How IBM Z Is Modernizing Mainframes with Skyla Loomis

Added: 2026-05-12

[00:00:00]and they absolutely need AI. Um, you know, if you're going to go and a lot of times they're very timesensitive. So, if you're going to go make a transaction, you want to catch that fraud at the moment that the transaction is happening, not 30 minutes or an hour later or even, you know, 30 seconds or 60 seconds later because by then you probably just had to let the transaction go because it took too long. And so, and once you've let fraud go, it's basically a loss, right? like it's just kind of a lot of banks look at it as the cost of doing business that you lose, you know, 8% of your revenue due to fraud in a year. Um, and that's a pretty high cost.

[00:00:35]And so, >> hey friends, you probably knew that text control as a powerful library for document editing and PDF generation. But did you also know that they're a strong supporter of the developer community?

[00:00:44]And it's part of their mission to build and support a strong community by being present, by listening to users, and by sharing knowledge at conferences across Europe and the United States. If you're heading to a conference soon, maybe check if text control will be there.

[00:00:59]Stop by and say hi. You'll find their full conference calendar at textcontrol.com.

[00:01:05]That's textcontrol.com.

[00:01:10]Hi, I'm Scott Hansselman. This is another episode of Hansel Minutes. Today I'm chatting with Skyla Lumis. She's a general manager of IBM Z software at IBM. How are you?

[00:01:19]>> I'm good. I'm good. Happy to be here.

[00:01:21]How are you?

[00:01:21]>> Yeah, thanks for hanging out. So, you've been at like IBM forever. This is really cool. I look at your LinkedIn, it is just filled with all kinds of experience. How do you get into like mainframe software?

[00:01:35]>> Yeah, you know, it's funny. You know, a lot of times you talk to someone who's been on the main frame and they've been in the main frame their entire career.

[00:01:41]Uh that's not my story.

[00:01:43]uh you know I I started out of college at IBM bounced around at you know different parts of the tech stack you know mostly on distributed in the beginning data some mobile analytics some cloud early days of IBM cloud and then you know I had this opportunity to work on the mainframe and it's you know such a storied platform uh within IBM and so much meaning for our clients that it was really an exciting opportunity after you work on new stuff and you're always trying to get new clients But it's also really refreshing to go to a platform where you have this rich history and these long-standing relationships with really important clients and you can make a difference you know for them and and in and what they do and what they do for their clients. So uh since then I've been about on the main frame about nine years or so and uh it's been been an exciting journey for me.

[00:02:32]>> So I'm going to ask a couple of ignorant questions because it's been a while since I've done any mainframe work and when I did it's it was invisible to me.

[00:02:39]I think one of the things that's interesting about the cloud and one of the things that's interesting about the mainframe is that people who work on them, whether they be a a young person just out of school who is working on the cloud or someone who's been around a minute who's working on the mainframe is in both instances we may have never actually seen the machine. I was talking to a whole class of young people recently that just started at Microsoft and they were basically you know they don't know why the cloud existed like what was the problem that was being solved for them has always existed and in the old days we used to visit the cloud we take them out to a data center and it's like Costco full of fridges >> like this is a dumb question but like have you seen a mainframe do you see them anymore they just in a cooler somewhere >> yeah no absolutely you know we actually have some uh social media programs hug your mainframe day. And if you look at LinkedIn on those days, you'll see lots of people going about and hugging their main frames in their data centers or where we have them. You know, certainly when we do launches, you know, the the box is certainly front and center. And there's a lot of thought actually that goes into, if you believe it, the the design of the door, uh, and actually making it kind of a very, you know, very powerful kind of impactful visual, uh, to, I think, represent what's behind what's behind that door. Mhm. Now, do modern mainframes exist or are all mo are are all mainframes legacy by definition or can I buy a new mainframe?

[00:04:04]>> Well, you you can absolutely buy a new mainframe. Um, you know, I think it's funny. It is this perception that I mean there's and we often see these like super old pictures from, you know, the 1960s or 70s of these computers that filled an entire room and that's what people's perceptions of the mainframe are because, as you said, they don't necessarily go and see them. But, you know, the mainframe is kind of like a car. Sure, the car was created a long time ago, but you go buy a car today, it's nothing like the Ford Model T that you bought in the early 1900s. The same is true for a mainframe. It was invented, yes, in the '60s as like the first, you know, modern computing engine, but the mainframe of today is the most modern cutting edge uh compute stack that you can buy on the market.

