Strategic AI and Cloud Solutions: GitHub’s Blueprint for Modern Development Success

In this episode, David welcomes Ari LiVigni, Senior Cloud Solutions Architect at GitHub, for an engaging conversation about cloud infrastructure, cloud adoption, AI, and security. 

Join us as we discuss:

• The importance of effective cloud solutions for data protection and security, including a look at tools from GitHub, AWS, and Azure.

• GitHub's significant milestones, including its acquisition by Microsoft and the launch of innovative tools like pull requests and forking.

• How GitHub Copilot enhances productivity by enabling developers to focus on code prompts while automating routine tasks.

David Joy:

What is up, everyone? And thanks for tuning in. In today's episode of Big Ideas in App Architecture Podcast we speak to Ari LiVigni, who is the Senior Cloud Architect at GitHub. Yes, that place that's become synonymous with checking in code and code collaboration.

Ari and I talk about his journey as a cloud architect to his excitement on the possibilities of AI, GitHub Copilot, and how it all comes together at GitHub. So, pump up the volume and enjoy listening in to this episode. All right, so welcome to the Big Ideas in App Architecture Podcast, Ari. How are you doing today?

Ari LiVigni:

I'm doing well. How about yourself?

David Joy:

It's been good. I've been just busy with some work. I know we had to reschedule a few times, but I'm glad that we kind of finally made it.

Ari LiVigni:

Yes, likewise. Yeah, same here. Oh, it's very busy with work and other things, for sure.

David Joy:

Brilliant. The last time I spoke, I think the last thread we left, it was lacrosse. I think we were talking about your kids playing it and your coaching and getting involved. How is that going?

Ari LiVigni:

Good. Yeah. I mean, it's off season now for both the college guys that I coach and my son, but my son also does cross country, so he's doing that right now. But he's been practicing for his fall tournaments for his club team Laxachusetts. So yeah, everything's been going well. He's been doing well. He's been growing into his body, getting bigger, so weightlifting and all that kind of stuff.

David Joy:

You told me last time I spoke, last time we spoke, "Go check out some videos and stuff." And it's a pretty interesting sport. I mean, every sport is pretty unique and obviously there is some respect that lacrosse has gained in my mind, so thank you for that.

Ari LiVigni:

Oh, absolutely. Yeah, I enjoy it. I'm glad you got to take a look.

David Joy:

So, for people listening in, I'm really excited to have Ari on. It's not every day that you get to talk to somebody who's working at a company whose product you kind of touch on a day-to-day basis. Right? So, everyone who is in the tech space who is involved with some level of coding or doing some sort of version control on the code they're building use GitHub.

This is a company that started off in 2008 and it evolved from Git, which was developed by Linus, who was also the guy behind Linux. And it's very fascinating how this company was formed and then got acquired by Microsoft for $7.5 billion a long time ago.

But it's been very influential over the last 10, 15 years to code developers and it's also changing. Right? So when I was thinking about you, Ari, and I was thinking what we should talk about, I was like, it's so awesome that I get to talk to somebody who works at this company and get to know their journey and what they do at a day-to-day basis. So, I'm really excited to have you.

Ari LiVigni:

Yeah. I'm happy to be here and happy to really talk about all things GitHub. I enjoy working at GitHub and working with our customers for sure.

David Joy:

Brilliant. Yeah. So, I don't want to butcher your interview here and your introduction. So, let's start with why don't you let the people hear a little bit about yourself and let them know what your role at GitHub is. Tell us a little bit about your background, some of your strengths.

Ari LiVigni:

Sure. So my name is Ari LiVigni. I'm a Senior Cloud Solutions Architect with GitHub. I'm based remotely in Worcester, Massachusetts. I've been with GitHub for about two years. I've worked at other great companies too, AWS, Red Hat and VMware all great companies. And obviously Red Hat, being an open source company really feeds well into GitHub as well.

