virtual machine because you want some specialties of the kernel or something that you can't get with containerization. But those are pretty unusual, and but there are definitely common use cases where it's just not valuable to anybody to actually convert this particular application to being cloud native because it does what it does and it doesn't fine. But if you can bring it in, then at least you can keep track of it in the same way and you can start to use some of the really nice points of an orchestration layer where if you want to do auto scaling and things like that, you can still deal with that, but you don't necessarily need to make a full move to a cloud native or containerized application. Okay, I see. Okay, hey, you just made me think of something else when you said something. To me too, I feel like we've gotten to this place and time because of some of these cool emerging technologies I was mentioning that I've been following for some years. Low code platforms is an obvious one. Intelligent automation, RPA, robotic process automation. I feel like these kind of technologies helped usher in where we're at today as far as these training programs because to me, they broadened the audience that is able to have access to this high productivity technologies that usually were once just reserved for those elite data scientists, those guys. Or like 4GL kind of software, right? And they're not very common anymore, but like PowerBase, was that one of them? Power something or another? And so they had these very specialized language models and all this stuff, but they made you a much more performant developer at the expense of universality, right? Yes, sure. There's always that trade-off. Right, so we're starting to find a much, at least I think, a much better happy medium where we can kind of get a lot of those advantages without as bad trade-offs as the old 4GL programming languages. PowerBase, that's what it's called. Oh, got it. Okay, okay. Yeah, and speaking of that, I mean, so some of these automation solutions like Microsoft Power Platform. Oh, like Power BI? Yes, that's right. IBM Cloud Packs for Business Automation. Well, this is my big... Salesforce Flow. Again, these are platforms that have... They include technology that just used to be reserved for the savvy, the pro developers and data scientists and now they're really expanding to much broader non-coding. Right, right. So I'm sorry to interrupt you. Oh, no, no worries. I was just... You know, it's like one of the things that I'm a bit like where we kind of, you know, a lot of software, right? It's a lot about pendulum swings and kind of getting it in the middle. And, you know, and Paz is something I really quite miss but also understand didn't quite hit the right sweet spot because one of the things about those kind of tools, it's really nice as a developer to be just kind of like, okay, here's your entire platform. This is all the stuff you'll need. It's just right there. The problem for most of the Paz was that those platforms were very difficult to build and maintain. And so what I'm hoping is we'll see a next generation of those with some of these things or, you know, or they'll kind of be delivered in a different way. So like, you know, some of the things you're talking about but also things like the operators, those are huge for that because you can say, oh, vendor, what is the best way to deploy and operate your application so that I can use it in my application? Because I don't really care about how your application works. I care about the value add I'm putting that's going to consume some service of yours. That's so true. And that right there is what I always say has what has been the barrier of adoption of a lot of this technology that we're talking about that people have come out with but the difficulty and great. Oh, there's that museum, cool. Oh, yeah. Yeah. Yeah. I keep driving around in like, I'm not in some ways I'm not really seeing a lot of this stuff, right? Because I'm paying attention to the conversation and the bikes and the cars and stuff. Well, you don't have to, you're not driving, right? So, you know, it's nice when you point out stuff because we do have the roof camera. So our audience should be able to see it too. So yeah, it's kind of a lot of fun. Yeah, Jan Vermeer is as an exhibit here. Oh, nice. Girl with the pearl earring. But anyway, so that is the difficulty we've got, you know, you've got slick developers, even low-code, no-code developers that are creating apps but then they throw it over to operations and it's like, okay, what do I do now? How do I move this into production? And they can do it, but there's a lot of refactoring. There's a lot of configuration. A lot of the difficulty has not yet been abstracted to the level it should be. And that's the real tricky part. And again, that goes back to our little theme here of needing more professionals to, or again, rescaling people to be able to. Yeah, one of the things that we were talking about earlier or like with another interview, we're talking about how when you, you know, it's so much better if you can actually have, you know, not necessarily the operators, like the operations team, but kind of someone who's in that neighborhood actually kind of reach into the development teams and work with the development teams with their first, you know, applications in those kinds of environments or, you know, or something like that. Like if you can kind of show them how to operate those development, you know, activities and then, and his point was this is the key, is then automate it as much as humanly possible. Right, absolutely. The benefits then, you know, you have way less of that problem where when they're bringing it in, okay, now we actually want to deploy it to production. Now you don't have as many problems that, you know, the operations team, for example, has to deal with. For sure. And that makes sense. Yeah, so they, having like, in his point, was really that it's not so much about DevOps as much as like making sure that whether you have silos or don't have silos, but his argument was silos are okay. But if you, it's more that make sure you're reaching across the lines. Yeah. That