Integrating AI and Software Development with Patrick Debois at JFrog swampUP 2024

This is Techron tv. Hi everyone. We're back here in Austin for Jfr Swamp Up, and we're having a, a just a fantastic first day talking to some of the speakers, some of the folks here at credit.

Uh, well, we just did Credit Karma. That's why I think some of the fun he, he presented as well. A lot of announcements and video partnerships.

GitHub Integration security. We're really happy to have this next guest on. com audience, he's no stranger.

Uh, actually for many of you in Cloud native now and even Security Boulevard, he's no stranger. It's the one and only Patrick Dubois. For those who don't know Patrick, we have him listed here today as a social impact leader.

Where did we get that? You said, well, where do we get anything today? We go on AI and we ask, I asked ai, give me a few title suggestions for an influential community leader.

And it came back with community engagement strategist, community impact leader, which I liked. It's always trying to be nice, Al always. It's always, it's always nice.

It's always nice. It's always nice. But by any other name, this man is a singularity in the DevOps space.

He's Patrick Dubay. He really doesn't need a title or an introduction. Patrick, it's great to have you here.

Thanks For having me all and I to thank you. Right? Yeah, Absolutely.

So actually, before we get into what you spoke about here at Jfr at Swamp Up today, let's give people a little bit, as I said, people in DevOps know Patrick Devo, but you've been on a bit of a personal journey yourself the last two years, three years exploring the intersection of AI with software development and operations. I know you and my friend John Willis and, and, and some of the other folks, Chris Re and, uh, um, uh, mark Kinkle have been doing workshops and, and a lot of, uh, sessions on, on, I call it AI in ops, but operationalizing ai, I guess is the word. Give people a little bit what, you know, your journey these past couple years.

Yeah. So, um, you know, it's been roughly, what was it, 2009 first DevOps days, kind of DevOps. That's 15 years of DevOps.

Well, I can tell you I got a tiny little bit bored. You doing the same thing thing? Yeah, The same problems, different ways.

But, um, and then during the pandemic, I actually, um, you know, everybody was at home. People asked me to do some talks and so on. So I thought about like, what about automating myself to giving a talk That was even before chat.

Right, right. Right before that first started And all the stuff, and then kind of like got into like virtual TV production and kind of more ai and it just crept up, right? Everything that was hard in there was getting easier with more AI and then chat GPT and you know, God, you know, the fence was open and then I just kept learning and I've, I've been, you know, very excited in the beginning about DevOps.

Sure. Learning new stuff that kind like, you know, faded up and down. And now I completely reenergized on the Jenny I and ops and I, you know, uh, there's a technical part, which is fascinating.

And then the other part is about the impact it will have on the way we work. And that's kinda like, you know, humans code, that's my intersection that I like. Right.

I love it. Absolutely. You know, it's funny you say that.

So my background security, right? 25 plus years in security, the best security people I ever met, they were never trained on security, never went to school. We didn't have school back then security.

But you know what they were, they liked to break things, they liked to break things and then put 'em back together so they were not so easily broken the last time, the next time The thinking mindset. Right? Right.

And that, that's what made a great security person. It's that, uh, quest. It's that curiosity that ignites passion that I think, you know, I speak to so many young people, my own children included.

How do I break into cyber? How do I, you know, how do I get my first job in tech? Why, why are you going to be a cybersecurity person?

Why do you want to be a developer? Well, there's 500,000 jobs open and they started $120,000 a year. If that's the reason why you're doing this, you're gonna have a long miserable career.

Right. Because going to work to do something that you're not passionate about that doesn't get the juices. Sure, sure.

But it, it is definitely a narrative that we like to take ourselves to kind of bring ourselves free. Right. But it is also a reality for other people that they, they don't have the luxury to do it exactly what they're wrong.

Right. So, and there's nothing wrong in kind of doing jobs initially that you don't understand and know Exactly. And they're not perfect fit.

Um, because it's a learning experience already, right? So Sure. Finding what you want to do and not wanna do, you know, as hard it may sound in the beginning, you dunno, right?