I'm very passionate. I would say the thing what I'm passionate about and what I get to do on a day-in, day-out basis is really increase and improve productivity for our customers by using our product and showing them the best ways, best practices to use it. And yeah, it really kind of keeps me passionate about it every single day. When I wake up I'm like, "Oh, here's another problem to solve and how can we solve it with GitHub, or Azure for that matter?" A lot of times a lot of customers that are using Azure on the backend and want to learn how the different services work. So, we also have to be informed and know about them as well.

David Joy:

Amazing. So, for people listening who may know a little bit about GitHub, tell us a few things that nobody knows about GitHub, like the way everything is set up and the way you kind of operate as a company.

Ari LiVigni:

Yeah. I mean, we have a really, I have to say very brilliant engineering team across for all our different products. I mean, I think a lot of people think of GitHub as like, "Oh, I put my source code up there and I check it out." We invented the pull requests, we invented forking. All of these things that are built in now are issues and discussions that really create a great ecosystem for software development from not only engineers, but I would say also for project managers or for people looking at a high level out of, "What are my engineers or what is my developer community, what are they doing?"

And I would say too, we try to emphasize on using GitHub in an open source or an inner source way. We use the term inner source because it's not really open to the public, but maybe you're inner source, you're keeping open source within your own company and everyone wants to have a view of that. I would say the other key thing that I'm very excited about, because I get to work with it and the new features and things that come out about it is GitHub Copilot. And I think a lot of people don't understand we're not just this platform. We have a lot of security built into analyzing code and looking at it as well as automatically fixing. I can go on and on about GitHub Copilot.

David Joy:

No, this is cool because I did want to get to talk to you about Copilot. I am right now really actively testing Copilot. I'm also using Cursor. I'm also testing open AI's new built-in code editor. Right? But the fundamental idea is that GitHub Copilot was one of the first ones to do it and present it. Right? So, I'm really fascinated by the AI part of GitHub now and what it's enabling. We can get into that as well. 

But let's just go back to what you were kind of talking about, your role, right? At GitHub, many times we are thinking, "Oh, this is part of my software development lifecycle. I'm coding, I'm checking out, I'm checking in, I'm committing my code. I will go in, put an issue, I'll pull some things, and push things."

These are things that on a day-to-day basis a developer is using. But you brought in the idea that there's so much going on behind that. You have a cloud infrastructure, you have to protect all of this data, you have to secure everything. So, as a cloud infrastructure person, how are you involved and how has the company kind of shaped everything to support that?

Ari LiVigni:

Yeah. I mean, I think what they've done is, they've created this ecosystem of that they have ... As a developer, you have all these tools at your fingertips, whether it's GitHub advanced security to kind of dive into more of the security part of your code, making sure we use Code QL to analyze, and Automatic or Dependabot is another feature we have that looks for end of dependencies or any vulnerabilities and reports those. And now we're getting to this point too where Copilot and the AI is there not only to identify these issues but also to automatically fix it. We have kind of a motto or a phrase that we've came about now with Copilot integrating with our security product is this, Found Means Fixed. So if a problem was found, it's automatically fixed without having to then, a developer has to go back and then look at it, look at the vulnerability.

And I think that's what it is too, is really getting the most out of our developer community without having to put more of a burden on them. They don't really want to have to always think about all the security things that go into their code. They don't always have to think about even generating code. And that's where Copilot comes in for a lot of that and automatically analyzing and fixing.

David Joy:

Oh, yeah. I mean, that's pretty cool because I have some projects sitting in my GitHub repository whose packages that I have not updated and I get these vulnerability [inaudible 00:08:19], alerts, sorry, saying, "Ah, update this." So, I see it. So, you're basically talking about how Copilot can basically handle this vulnerability and tell me that it's fixed.

Ari LiVigni:

Yeah. And even from the normal, what we call flow state or normal developer flow, I'm guilty of this as much as anyone else of putting a pull request together and not putting a lot of details in my pull request summary and not really giving my peers that info. I think we're all, in fact, I think there was an analysis done in our last universe in 2023 that only 40% of people actually fill in the pull request summary.