it's not, you know, it's not a hard division, you know. You're right, and silos are okay because guess what? That's the reality. Mm-hmm. And I don't know, oh man, probably be hated for saying this, but DevSecOps, I don't know, man, that's just not happened yet. So because great, you know, we want to tell developers just put aside your day job and learn security now. Right, I mean. It would be great, but it's just not practical. It doesn't seem like it's flying and maybe it shouldn't. Maybe this is stuff that needs to just be snapped on or built on. Or it's, yeah, or I mean, even if it is like, you know, security can be difficult, right? Because it does need to be integrated at kind of a deep level. Sure. But it doesn't necessarily mean it needs to be the same human. You know, it's like, one of the problems I have with this whole concept of, you know, full stack developers is, you know, I remember the days when people were full stack developers because we had to be, because there was no striation between skill sets. And so I just had to do the, you know, massively ugly buttons on my websites, because there wasn't anybody else. So you mean like as in front end and back end? Right. So like, yeah, exactly. So, so what's funny is like, no, we have different skills. What we need to, I think we need to do a better job is figuring out how we work better together without kind of the concept of it's not really a handoff, which is kind of comes from a waterfall model where, you know, if we, this is why I often complain that, you know, what we do in software, the software world is not engineering. It's much more like writing a book. Because what we want to do is we want to have a relationship between the UX person. So like, for example, I had a student team really recently where the UX person had done a design and then the engineer came along and said, you know, hey, this framework that we chose is probably not going to be able to meet that design. And I'm like, okay, did you talk to the UX person about changing the design so that it can work with the framework we've already chosen? And he was like, no, is that an option? I'm like, yes. That's the point, right? Is that, you know, we don't know without talking to the UX person. We have to play nicely together. Right, whether, whether it, like, is there some really specific reason why they chose those particular icons? If there was, then we'll go change the framework. But if there wasn't, maybe they can just change the UI, you know? And I think that's the problem. It's not handoffs. It's, we gotta have conversations. We gotta work together and be like teams. It's much more the agile scrum thing I think than, than kind of true DevOps models. Yeah, that's really interesting. There was an interesting announcement that came out a year ago at one of the big conferences, AWS Studio, Amplify Studio. I remember both, both those names, but I'm not sure which one you're referring to. Okay, it was, it would be Amplify and what it, I'm pretty sure I'm remembering it, but the point is, for the first time, it was saying to front-end developers, hey, guess what? We are now giving you access to back-end development. And I looked at that, like, what? That's unheard of, but pretty amazing. Because again, that's the really tricky. Well, there's another product that we've been experimenting with at school, which is, I think it's for Figma, but it's basically a plug-in to Figma which will, based on a UI, generate an application. And again, kind of it's more like, and this is, like I said, I brought a bunch of code generators and one of the things that we learned was that it's really good to generate like the base, right? It doesn't have to be right. It can be the 80-20 rule where it's just like the right architecture, exactly. Which then you can modify to be the right thing. Right. You know? So I think it's super interesting and I'm curious, this is why, you know, it's kind of interesting to be kind of on the university side because, you know, we have researchers who are way far ahead of where we are and we have all these, you know, we have a lot of students who want to get into industry, right? So they're doing stuff that's very common day to day and so I kind of see a lot of the whole gamut. But yeah, for sure, it's great. Yeah. Really interesting. Yeah. So, let's see. What else do we want to talk about? So what would you say has been the highlight of your trip to, you know, Amsterdam, KubeCon this time? Oh boy. Yeah. For sure. I mean, what can I say? This is my first time in Europe. I always go to KubeCon. Oh, in the US? Yeah. Yes. But it's great. It's multi-vendors here. So you get access to everybody. It's not just a one-vendor conference. Just the different discussions with everybody, just in the lunch room and some of the user panels and things I've been seeing and keynotes. I don't always understand all the different project updates, but it's interesting. Oh, definitely new, new open source technologies. And you always just walk away with a handful of these things. You know, I mean, backstage it's cool getting to know that better, that technology, and where that might go. Backstage for the last, so I do the show monthly. And for the last, like, five episodes, so the last five months or something, it's come up just kind of organically. I thought I was being original. In every one of them. No, it's just crazy. Is it, clearly, it is meeting some need? Yes, I do need. Excuse me, that people are starting to notice that they have a need for. And so that's been really interesting. And I've been particularly noticing it because I hadn't heard of it until I went to Detroit for KubeCon North America and I came back and I want to implement it at the school for the projects we do because we have, you know, we have very high turnover, right? Every semester we turn over 300 employees. You know, because of all of our student teams go away and then we get another crop, right? But the projects live longer than that. And so we have a very hard time of transitioning from one group to the next because not only do you have typical transition problems, but the students might just be gone, right? They might have graduated or