[00:04:51]Yeah, I think that's an important thing to remember that, you know, I I I think, you know, Ford Mustang and I think like a 60-year-old car, and I'm sure they're beautiful, but I can buy a Ford Mustang today. And I think the thing that's most significant, even though IBM mainframes have been around since like 51, 52, is if I understand correctly, the Z in IBM Z stands for zero downtime. Is that true? Well, the Zeke can uh you know, there's probably a lot of things the Z can stand for, but we do have one of our claim to fame is 89's uh you know, just inherently in the box. And I don't think there's really much else that can compare to that.

[00:05:26]>> That that is Yeah, I I can think about five being super challenging. Eight, I can't even conceive about what is that even me like a minute and a half every couple of years.

[00:05:36]>> Yeah. Yeah. I think even long I think it's like 300 milliseconds a year or something crazy like that which you know turns out to ba basically practically almost never >> and now by default they they run virtualiza virtualized right everything is virtualization is required on on an IBM >> yeah I mean there's lots of ways that you can carve up the compute right and make it make it accessible you know you can run zos on on the platform you can run Linux on the platform not everybody realizes that so there's a lot of a lot of options for how you can leverage you know this highly resilient uh compute capacity that's you know you know it's really meant to be extremely optimized as like a full stack and run very dense and run very hot very reliably versus kind of the distributed scaleout model that you typically see in the cloud.

[00:06:28]This is really more of a scale up kind of model. That is a really interesting way to look at it. Like it is like a a cloud in and of itself. But to your point, it's for big bursty batch heavy work and not little tiny HTTP get here and there kind of like work that like most modern cloud systems are doing as they're managing web servers. But do people run like a web server on do they treat a mainframe like a cloud or that's just simply not what they're for?

[00:06:56]>> Yeah. No, absolutely. I mean, uh, yes, people of course run batch, but there's, you know, a ton of online, uh, transaction processing. That's really a lot of the bread and butter, uh, of of what the systems are used for where speed, uh, you know, with ACID properties, low latency is all super super critical. And so, um, you know, we absolutely can run web servers, we run databases, you know, all all the same things you run everywhere else. Uh, you know, you can you can run on the on the Zplatform also.

[00:07:25]>> So then when would I pick it though? I think the part that would be confusing for folks that are maybe listening is that everyone's so used to just throwing it in a cloud, one of the many clouds, and I don't know if I would if I were faced with a problem, how would I decide, you know, this is an IBM Z16 sized problem?

[00:07:43]>> Yeah. So, you know, we talk a lot about fitforpurpose workload choices and, you know, like IBM, you know, your your Z main frame is is kind of like the race car. Uh so you don't you don't need it for every application for sure if you're going to have uh something that is you know maybe smaller volumes uh maybe perhaps more experimental uh you know you may not especially if you don't already have a platform if you don't already have Z you know you may not choose to put it there uh but you know the thing about the Z platform actually is it gets uh it gets better and gets even more cost-ffective the more you put on it right and because it can run so densely and so hot a lot of our clients run their boxes extremely hot. Um, and so, you know, thinking about those like core transactional workloads where you really need the resiliency effectively your brand trust is going to be at risk if this goes down where you want to access to perhaps the data and the data gravity that you have in the platform already. You know, a lot of times, you know, why would you make a copy and introduce security risks, you know, cost of uh, you know, storage costs, movement costs, you know, and all of those challenges, you know, the the compliance cost and then managing this data somewhere else when you've got it on the platform and just being able to really leverage it. So there's really kind of a gravity that I think forms around these platforms where clients already have these existing workloads.

[00:09:07]And so anything that needs to kind of take advantage of that transactionality or the data that you may already have there, that's where you really want to kind of keep extending and leveraging the platform versus trying to create some, you know, replica somewhere else.

[00:09:22]>> Do large systems mix and match? Like if I were building my um I don't know like an airline, I would have parts of the workload of managing the airline run on a mainframe and then parts run on a cloud and they would live together in the same data center or near each other.

[00:09:37]>> Yeah, absolutely. I mean, we really think of this as part of a hybrid cloud architecture. And like, like you said, like you're probably not going to put your mobile, you know, part of your mobile app or your, you know, your your web page necessarily hosted on the main frame, but it's if you're going to go check your bank balance, that's going to call back to the mainframe. And so what we've really done a lot of work on is how do we enable the technologies that allow the platform to seamlessly integrate into that hybrid cloud world.