So, it tells you, and now GitHub Copilot, you can actually click a summary of what you've changed in that code. It will look through and tell you. It'll outline the files that were touched. And there's another time-saving thing that we all dread, like what am I going to put in this summary? I'm too close to the code. Am I going to miss something that my peer needs to know to really review this thoroughly? So, that's been a huge productivity boost I know for myself and I know for customers that we work with on a day-in, day-out basis.

David Joy:

Amazing. Well, that's really awesome. So let's dive into GitHub itself and the way it's available to us. Right? We go to GitHub.com and every person can create an account and that's their private account. It's sort of publicly available if you keep public repositories. But then we also have the opportunity to create organizations. And that's where you work with customers, right? The customer org and you have their projects and things like that. Is all of this running on-prem, all of this running on the cloud? How is it?

Ari LiVigni:

Sure. Yeah. So, we even have this concept of enterprise, so there's an enterprise customer and then they will have maybe multiple organizations as part of their enterprise as well, and it can run on GitHub.com and we're maintaining that infrastructure. So, obviously very cloud orientated, off-prem. But then we have our GitHub server product, which they can just run in-house. We're getting this data residency where we'll have multiple cloud locations if people still want to use cloud, but they need it in their region. So, we're doing a lot of things to expand our footprint there and still offer on-prem solution, a hybrid solution or a completely cloud solution where we're maintaining all of that infrastructure for folks on GitHub.com

David Joy:

Right. And that addresses a very important question, right? For every enterprise, as much as keeping transactional data is important, keeping the code of those applications is equally important because there are developers working on it, they have to make sure things keep moving, keep working exactly how they want them.

So, it's very critical for this data, this application data, these code repositories to be available. So, resiliency is important. So, how do you do all of that resiliency and high availability on GitHub as well? Is it a multi-region setup with bunch of load balancers? How do you go about that?

Ari LiVigni:

Yeah. I mean, now we have the main, really it's located here in the US. But the idea is with this data residency that we'll have presences and just how AWS or Microsoft does for Azure, where we'll have these multiple sites where they can use those as a local cluster to use and get access to, to use GitHub.com there, get better throughput. And also in certain countries they can't have that data be somewhere else in another country, another location. So, this data residency will help them fix that problem for them.

David Joy:

Nice. That's awesome. Yeah. I'm really excited for that because I think going forward in the future, everybody kind of wants data residency, there are requirements around GDPR and DORA and CCPA, all of them are adding to the complexity of how enterprises want to interact with any kind of data that they are kind of putting on the cloud. Right? It's very interesting. So, let's dive a little bit into your own background. You've been working in the space for a while. What made you passionate about cloud infrastructure? And I know you actively also worked on DevOps and communities and stuff, so tell us a little bit about that.

Ari LiVigni:

Yeah, I would say really the digital transformation and the consumption based model of that kind of shift. I hate to date myself, I've been in the industry for over 20 years. I'm one of the old guys, but being this on-prem, maintaining servers and connecting them for people to use, doing capital budgets and really shifting to this model of more of an OpEx of okay, only pay for what you're using and you can ramp up projects much faster that way than having to plan six months to a year out of what resources you may possibly need, that may sit idle, or may not or you may not have enough of. 

And I think that's what really made me passionate of getting into the cloud space was it was kind of the perfect, the nirvana of computing where you had to worry less about infrastructure and really concentrate on the core development of your applications, your services and your software for your end customers. And really that's transformation. That digital transformation, I think, really made me excited about the field and getting deeper into that. And for DevOps especially, it really solves a lot of those problems.

David Joy:

Nice. That's awesome. So, did you start off working on any particular cloud at first? I know now you're working on Azure because of just the way it is, but did you start on AWS or what was it like?