You're young and you're Like, that's what I tell young people. You Kind of have to learn, just do something. Um, but it feels a very unefficient time to do so.

And obviously you want to immediately get to the goods and, you know, the younger generations, it's like the instant feedback. Yeah, That's, well, they wanna be president tomorrow, You know, social impact leaders. Yeah.

Social impact. Exactly. Yeah.

Look, I get it. I get, you know, uh, my, my, uh, both of my boys once finishing law school, one just graduated college, and he's a, uh, sports broadcaster here in Texas, out in the big country, the heart of Texas. And, and he's like, well, dad, I wanna put my time in here and go to Pittsburgh or New York or these big cities.

So easy does it, you know, don't, you can't make wine before it's time. You gotta, you gotta learn your craft. Yeah.

But I think we have to recognize that, uh, they don't necessarily have to go through the same training exercise as we did. No. It's a different World.

So they can pick up where, you know, in the middle maybe of our journey, uh, you know, because we're building a lot of abstractions on extractions that kind of helps them kind of get into a certain space. So, and then like you say, they can bring in their passion and kind of build from there. They'll learn eventually kind of the harder parts that go from there.

And I think it's still, you know, even though there's a lot of narrative of like, you know, is our profession dead and developers are we still needed? And kind of that, well, well, I, I still believe that it will not hurt to be, you know, a good professional that has a lot of knowledge saying, well, I have to train for this more than ever. I believe, You know, I, I remember when I was a young kid, my dad told me that being smart never goes out of style.

Right? Being smart never goes out of, you know, in terms of being successful in life, because there's always opportunities that will present themselves to you. It's recognizing those opportunities and capitalizing on those opportunities that determine whether you're, and how you de define success is another story.

Yeah. Whether it's money or or something else. But it determines your success.

Yeah. So, yes, being clever, being, well, I'll tell you, I didn't recognize the signals, right? No.

2009, first level day. Right. You should have retired already The first time.

I was like, okay, that's it. That was nice. Let's call it, oh, we don't do it again.

It was kind of by accident that, you know, not accident, but side guys. So the community took this further, right? So again, it Is, you never know.

You never know, right? So You just don't, I mean, look, you're talking to someone who went to law school and they hated every second of it, and then discovered computers, right? And said, what, where you been my whole life?

You know? Yeah. So it it, it is like that.

Anyway, let's talk. You're at swamp up. You're, you're presenting, give folks at home sort of a, a peek into what you're presenting.

Sure. Um, And First of all, I get the AI fatigue, so I apologize for that. But again, it's my passion as we this done.

Um, I presented a story about bringing geni applications to production, and especially let's say more like, uh, a year ago, a year and a half ago, it was other chaos, right? We all came from, Hey, this is Chachi pt, and the boardroom just said we wanted yesterday. And everybody was seeing the possibilities, the money, the impact the customers.

And we as engineers were like sweating as hell. Like, How do we do this? How do you do this?

Right? How do we do this? So the story is about, um, well, we had a set of principles in DevOps, you know, fast feedback, our pipelines, you know, working things more progressively.

And, uh, the, I talked about how we apply those principles to delivering this new type of ations. Uh, for example, how do you test something that is non-deterministic? Like a lot of the engineers had a hard time figuring out how we do this.

It turns out that there are some strategies that are, you know, you know, we use another model to test the other model and, you know, kind of things. How does observability look like? Oh, you know, we log every call we do tracing not API tracing, but from tracing.

So that wasn't the story that I brought, is that you take all the principles and you apply this to this new kind of emerging field, uh, to kind of get better at performing. And, you know, over the year and a half, a lot has changed, right? The care is in the beginning to do it yourself.

Now, vendors stepping up, you know, having solutions for a lot of different things, um, that's becoming more helpful. But still, it is an emerging thing. A lot of people are struggling how this fits in.

Yeah, Absolutely. And, and this is in line with exactly what you've got been going around the world doing. As a matter of fact, you know, people are watching this on Wednesday, on Thursday.