[00:10:04]you know whether it's REST APIs whether it's Kafka or you know we just purchased Confluence so you know we may have that prefer that flavor right of of Kafka out there but you know all these you know common ways of interconnecting systems absolutely apply just as effectively to your IBMZ platform and can participate natively that way. Yeah, I used to work before uh my day job at Microsoft, I used to work in banking and we would always we would have a a big front end that was running uh on x86 at the time that would handle the web pages for retail online banking and then we would interface with the mainframe via whatever technique that that bank's mainframe would use and it worked it worked very well and then bill pay was all handled at like the 2 a.m. nightly bill pay run.

[00:10:52]>> Mhm. And uh I would say that the machines that we were running for the web servers would I I would be lucky if we had two or three nines at the time.

[00:11:00]This is a 25 years ago. But we never really thought about the mainframe going down. It was always there just humming along. Now you say hot though. Let's talk about that. Hot for for context for our listeners means that you're running those CPUs at 90% at 80%. You're you're not idle. If a mainframe is idle that's a problem, right?

[00:11:19]Oh, well, it's certainly not a problem necessarily, but you're probably not getting the most, you know, your maximum cost efficiency and value out of it. So, uh, you know, and these boxes can be, they basically come fully loaded with all of the capacity on it and you could just turn on the capacity that you need.

[00:11:36]So, you know, within the box, you can, you know, set the capacity level and just pay for the capacity that is appropriate for your business. But if you need to spike or scale, right, there's that possibility of expanding up into kind of the dark capacity that basically is there, just not, you know, always enabled and that you're not always paying for within the platform.

[00:11:55]>> Could you expand on that? What does that mean, dark capacity? Because if I think about running something hot, I feel like there's not enough headroom for another burst.

[00:12:02]>> Well, so that's where it's there's there's the full box that's available that we ship that has basically everything, you know, most of what's possible in it. And then you may choose to run it at, you know, just turn on or enable say, I don't know, 60% of what's possible on the box. And so within that 60% that you've sort of entitled yourself to, you can run that at 80 90 95 97.

[00:12:28]>> But then if you need to turn on more, right, that's sort of that dark capacity that's still available in the box, but you're not actually paying for it because you haven't turned it on. You can then actually vertically scale up and turn on more capacity.

[00:12:39]>> Interesting. Okay. And that that kind of explains that the the mainframe is it has a cloudlike attribute but it's onsite right so you've got self-provisioning of resources you've got scalability within itself you've got a extra fifth gear as it were that you can pop into >> yeah and we find we found some clients you know there's obviously a lot of the core systems that run on zos but I mentioned Linux earlier and we've seen a lot of clients actually as their data centers are getting full as their power consumption is maxing out and they still have these needs for, you know, more database uh compute or more AI actually turning to Linux on Z in their IBMZ mainframes and you can get a box that's like call it Linux one that's just 100% Linux if you don't have ZOS or you could just have an LPAR it's one of those virtualized kind of slices of the compute and have that be a Linux um Linux-based LPAR and then within that we've seen a lot of clients really um do a lot of consolidation around databases.

[00:13:40]So, uh, City Bank was a big example.

[00:13:42]They actually made a big splash at a MongoDB conference a few years back.

[00:13:46]They moved all of their MongoDB from x86 onto Linux on Z and they had, you know, over 50% power savings. Um I think they had some license cost savings and they were able to achieve some regulatory requirements uh around cyber resiliency and you know uh basically immutable copies backups that could be totally separated that they weren't actually able to achieve on x86. So they cost savings you know they got better performance they had less power and they were able to achieve things that they couldn't do otherwise. Now that term that you just said, LPAR, that's logical partition. That's like a virtual machine. It's a section that you are logically as as is the title logically partitioning. You're setting aside a space. But then you mentioned regulatory things. Are mainframes somehow special or treated differently. If I uh often we're we're told if I want to get past a certain regulation, I might need to move a workload from one part of the cloud to another to make sure it's on a physically different machine. uh is a logical partition special in a mainframe that I can run you know workloads that are that need to be separate from each other but they are separate both logically and physically I guess I'm not understanding the >> yeah yeah I mean you can you can certainly um have separation uh you know within the mainframe through some of those virtual virtualization techniques I think often you know the the mainframe is a good place for those regulated workloads in part because I mean in most of these companies is it's already established. Uh you don't have to go establish something from scratch. You know, the processes are there. The you know, the the the way that we do proof points, everything is is is basically there. And so you kind of almost get it for free when you bring a workload because you're already taking advantage of that investment that's already been made um within the enterprise on that platform. And so you kind of just inherit you know that aspect of the qualities of service as long as well as you know the others that are part of the platform.