Ari LiVigni:

Yeah, I started really working with AWS a lot. I had a lot of interaction from Red Hat because we used a lot of AWS on the backend of OpenShift, our Kubernetes offering that's there and still growing and doing great. It's a great product, I love it. But that's how I got kind of introduced into AWS and really using it on the backend. So, that's my first cloud. I would say probably the one I know the least at this point is probably Google. I don't really use Google too much, but Azure and AWS, I feel very comfortable in all the services that are offered in both of those.

David Joy:

Right. And so, I'm going to ask you for an opinion and you can give me a politically correct answer if you want to, but which one do you prefer personally?

Ari LiVigni:

Yeah, I would say both AWS and Azure are very comparable in the services they offer. I think Azure has some better solutions there when it comes to FedRAMP or governance or more stuff when it comes to the PubSec space of that, that I didn't find a lot of that easy to use on the AWS side, so I'll give a little bit extra points there for Microsoft Azure, but that's what I would say.

David Joy:

Yeah, safe answer, I'll take it.

Ari LiVigni:

Got to stay safe.

David Joy:

I know. It's very interesting, Ari, when I was also researching you, I went to check out some of your previous work that you've done and you're a big passionate person in CI CD and Jenkins. 

Ari LiVigni:

Yes. 

David Joy:

And is that what led you to also get to GitHub? Tell us a little bit about that.

Ari LiVigni:

Yeah, I would say, I used to work, I worked at a startup way back when and I got exposed to Jenkins, which was at that point was Hudson, which is what know Kohsuke created this ... We worked at some Microsystem, so I know the whole history. I love the history of software and where we've been and where we are now. And first we're managing this right on bare metal machines, then it's VMs. Then we actually got to containers, of using a containerized or container orchestration. And I think what I loved was really, that was the old way of thinking about it and we needed to really come up with a way where it was more cloud native, more oriented. That's kind of where I got into the Tecton space of here's something that's pipelines and CI CD are first-class citizens built into Kubernetes. 

Tecton is the open source. OpenShift actually has their own version called OpenShift Pipelines, which is built on Tecton. And I think that's what really kind of was like, wow, we're really getting to the point where now as developers, testers, anyone that's using CI CD in any way shape or form have to worry less about the infrastructure and more about like, "Well, how do I just ... While I'm not only coding the software, but I want to write a quick pipeline to test out this, how would I do that?" I could have that embedded in my code and I think that's the evolution of where we've gotten to from where we started.

David Joy:

100%, yeah. And as you were saying, you've been in the space for 20 years and there've been some really interesting changes or paradigm shifts. We went from servers to VMs to containers. We went from databases that are traditional to NoSQL to now distributed SQL that we lead in the industry with, with CockroachDB, and then we also have this complete transformation of AI that was in 2013 to where we are right now. We have a magic tool right now available, right? And so we'll not talk about AI in this context because it kind of overshadows everything. What, according to you, really was the most significant change? Obviously, let's talk about the cloud as well. Which one in your opinion was the most impactful transformation in the last 10, 15 years that you feel was really significant for the space?

Ari LiVigni:

Just in cloud in general or CI CD or ...

David Joy:

I would say cloud, CI CD, Kubernetes in all of this, which you feel ...

Ari LiVigni:

I would say something that blew my mind, and I got to see this first in AWS, was using Lambda and serverless. Here, I don't have to worry about not only just the infrastructure, but I don't need to set up my machine to have libraries or it's like I picked the version of the code that I want, the language, and all I have to worry about is the code and write libraries and write code there and then have that seamlessly integrated with a DynamoDB on the back end or an application gateway on the front end. And I got to experience that working with a customer at AWS was like, "Wow, it just simplifies everything." And that was when I was like, wow, this is just ... It's like a light bulb moment. 

I think obviously not talk too much about AI, but I think that's our next killer app or our next amazing feat. It's almost like when the internet came and then VMs and virtualization came. I feel like those two things combined of where we put less of the burden of any of that infrastructure on our developers or people maintaining it and it's really kind of offloaded to a cloud provider. We've really moved very far ahead.

