Harness’s Nick Durkin on the State of Developer Experience

This is Techstrong tv. Hi everyone. Welcome back to Techstrong tv.

You know, we talked just back from RSA last week. I thought maybe I had enough of security, but how could you ever have enough of security? So in that note, let me, uh, introduce you all or reintroduce you to my friend Nick Durkin.

Nick is Field CTO at Harness a company. I actually participated in a panel we did over at RSA. We're happy to have Nick and the harness team working with us.

Hey, Nick, it's great to see you. I'm sorry I didn't see you in person, but it's great to see you here at Textron tv. Absolutely, Alan, and thank you for having us on.

Genuinely appreciate it. Uh, it's much. It's actually our pleasure to have you on.

Hey, Nick, we're gonna jump in and talk about something that no one talks about today, ai. But before we do that, why don't we talk a little bit about you and give people a little, your background and just in case there are folks out there who are not familiar with Harness, maybe give 'em a little of the hardest story as well. Sure.

Yeah, no, look, um, I started my career working in the banks, uh, doing some interesting things. 7 billion valuation and doing some fun stuff here. But ultimately, lover of tech, um, advocate for the software developers and the operations folks around the globe trying to remove the worst part of people's jobs.

And when you think about Harness, I think that's where it comes down to. We, we built the modern software delivery platform designed to remove the worst part of people's jobs. No one wants to sit there and babysit deployments.

Nobody wants to sit there and wait for tests to run, try to figure out which systems need to be running and operating. So use artificial intelligence, use machine learning to think about things like your best engineers so they don't have to Right? Remove that worst part of the job.

And, and, and look, AI is a hype term, uh, currently, but we've been doing it for the last seven years. We came out as the first continuous delivery platform using artificial intelligence and machine learning to remove that worst part of the job seven years ago. Great.

Very cool. And they really have, I mean, harness almost, not, not redefined, but modernized the whole CICD now we call it of course, soft, you know, the pipeline, the software and pipeline, the software factory. And, and, and also let's not minimize software supply chain security and the whole issue of dev set ops, you know, within the, the de development and delivery and deployment of software, right?

And, and that's part of the harness mission today as well. It was actually on the panel we spoke about at RSA. I had Sid, uh, you know, your product G four on with us.

And it was about how, you know, DevSecOps is going forward as part of this whole DevOps CICD cloud native. You know, it's all rolled in. They, they, they're very interconnected today, right?

These are not separate silos anymore that just exist, you know, in a vacuum. The way, the way we think about it was everyone tried to shift left, but unfortunately they shift the workload left. And that was the wrong thing to do.

You know, adding more burden to developers, um, already, you know, they're already having to build their own tools. They're the only engineers on the planet that have to build their own tools, that have to go search for documentation, other things. So to add more workload to 'em just is insane.

And so what we try to do is shift the information left, make sure that everyone has the right information at the right time, and that's security information that's making sure we know that our, our artifacts are at tested, that they're built appropriately, that they have the appropriate security scans. That's also about cost, right? Engineers, if they don't, if they don't know the cost, they can't even do anything about it, right?

And so bringing the right information at the right time, whether through the build, through the deploy, through the, uh, code scans, again, through the attestations, all of that needs to work in harmony. And the only way you do that is you stop that infighting by empowering everyone, giving them the information when it's valuable. And I think that's one of the key things why this platform is so successful and so powerful is because it does, it gives people the right information when it's valuable.

I love it. Good stuff, Nick. io, yep.

Yeah. Yep. Dot io.

Absolutely. Cool. All right, so another thing that was on high visibility at uh, RSA of course is ai, right?

Everybody's talking about gen AI and ai, and is it gonna replace your job? Is it gonna make you more effective? Is it gonna make you 10 x more efficient?

Um, what, what should, what should we have AI be doing and what shouldn't we have AI be doing? You have some thoughts on that, Nick? No surprise.

I, I do, I do. It's, uh, it's, it's one that's honestly, you know, for the last seven years I've been preaching it. And if you're gonna use artificial intelligence, if you're gonna use machine learning, use automation, right?

We should be doing it to remove the worst part of people's jobs. If we're using it to remove the best part, you're going to be fought, right? You're going to have conflict.

You're gonna create, you know, un unnecessary intentions as opposed to if you use it to remove the worst part of someone's job, they'll gladly adopt it. No one wants to babysit deployments. No one wants to wait for tests to run.

No one wants to figure out which system should be turned off and turned on to save cost without impacting customers. All these things, let's use artificial intelligence for, and at that point, you think about the toil when developers are spending only 40% of their time actually writing code, right? That means 60% of the time is just toil, it's wasted time.

And, and the reality is they shouldn't be having a context switch a thousand times over. They should not to be writing these tasks to do all these pieces. If we start removing that worst part, now we're doing something beneficial.