[00:15:48]>> Now when I think about a main firm I certainly don't think about AI uh forgive me in my ignorance but you know I think about airlines and banks and big important stuff and power systems and things that needs to run uh forever. How has a mainframe been introduced to AI?

[00:16:04]like are you literally running inference on things like an IBM Z16 uh or are you know where does AI fit in and when did that shift start to happen because you've been there while this happened like you were at IBM during the rise of AI are they were were IBM Zclass mainframes just ready to run AI like do they they don't have Nvidia cards how does that work >> yeah so you know I think we actually have been ahead of the curve on this so with IBM Z17 which we just announced Last year was actually our second generation of mainframe that has specialty AI chips built into the box.

[00:16:40]So we first introduced this with Z16 I about four years I guess four years ago.

[00:16:46]>> Uh and and there we introduced something called the Telm chair and that was really more of your traditional machine learning I would say as opposed to you know generative AI models. uh but we really introduced this because we saw this need for AI to be much closer to these workloads. So you know we think about it in somewhat I suppose pragmatic terms right and really focus on how we can deliver value for our clients. So they've got these workloads that are running on these systems and they absolutely need AI. Um you know if you're going to go and a lot of times they're very timesensitive. So, if you're going to go make a transaction, you want to catch that fraud at the moment that the transaction is happening, not 30 minutes or an hour later or even, you know, 30 seconds or 60 seconds later because by then you probably just had to let the transaction go because it took too long. And so, and once you've let fraud go, it's basically a loss, right? Like it's just kind of a lot of banks look at it as the cost of doing business that you lose, you know, 8% of your revenue due to fraud in a year. Um, and that's a pretty high cost.

[00:17:51]And so actually bringing AI to the platform is really critical because now you've completely cut out network latency. You're able to run it at the source of the transaction is happening and you can get you know less than a millisecond or singledigit millisecond kind of response times to those inferences and have it and and be able to then score and do it in 100% of your transactions. So then you can actually you know start to catch massively more amount of fraud and stop that which becomes you know a huge business impact and again a return on your investment for the platform and for the investment you know and I think the same is true of the data a lot of times people are copying data out and you know there's capabilities that we've built into DB2 on the mainframe to allow you to do like semantic analysis to be able to find similarities to apply patterns and classifications and you know really be able to do that on platform um and and again kind of at a lower cost you know for the use cases that matter for these enterprises and so we introduced Telm in Z16 and then in Z17 you know our latest generation of the mainframe uh we introduced Telm 2 which is more powerful can handle larger models we also introduced spire cards uh which are um PCIe attached cards that you know you can run gen you know small small to medium not not your you're not going to run a frontier your model on a mainframe, but uh you know a small to medium model kind of more special purpose uh for the use cases to be able to bring Gen AI to the platform.

[00:19:23]>> Yeah. And the density of these, like it's I'm having trouble getting my head around it because again, you just see the giant black refrigerator with the cool front door, but the the Spire cards have like 128 gigs of of memory on them and 32 cores and you can fit 48 of them and and that's just such an insane amount of processor power. You just plug the plug up 48 across all the IO drawers. That just blows my mind.

[00:19:49]>> Yeah. Yeah. Know, it's pretty cool. And uh you know, we're we're definitely seeing clients get excited about that.