David Joy:

We'll get right back to today's episode in just a moment, but before we do, I want to let you in on a secret, big ideas like those you hear on this podcast every week, don't need big databases to start. With CockroachDB Cloud, you can bring your apps to life quickly and without upfront cost, architected on the same resilient distributed SQL platform that industry leaders trust with their mission critical workloads. So sign up for free and start bringing your big ideas to life today at cockroachlabs.com/bigideas. 

Agree. No, I agree with you. I think this is a great segue to where I wanted to go, is talk just in context about our love for GitHub Copilot, right? And GitHub essentially you can say is a target audience, target users is developers and the software development lifecycle. And code is completely the center of attention right now. Pretty much every model out there from Sonnet to OpenAI mini models, they're all just really focused on improving how we code. And I think the implementation of Copilot also kind of changed the game in how I interact with the code and how I communicate. So, tell me a little bit about that and what you are working with your customer on.

Ari LiVigni:

With Copilot in particular-

David Joy:

With Copilot, yeah. 

Ari LiVigni:

So, I would say this idea of really, instead of worrying so much about the details of the code, how do you prompt Copilot maybe in Copilot chat to get the responses that you want? Prompting becomes an art of how you're going to talk to the AI to get a more complex, robust answer output back. I would say the other part of that that I've noticed when I'm coding in the IDE in the editor is the comment driven development, where I'm writing a several bunch of comments and I'm kind of narrowing down of exactly what I want to get out of that piece of code and then Copilot is spitting that code out so I can review it, accept it, maybe I want to ask another question of it. 

But I think as GitHub, we've done a really good job of integrating that in the IDE. It's integrated in GitHub.com and it becomes that tool that's alongside of you. I mean it really is Copilot, a peer programmer that's not a human being that's there to help you out and kind of drive you to your end result much faster than if you were coding something on your own. 

I would say the other big part of it that I've noticed with customers, you'll have customers that are maybe proficient in one language but and they need to then translate that into another, and instead of having to go out and now I have to hire a whole bunch of developers or find developers that know that language, they can have Copilot guide them through like, oh, this is what it would look like in Node.js. This is what the translation would be. It's saving time, money, all of those things, but it's also growing the skillset of your own development community within your company. And I think that's just fascinating to me.

David Joy:

No, I agree. I think that's the cool part. When I first tested Copilot and I was writing a really, really basic function, so I wrote it first in JavaScript and I was like, okay, let me see if this can be converted into Ruby, which I just completely don't understand. And then it just converted the entire code and I was like, this is phenomenal that I can do it through the IDE. And it's fascinating and not just the fact that GitHub is becoming my solution to code, the way it can now be integrated, say with Netlify or [inaudible 00:23:02]. That is also very powerful for me now because I can completely do an entire CI CD through these solutions and it's completely automated for me, so that's fascinating.

Ari LiVigni:

Yeah. Yeah. I mean the Copilot, it's extensibility now. We have Copilot extension, so now you have these third-party tools like Docker or Sentry that are first-class citizens within the IDE that you can ask it questions of Docker or you can write your own extension. So, maybe for you as a company you have a DSL or a language that you want your folks to learn from and they can ask it questions directly. You can build that in and have the ability without having relying on GitHub to do that for you. You can do it and integrate it. I actually just had a session on some of the things that were coming down on that. It's just fascinating to me, as well as playing with some of the models that are available, the different LLMs that are available. We have the ability in our marketplace where you can play with them, you can install them and mess around with them. So, we're not just kind of, it's a full core press GitHub Copilot. It's like, oh, where else can we take this even outside of that further to help our development community?

David Joy:

Right. And one of the questions I had was, one of my struggles though where I'm coding with an AI is, and I think everybody's trying to solve it, and some people hacked it, was like I would ask it to code something and then it would code something. That was like 80% what I wanted and then I'm being lazy. Instead of putting that into an ID and making the change myself, I'm asking it to make the change through a prompt, which is a language, like my English or something, and then it rewrites that entire code. And I was like, "What if it could just cache all of this and just make change to that part that I'm asking it to?" So, is Copilot also doing some kind of a mechanism where it's caching some of these prompt and prompt requirements and then not hitting the entire code?