You're doing something here in Austin, right? Yeah. We have a workshop, uh, where we ex explain this new hotness in the delaying language of the DevOps engineer, and kind of, we show the code, we show that there's no magic, uh, how it actually works, because there's a lot of mystifying, you know, ai ooh, you know, kind of that worried us.

But, you know, we're engineers. So I, I, that's kind of what we're doing to explain that to the folks. Is it too late for people?

Maybe locally. It's gonna be challenging though. But we're running one in London, so if you're seeing this one in London, uh, that's kind of the end of, uh, uh, September, and likely it will putting on more, uh, you know, over The world.

How, how do people find that? Just find you on LinkedIn. Yeah, they watch me on LinkedIn and my website's called like Jedi be like in the store.

That's true. Jedi. Like, like the Return of the Jedi.

Yeah. Yeah. Young and dynamic it in Dodge.

Anyway. Okay. I love it.

Um, is this your first swamp up? It is, yeah. And I've been hiring this event for a long time, never really worked out, but, uh, I must say, you know, I travel a bit to events and, um, what I like about this event is that, uh, and you saw this in the keynote, uh, it's, you know, a lot of these product companies, events, they're all about the fan boys and the partnerships, right?

And I get it. But this one is, it's down to earth. Like, I, I remember there was a discussion about a slide saying, coming soon, and they didn't like this because they wanted to mention that we wanted to be there.

We don't wanna Sell You stuff that's not there. And I like this. Right?

Right. It's a realistic thing. It's kind of like no fluff.

Uh, and that's what we need, right? That's kind of the style i, We need. That's what we need.

Badly, badly. You know, I, I will tell you, I'm not the DevOps expert. You are, obviously.

But when we talk about a DevOps culture of blameless, of, of working together, of breaking down silos, and I, you know, from where I sit, I get to talk to all, I talk to all these companies out here and others, and everybody gives lip service to culture and then show me your tools. Right. You know what I mean?

Well, yeah. The culture that Shlomi and yv and, and Fred and the team here have built, it's a great culture to work in. Agreed.

Yeah. It's a high bar. People work hard and they, but they are so customer, their, their feedback loops from their customer to them.

I, I can't come how many times they ask like, please, customers tell us what to do next. Right? That was very refreshing.

Absolutely. It's not a broadcast, it was like a two way. No, no, it, it, it's a very, I I le I was telling someone earlier, a senior person at IB M1 time was he, I learned the lesson from him.

He said, if your best customers say they're your customers, who failed? Your best customers should say they're your partner. Oh, Now, And like, so for instance, I was talking, I did the Credit Karma guy, just, I interviewed the credit called jfr, their partner.

When you talk to their customers, they call Jfr their partner. I think that's an important, you know, it's an important distinction that they've instilled that cultural relationship into that Good for That. Yeah.

And it's interesting. Um, I, I learned over my career, like you imagine you have, you know, initially it was about deep dev and ops collaborating. We're getting them together, you know, kinda like, ah, you know, kumbaya in the room, right?

But, um, then it shifted to, well, your ops team or your infrastructure is not anymore in your building. It's moving to the cloud. Yes.

And then I thought like, Hey, hang on, this is the end of DevOps because everything's behind an API, we don't talk to each other anymore. Yeah. Guess what I learned?

They communicate as a partner. Right? So that kind of close collaboration, even if you're not under the same kind of, you know, authority, that is still key.

So you can apply the same thing with DAVO ops being in a very good relationship with your partners. Yeah. So that's kind like a suppliers or it's a sauce company.

And that kind of feedback loop, like you're looking for a feedback loop that Works, that that's how you build a successful, and, and they're, the proof is here what they built. But I think yo, one of their co-founders, their CTO said it today, they're seeking to break down silos wherever they can. No silos.

There's a lot of them. Yeah. Yeah.

And one of the silos I mentioned in my talk was, um, the data scientists for a long time, they kind of were somewhere in their data lake, in Their own right. In their own lake, their own Brand with a swamp. Right?