[00:19:55]We're seeing a pretty big um takeoff in terms of folks buying cards, reserving space, and and kind of getting them installed as they're bringing in this next generation of the mainframe. You know, and I think we see a couple different ways that folks are thinking about using it. You know, one is uh we're really focused on how we continue to bring uh make it easier for folks to work with the platform. So think about your operators, your DBAs, your security administrators, all those folks who maintain and operate the platform and and we know that uh you know the mainframe is highly uh differentiated in a proprietary way in terms of how it's been built, but some of its interfaces are also proprietary and they're not quite so differentiated, right? Like some people don't will kick and scream if you took away their, you know, green screen or ISPF panels, but it's not and we don't want that to be the way that you have to work with the platform in the future. Um and and today, you know, there's a lot of ways that you can work with the platform without ever touching, you know, a green screen actually. And and part of that is how we're bringing Aentic AI to the platform for those operators um and for those folks that work with the platform. And you can run that leveraging your spire cards. We also see clients, you know, have business use use applications was which was kind of where I started with this and where we focused the beginning of our AI journey around is how do you actually bring more business operations agents again tied to that data gravity or that transactional gravity that you have on the platform and be able to leverage that. Um, I want to shift from running running AI models on the thing to understanding how AI is going to change how developers work because I was around at the banks when we were finding old people, pulling them out of retirement and they were helping us with Cobalt and I was also in the banks during Y2K and I you you know you've been pretty vocal about the like Watson code assistant for Z being at a developer assistant and you've been very clear about it not being a replacement. Where do you see AI as it relates specifically to Cobalt and how is it going to help? I mean, there's young people now getting into Cobalt. Like Cobalt's a growing, shockingly a growing business and like if you're you want to do some interesting work, go learn go learn Cobalt. Where do you see AI fitting into that?

[00:22:05]>> Yeah, I think AI is a huge unlock for for development. You know, the I think the challenge with development on the mainframe historically, it's not really been the language. It's been a the complexity of the applications and the fact that you know in many cases they've been around decades and are complex and don't have good documentation or tests but b it's been the tools that develop and that's really been the hardest friction point are the tools that developers have had to work with and so we've been on a journey for quite some time now to completely modernize the developer experience so you can use VS Code you can use Jenkins right you can have a fully modern CI/CD pipeline with automated tests, deploy code with Anible, right? Like every, you know, use Artifactory, all the things that you would do on, you know, a distributed platform, you can have that exact same developer experience with your traditional ZOS applications and clients that have invested in this because, you know, it's a tool change, but it's also a cultural change, right? You know, I think if anything that's probably the hardest part of it is the cultural change aspect and but once you do that, we've got clients that are deploying code changes 20 times a day on a traditional ZOS kicks DB2 banking app, right? Like you can have the same agility and same modern tool tool, you know, tool abilities. We also have other languages on the platform like Java and other things, right? And we've seen a lot of folks do as they operate incrementally as they modernize incrementally transform their applications to have a Java interoperate with Cobalt. And we can do that in the same transaction scope. We've invested a lot in this interoperability capability. And so at what AI really does is it enables you know you've got this modern base of your tool chain and now you think about these applications and they're kind of like big tangled up balls of string and folks can't see through them. They're not sure what's connected to what. They're afraid to snip out a piece. They're afraid to kind of pull on a string. And what AI does is really kind of helps give you that X-ray vision, I think, into that application. And what we've done with AI is really a combination of, you know, analysis that is not necessarily AI based, right? It's more static analysis and some dynamic analysis we've built in, but you know, uh, declarative, I would say. And we use that actually as context to help inform the AI and the AI models to do their jobs better, which I think is what differentiates our stack versus just any other AI stack that you're going to get out there. It's because we have this base knowledge of how everything is interconnected. What transaction on kicks may hit a DB2 record over here, you know, impact something in IMS and come back around, right? That's the type of stuff that's hard to just know because a lot of times these code bases are tens or hundreds of millions of lines of code and you can't like wrap all that in your head and even the AI models can't quite consume that quantity either. And so when you combine this like analysis and this declarative understanding of the applications and use that in the contexting and in the prompting and how you um you know do documentation around the application it now totally empowers the developer to you know not be afraid to find that dead code and cut it out not be afraid to you know rearchitect this section of the application to make it I won't say microservices but more serviceoriented Right. And and really helps you to have that confidence as you do those changes so that again you can move with that same agility on the platform.

[00:25:42]>> Yeah. I feel like people who are getting their feet wet in AI think that we're we the AI people, you know, I'm an AI person at Microsoft and you're an AI person at IBM of sort are saying that like these LLMs are going to solve everything. But your point about static analysis is so important. The LLM may be an orchestrator or a coordinator, but that does not mean that the software development life cycle changes. That does not mean that static analysis changes. It's the tools around that and then the orchestration of those tools that that is the the real unlock.