Ari LiVigni:

Yeah. It does cache some of that prompting, but it gets garbage collected over time. But it does cache some of that. And I would say some of the thing that you're talking about there is we have the ability where, let's say you have the code, you took one piece of that code and then you asked it a follow-up question, and we have the ability now to look at what's already in your ID in the editor and then actually just make those subtle changes and then accept it and then you're done. So, we do have that in the latest and greatest. We have that ability now to do that, obviously. 

I would say the other thing that's really what GitHub does and its philosophy is, it's very much a ship to learn. So, we ship features that may be in an alpha, beta so customers can get their hands on it and give us that feedback and we get a really good feedback loop on, well, this is valuable, this is not valuable. Or we would hit a lot of our development community if we added this feature. And I'm always amazed of how we just iterate so much that we get a lot of that quick feedback loop to make improvements in the product.

David Joy:

Right. I mean that's awesome. No, I've been using Copilot for about a week and I have to try some other things on it because part of what I was doing was testing. I do not know how much you know about Cursor AI. Do you know about Cursor AI?

Ari LiVigni:

Yeah, I've heard of it, yes. I haven't used it per se, but I've heard of it, yes.

David Joy:

Yeah, it's also pretty fascinating because one of the things I like about Cursor is that you have the code, obviously code comes out and then you can select a bunch of code and then bring a prompt right above it and just type on top of it and then it'll expand just that section. It's also fascinating. Might be good idea to add as a feature in Copilot, you know?

Ari LiVigni:

Oh, thank you. I'm going to go back to our engineers and give them that tidbit. I'm sure in a lot of, obviously because I deal with customers and a lot of our field stuff, I get a lot of those competitive reports and we do have stuff on Cursor that I've read through. So, I'm sure our Copilot engineers, our GitHub engineers are actively like, "Okay, let's add this into us and into what we have for GitHub Copilot." And I'm sure something like that will be there before you know it. And I'm sure with GitHub Universe coming up October 29th and 30th, I'm sure we're going to have a whole bunch of new announcements there that our customers and our community are going to love.

David Joy:

Very good. Are you going to be there at Universe?

Ari LiVigni:

Yeah, I'm doing a session with my peer, Brian Sun. We're doing one on, let's build an app with GitHub Copilot in 40 minutes. So, we're doing a sandbox session there, so we're looking forward to that. And yeah, it's going to be a great ... I went last year and luckily I submitted a proposal that got accepted for this year and yeah, I'm really looking forward to all the announcements and what we're going to be doing.

David Joy:

That's awesome. So everybody listening in, if you're going to Universe, make sure you meet Ari. He is going to be presenting during Universe, but just go follow him. I love your energy, Ari, this is brilliant. So, on that turn, I just had a question, just a follow-up is like, can we bring ... I did not test this so far because I was using the existing model. Can you bring your custom large language model locally and connect that to GitHub Copilot?

Ari LiVigni:

Yeah, so one of the big announcements, and we've kind of I think done a preview of it is you can develop custom models or fine tune models. So, this way if you're writing in your company, your own DSL, your own language that's specific, you can train the LLM on that to then use that with Copilot. So, I haven't played with that a lot. I know it's out there, but I'm sure at Universe they're going to show some amazing demo that it's going to do everything. So, I'm looking forward to seeing that. I was blown away last year when we announced workspaces was like ... I think from a project manager standpoint was amazing because you can open this workspace and say, "Here's the problem I have. How can we break this down to epics or stories to develop this application?" And it will go off and do that for you all. and create GitHub issues. So, those are some of the things that we did last year, so I can't imagine what we're going to show this year. It's going to be amazing.