they're just not available because they're doing 37 other classes or whatever. So we really want to implement something like backstage and we try to do good hand-off documentation and stuff like that. But even just keeping track of where all the pieces are is difficult. And what you're describing isn't just unique for a university either. I think that would resonate with a lot of Red Hat or general enterprises and customers is the turnover and wait, when Joe leaves, this place will fall. Right, right. Nobody knows where... I did more than one consulting gig where what we were trying to do was build the thing to make sure that when Joe left that company didn't fall over. You know, normally a lot of spreadsheets getting converted into like an application was, yeah, mind-blowing. But yeah, so like I said, we have this very pronounced turnover problem that I think is significant or hopefully significantly worse than most enterprises. And so maybe we'll get to a solution that is actually consumed by other people. Let's see what happens. Yeah, that's great. But definitely the dominating thing, no big shocker is Gen AI and ChatGP too. Yeah, yeah. But you know, we're all enjoying getting to know it. It's definitely on my agenda of being included in a lot of my reports and doing a lot more coverage in that area. And so I'm enjoying that. I like getting everybody's take. Yeah, yeah. It's really interesting. I mean, it's funny like I've been... I've also been like kind of looking at a lot of kind of teacher positions on it. And one that I thought was really interesting was, you know, ChatGPT like for writing assignment, you know, in theory can just generate it. But what this, I think it was an English teacher found was that, you know, it's like, especially in the middle, it's kind of like, and so what she started doing or what she's experimenting with is that she's now giving all of the good parts from the ChatGPT generated answer to a written, you know, writing prompt and removing the bad parts and then asking the students, okay, now your class, the work is to fill in the parts that ChatGPT couldn't do well. So it's kind of like, it's kind of like the programming here's your boilerplate, now make it into a compelling argument or, you know, I think it's like a persuasive paper. You know what, I like that because it's also realistic and I'm not trying to say plagiarism is going to be okay one day, but I mean, my background is journalism and analysts. And being a really good researcher is important. So, yeah, in that capacity. The other thing I was going to say about it is it's really interesting. I'm hearing an overall theme by a lot of the vendors and executives speaking about, okay, Gen AI is going to take years. This is going to take a long time. We need to be careful. It's got to be explainability and I was listening to another panel today saying it, and as I did, I got this text from my 26-year-old son that said, Mom, look, ChatGPT is now on Snapchat. Is this guy saying this is going to take years to become part of our products? No, it's not. I know they have an agenda that they want to be the ones to be providing it, but you've got to... You have to move faster than that. You're going to have to move faster than that and just recognize that in some capacity, it is here today. But be the thought leaders on how it's going to be implemented, the bits that are going to be implemented today and then the bits that we want to be careful and move more slowly on. A vendor certainly can serve that role of being the thought leaders on how we go about that. And that's why we put out that policy as the faculty, right, is that we said we're an unusual unit. But because we need to embrace this now and we need to embrace it on our terms and then we can embrace it and say these are the parts we understand and these are the parts we don't know yet, but here's our current thinking and then we can evolve that over time. But if we kind of pretend like, what was in Pandora's box? I can't remember now, but if we try to pretend what Pandora's box wasn't open, we're going to be in trouble. Absolutely. No, that's a very good way to go about it. And that's a realistic attitude too. That's probably because you work with... Students, you work with all kinds of... Well, in our particular unit, like we focus on... So we do a data science major, but the major kind of has two tracks within it. One is you want to become a PhD in data science and you want to advance the next piece of machine learning. But there's another track, which then that's the typical academic track, right? When you go to university, what you're really going there in a lot of ways for is how to be like an academic, how to advance that field, whatever it is. And then... But we have... We've introduced another track, which is the practitioner track. And so the baseline, a lot of the baseline stuff is the same, but then your kind of like half stone type work and the major... You know, your later year work is more on the consumption or usage and proper usage of things like TensorFlow or whatever, rather than building the next TensorFlow. Like the practical uses, in other words, you mean? Okay. Just in case the audience doesn't know, global data is an analyst firm. I'm meant to see that. Yeah, and that pretty good size too, as I recall. It's good size. We are global data is based in London. We are... We do data analysis. We consist of a number of research and analyst firms that do a vast array of materials, data for our clients, for our customers, from surveys to thematic, big, massive, I call them monster reports on topics, to event reports, very timely, very, very quick. That just helps give enterprises some direction on where the market's going and where their opportunities are. So not everyone has to keep up with everything all the time. That's right. That's what we do. Yes, exactly. Well, thank you so much. Really appreciate it. Sorry about the slight confusion about where we're going to end our drive. But at least you got a nice little tour of Amsterdam. I loved it. Weather got much better than it was this morning, which is nice. But yeah, thank you. Okay, it's my turn to drive now. Yeah, exactly.