[00:26:14]>> Yeah, absolutely. I mean, you know, certainly there's this view that, you know, AI is going to change everything and and in some ways it will, but I also think that going back to some of the basics of like what makes good agile development practice, uh, you know, what makes good engineering and good architecture become even more important, especially if you're going to be developing code at a significantly faster rate and pace. uh you know like we as humans aren't going to be able to necessarily consume or put everything together unless we have the right you know kind of guardrails we have the right architecture design that we're using really as input into this AI so that like we're driving the AI to do what we want as opposed to the AI telling us what should be done.

[00:26:59]>> Yeah. Yeah. I want to go back. You mentioned Java like I feel like in 2026 like a lot of people don't realize that Java is on mainframes and like why is there an awareness gap in 2026?

[00:27:10]>> You know that is that is a great question. I I would if I knew the answer I would fix it. You know, I I it's funny at the Z16 launch four four or five years ago. you know, I went there and I was talking to a client and at I think at that time Java had already been on the platform 20 years and I was talking to someone and they didn't know um you know even today there may be folks we've been over there Java's been there over 25 years and folks don't realize um you know we've invested in Java on on Z for for a very long time now in this interoperability for a very long time and we actually this is part of the beauty of the full stack of the platform is we actually optimize machine level instructions to make Java go faster. So things like sorting on Java or we actually reduce the amount of time that it takes to do garbage collection on the platform. Like Java runs better on the Z platform than any other platform.

[00:28:07]Full stop. It's more performant. It's more efficient. So, you know, people a lot of times think, oh, I'll move it to Java and I'll move it off. Well, you know, hey, you've got that option. But, you know, people think that modernization is a language thing. It's actually the value of the Z platform is is not really about the language. It's about this full stack optimization top to bottom that we do. And you may be able to, you know, move an application that's Java to another platform, but is it going to perform the same? Is it going to have the same resiliency? Is it going to have the same latency when it's now disconnected from the data and everything that it's close to? And so it's really about these architectural decisions and again what is the fitforpurpose platform based on your use case that you really need to think about as opposed to just a language swap which you know at the end of the day Cobalt is not that hard of a language pretty readable and pretty understandable.

[00:29:01]>> Yeah. Yeah. And not just Java. Shout out to people who are running net on their uh their IBM mainframes as well. So >> there you go.

[00:29:07]>> As a fan I have to say I appreciate that.

[00:29:11]I'm curious as someone who stayed some stayed at IBM for so long like what kept you excited about staying there because a lot of people are doing the one you know 18 months here and I'm proud to announce and their LinkedIns are filled with them moving from one place to another but you stuck with it. Yeah, you know, it's, you know, it's not something I initially planned uh starting out. You know, I think a lot sort of the story of a lot of folks at IBM. You think you're going to join, maybe stay for five years, go somewhere else, and uh then you stay and then, you know, you think later on, oh, maybe I should look around and go somewhere else or someone pings you about an opportunity, and you look into it, and you do, and then, you know, you look back at what you've got and and you decide to stay. Um, and so for me, I think it's been a continual series of conscious just choices to remain at IBM.

[00:29:56]And I think it's it's a few things.

[00:29:58]Partly it's, you know, people always say this about where they work, I guess, but it is the people you work with and the culture that you have. Um, you know, IBM is a big place and there are a lot of amazing people that I've had the opportunity to to work with and learn from. And we actually have a constant infusion of new talent and new perspectives coming into the company through the acquisitions. and I've had the opportunity to work with a number of people through some of the acquisitions that we've done as well that have you know broadened my horizons and given me almost a startup experience uh in some cases and I think it's also the the opportunity uh you know I've done a lot of things in my career at IBM from you know database engine optimization to you know analytical queries and mobile applications and cloud and you know now everything that I get to do on the mainframe is really kind of a a microcosm actually of the entire software industry because everything applies to the mainframe and I get to focus and choose what is most relevant and so for me it's been about that opportunity to continuing to work on great stuff the people I get to work with and then the clients um we have really deep longlasting relationships uh with many of our clients and so it's kind of this like multifaceted kind of community I guess that you get to be a part of and I've always felt like I've had the opportunity to keep learning and progressing. So, it's been it's been an exciting ride.

[00:31:21]>> That's very cool. Well, thank you so much for hanging out with me today.

[00:31:24]>> Yeah, thank you. This was fun.

[00:31:26]>> We've been chatting with Skyla Lumis, general manager of IBMC Software, and this has been another episode of Hansel Minutes, and we'll see you again next week.