David Joy:

Oh, yeah. I still feel like it's like we're still scratching the surface with AI. I won't be surprised that the amount of usage on GitHub increases because coding is becoming a commodity now. It's something anybody can do. The language to code is no longer Python or programming languages, but just basic English. And anybody who has experienced the app on a phone knows what they want and they can just communicate that. And so, I think that would add to the scale at which GitHub will have to adopt to these users trying to use the platform. I know we were going to talk about your cloud experience and DevOps and everything, but we went a completely different direction.

Ari LiVigni:

I'm so passionate about GitHub Copilot, like I said, I could talk for hours about it.

David Joy:

No, no, no, me too. I mean, I'm an AI person who loves to talk about AI myself, so it's awesome. I mean, let's take a pivot and I just wanted to get your perspective on, you've been in the space for a while. Some of the really interesting things that you have seen in the cloud space, for example, what are some of the mistakes that people do when they're starting on the cloud trying to modernize and move their CI CD and their applications to the cloud?

Ari LiVigni:

Sure. Yeah, it's a good question. I would say a lack of ... Sometimes you'll have a C-level executive or someone at a high level that's like, "We have to go to the cloud. It's a mandate, we have to go." But what does that really mean? I think you need to really ... Sometimes it's a lack of a clear strategy or planning, maybe an inadequate cost assessment of capital budget to OpEx. These are certain points that you need to really think about before diving into that space. Maybe it's security and compliance challenges that if we move this into the cloud, is there going to be a compliance issue, legacy applications that can't move yet, or maybe it's going to take ... I always look at moving to the cloud, it's a journey and there are short-term goals and there's long-term goals, and you really have to do a good thorough assessment.

You have to make sure that you are taking into account that you also want, when you're even deploying a cloud solution, you're pushing your workloads into the cloud, is that you want continuous improvements and feedback loops. So, if something's not working or something's not a good fit, it's okay to leave it on prem until there is a good solution there and not to force that. I would say too, cloud native design principles, taking that into account of like, can we containerize this workload? Do we just need a Kubernetes cluster so we can orchestrate those containers and have our platform run through that? Is serverless an option for us? So, I think a lot of folks, they don't plan ahead or scope and assess thoroughly enough. I think it's so important to do that and then also revisit those plans as you go of maybe some adjustments need to be made as you go along. I think that's what gets missed some of the time.

David Joy:

Yeah, no, I agree. I think many times people have this insane idea in their mind that, oh, I can just move to the cloud and I'm going to be fine, and it'll take just two weeks. But that's not really the case, right? Because you have to look at where you are, where you want to be, and you have to understand what that translation looks like. And also in today's day and age, scale and high availability is pretty much critical. And so pretty much everything that, as you said, you kind of miss out on those things and adopting to those become really important. So, one of the other things I was really curious to kind of know and learn about was, with all these changes that have been happening in the space, how do you personally keep up with keeping up with the times? Because there is so much information out there.

Ari LiVigni:

Yeah, I think it comes up with reading what new things are coming out, listening to podcasts such as yours, blogs, reading as much as I can. I kind of subscribe to certain types of articles. So, I'll have it pop up on my feed and I may not read the whole thing, but I want to get an overview of maybe some new technology that's coming out. And that's what really helps me keep on top of it. 

I'm also really fortunate enough to work with peers that they themselves are on top of stuff, so they'll share that with the rest of us and really give us some high level TDLR are on what it is, and then we can dive deeper if we want. I think that's what allows me in this fast-paced cloud world of how to adapt and keep moving or look to the future. And it may be something that's not here now, but it's coming down the road in six months to a year or to two years. And I think that's what allows me to do that. I'm always listening to some type of podcast or reading a blog or an article, and that really keeps me on top of what's going on, latest and greatest.

David Joy:

That's awesome. So as we come to the close, I want to be cognizant of the time, but I'm really curious, what's your advice to young people? I know you have kids who go play lacrosse, you give advice to them obviously, but what's your advice to folks who are trying to get on a journey, a similar path as to yours, as to what should they go and refer to, learn that they can follow the path of a cloud architect such as yours?