Right. But, uh, Had their own swap Is now interesting to see them coming closer to production. Yes.

Because we add a lot of demands, like, yeah, what are you doing? Is this valuable? Where's their ROI?

And all of a sudden they're like star of the show. But you know, that friction about like, Hey, I live on the side, I have my own ways of doing things, and then, hey, this is production. You should do things like that.

That's a friction point that I saw. And, uh, And it's going west, But this is the shift. Right.

Well, but I, you know what, we need to be shifting everywhere, right. That I'm, I'm, I think their tagline I saw here was ops, all ops everywhere. Every ops, every ops.

It's every shift too. Because shifting left alone is not, we've of course played that out. Um, I, I had another, you, you had a lot of stuff packed in there that we're not gonna have time for it today, but 70 slides Platform engineering is another.

org yesterday, but they're, they're represented here. They're part of the community. That's what it was.

The data scientists were hearing, well, will AI fundamentally change their job and their role? Right. You know, will today AI be the data scientist?

Sure. Yeah. Well, AI could be develop developer, it could be develop person, right?

Everybody's scared. It could be Obama who Obama raises you, right? Yeah.

So in DevOps, one of the early papers that was very instrumental was called the ironies of automation. And so if you automate the simple parts, the hard parts get hotter. So, you know, you do the automation simplicity for your developers, but the data centers are a really tough job to run, right?

Yeah. So kind of, there's a balancing counter effect. It also about the cognitive load that you put on people, you know, much like somebody in an airplane, they have like too many knobs.

It's all automated, but you know, they still need to have a feeling of in control. So kind of, that was a very instrumental paper that eventually led to a lot of thoughts in the resilience engineering, SRE and kind. Sure.

Right. We're kind of seeing this similarity in gene ai. So imagine I, you know, I have the most powerful coded store right now powered by AI and is generating all this stuff.

Guess what? I still need to know what good looks like, right? So I need to be the senior with the knowledge, with all the training, who's Watching the watchers Indeed.

That can kind of give that feedback. So that's becoming more important than ever. Sure.

Uh, and that's kind of, uh, the, the shifting that I want people to understand is that it's not gonna take our jobs because this technology is not perfect much. I know. It will it be anyway, right?

Yeah. But it allows us to scale in a fast feedback tool. But then, you know, I always, uh, make the similarity with, with the pilots, well, pilots need downtime, right?

And they need, and they Like to have that autopilot. Yeah. So kind of that similarity is the same here.

Now then comes the questions, and we mentioned this in the beginning, what do you old folks do? They don't have the knowledge. Well, you know, all the money we say is you have to put in training.

Right? Well, but that's, but that's the, the, so for anyone listening, I, you just heard Patrick himself say it upskilling training, keeping current with the latest stuff. If you're passionate about your job, you do that.

But even if you're not passionate, but everybody has pride in wanting to do what they do. Well, that's gotta be part of your, of your regime, of your everyday thing. Yeah.

You gotta stay current. And I follow the path where I said, um, you know, for me personally, what worked was every three, four years I change, I try to change my role and my domain, right? Because for me, what's important is learning all the time something new.

It also, you know, keeps my passion go right. Well that's your passion, but It brings the richness every time I got some new skill to the next thing. So that's how I roll it in.

But other people, you know, that want to go deep, A lot of people get, well, it's that classic T person versus the broom and all that. But we even getting Different in a way that, you know, we add, um, those t-shaped, uh, folks and, you know, full stack developers, but now maybe you have to let go of, you know, the coding and you have to understand what domain you're doing, right. So you kind of have to Even, you gotta gotta stay current and, and move with the times.

Hey, look, I agree Patrick, we're probably over time, but as it's always a pleasure. It's always a pleasure. Always, always a pleasure, man.

All right. Um, this guy is a treasure. Patrick DUIs here at, uh, swamp Up in Austin.

We're gonna be back. We'll be wrapping up our day coverage day one coverage in a little bit. But stay tuned.

You're watching Text Drunk tv.