Ari LiVigni:

Yeah, I mean, I would say Microsoft has a great ... The MS Learn and a lot of it is free, is you can really dive into not just reading about something, but really using a code space or repository and dive into that to play around with it. And I think that's how you learn, at least for me, that's how I learn the best is really hands-on of, okay, let me dive into this and do a sample workshop on how this works or a hackathon. I would say that you almost goes way more miles than just reading or watching a video is really getting your hands dirty and getting deep into it. So, that's my big advice. I know that's what has helped me over the years kind of stay on top of things, the many years, but that's what I would suggest and advise folks getting into the field.

David Joy:

Very good. So, how do you prevent people like me from getting lazy and staying up to date with code because I'm just going to use an AI, right?

Ari LiVigni:

Well, I mean, it does. We have something we use internally called Rewatch, or even our Zoom meetings that we have, the summaries that come out of that or the AI generated stuff that comes out of those meetings, it's great. It saves me so much time that I don't have to listen through a whole video or even ... I guess it's not even just AI, but the idea that it kind of summarized different sections of the video. It knows topics were switched. So, if you just want to go to that one section, you can do that. So, in this world, I think we all have limited time of what we can spend on different things. I have kids, I have a life outside of this, so how do I balance it all? And I try to lean on AI as much as possible so I can be as productive and as efficient as I can be

David Joy:

For sure. No, I agree. Well, this is awesome. Ari, it's been such a fascinating conversation. I absolutely loved having you on the episode today. For everyone listening in, go follow Ari. Ari, what's the best place to follow you and track what you are working on?

Ari LiVigni:

Yeah, I'm on LinkedIn, obviously GitHub as well. Just look up my name, Ari LiVigni. It's A-R-I L-I-V-I-G-N-I. You can find me on YouTube as well. I have presence pretty much almost everywhere, but on LinkedIn too, I do post a lot of things about some of the innovations we're doing at GitHub or GitHub Universe. But yeah, you can find me in any of those places.

David Joy:

Awesome. And go to GitHub Universe. Ari's going to be presenting there, learn how to do copilot in 40 minutes and build an app and hopefully build a $1 billion startup out of it.

Ari LiVigni:

Yeah, yeah. Our app that we're going to do, and we're going to make it kind of fun, is we're calling it the OctoFit app. So, we're going to create an app that's for exercise and keeping track of working out. So, we're going to make it fun for folks in the audience.

David Joy:

Well, that's awesome. So, I hope this is a fantastic demo for you. Is it like a full session or a demo where you're going to walk through everything for everyone?

Ari LiVigni:

So, we're going to do a brief intro. We're going to do a demo, and then it's actually going to be a workshop. So, there's going to be people hands-on, on keyboard working through the solution. We are time boxed to that 40 minutes. So, the idea is we're going to show them so much to get them started, and it's almost they can take that homework home and play with it more in their own environments if they so choose

David Joy:

Amazing, amazing. And all they need is a GitHub Copilot IDE to get started with this.

Ari LiVigni:

Yeah. All they need is a GitHub account and an IDE, and then, yeah, and that's all they need. And on-premise have the machines provided for them with the Go Copilot license, as well as a test account there, but in their own use, they can use GitHub Copilot free individual version for now. A lot of those features that we used to have at the enterprise level are now offered for free in business. So, it doesn't mean you need to be in the whole GitHub ecosystem. You can use it within your own IDE and take benefits of using Copilot and Copilot chat right there in your IDE that you're used to using every day.

David Joy:

Nice. That's awesome. Well, if you are a customer of Ari or if you're somebody interested in using Copilot, wants to learn, hit Ari up on LinkedIn, as well as go check out the Universe workshop presentation that they'll be doing. It's been an absolute blast talking to you, Ari. I know we had a completely different plan, but we went totally in a different direction, but I absolutely loved it. It's fantastic.

Ari LiVigni:

Yeah, same here. I loved it. I enjoyed it. Thanks for having me on. I appreciate it.

David Joy:

Thank you so much, Ari.