What's interesting though is that most of the industry is focusing the time on what engineers and what software developers love the most, which is writing code. And so I find it ironic that that's the first place to attack when yet there's so much more workload around what the software developers have to do, what, uh, these engineers have to actually create. And if we can go remove that toil, now we can actually start getting the benefits.

And, and I see it in a few ways. You know, one, you get junior developers become better quicker 'cause they don't have to, you know, move around all the different tooling and work that, but your best engineers become great. They become amazing.

They become the best because now again, you, you give them the freedom to do what they love. I, I agree with you a hundred percent, but you know, and I've been through more than several tech innovations in my career. I will tell you that I think AI has a potential to be, I mean, on par with the commercialization of the internet in terms of civilization changing.

But, but that being said, they all have a similar kind of evolution, which is, you know, at first call it a cheap parlor trick or the, or the magical factor. Like, wow, how did they do that? Right?

Uh, then it's the, okay, early commercialization, let's call it, let's go after the lowest hanging, easiest fruit that we can find. Then it's sort of the optimization sets in. Let's get more bang for the buck.

Let's, you know, not just do what's easy, but what's right and best. But that's a process, right? And, and so when we, when we look at this evolution of, of AI and how it's, especially as it affects like the IT stack and software development and stuff like that, you know, I can't help but think Nick, we're way, way early in this.

I think, I think we are early in the sense of the generative AI portions of it. We've been using, you know, artificial intelligence neural networks to determine things to think about things like your best, er, we've been that for a long time. And I think in the generative case, that's where, where, where we're we're, where we're early, and if you look at, you know, we just ran the state of developer experience report, you know, went out to over 500 engineering leaders and practitioners really looking at, you know, what does this look like in our world?

How is this going to affect us? And you think about it, just the burnout factor, I think, what is it over half of the developers attribute leaving their jobs to burnout. Why?

It's because of all that excess. And so we start using it again. If we start, you know, in order to get this adopted in order to your point, to have it not be the early phase to get it adopted, we have to use it to remove the worst part.

If we go and attack this and we go attack, you know, how many million developers in the world by saying, here's your job, it's going away that that doesn't win for anyone, right? That instills fear, more burnout, all the different parts and pieces. So I think it's about how we actually attack this problem, how we go work together in solving it.

Sure. I absolutely, but I, it's funny, I just got done the recording text on gang for tomorrow, actually. Well, by the time this plays in Edwin Leonard already played.

And, and you know, one of the things we, we talked about was developer burnout and developer, you know, uh, with the rides of AI and no code kind of stuff, citizen developer versus professional developer. Do we need a developer for every app? Do we need an app for everything?

Right? That, that was kind of the discussion. And I think right now we're still asking those kinds of questions.

Is this worthy of a developer's time and effort? Because you're right, one of the things we've done, and it's not just security shift left, it's the whole shift. Left movement is put more straws on that camel's back.

And, and it's funny when you think of it. So that's the most overworked, highly paid position on the, you know, on the scale here. So let's put a little more work on that one where, you know, with ai, I mean, certainly the promises is that we do make it more efficient and we put less work on there on their, uh, plate.

But as I said, I, I think it's a bit of a process to figure out what, what work is less, right? What, what is this, what is this developer consider to be sort of redundant, boring, low level, crappy crap that they don't want to do, right? Yeah.

And, and it, and it may not be, you know, uniform across every developer either. I think that's why you have different, you know, different categories. I mean, you have folks that specifically you spend their time and wanna wanna work on DevOps, you have security engineers working specifically on that.

And I think to your point, it's about empowering everyone to do the right work, to do the work they love, but do it in a way that's meaningful that doesn't cause infighting. And you know, you look at how many hours people are putting overtime in, I mean about half the days of, uh, of the month people are putting in overtime, you know, that's what we looked in the survey. And the reason is, is 'cause there's that much work to do.

So there's the work. We know that the work exists. And I think to your point, if we don't shift it left meaningfully, if we don't shift the information left, we don't give people all the right things.

That's why we get rework. That's why we get people, you know, context switching multiple times a day because people are shoving things back at them. And I think if we do this appropriately, if we make it so that we get the information at the right time, we remove that.

And we've seen that, and this is, this is those benefits that you gain is that you remove the back and forth. No other engineer, you know, I posted up on, on on Twitter or on X and I said, Hey, let's, uh, I think I did it on LinkedIn as well, and I said, let's, you know, start treating software engineers like every other engineer at our company, like our electrical engineers or mechanical engineers or chemical engineers. And people go, well, in what world?

They go, softwares do that. We have at the best. We get all the free food, we get the snacks, we get, we get to work from home.

All these fun things. They're like, great. I was like, so you, you know, you have to build your own tools, right?

You have to search and hunt and p for documentation and 20 times a day you get interrupted for, for changing, you know, the context. I go in what world is that a good way to operate? And so we start actually treating engineers like the rest for engineers of our companies, like our software engineers, genuinely, I think that's where you start actually seeing the success.

You allow them to do what they're phenomenal at. So couple of, uh, surveys have come out, how much time did developers actually spend developing? How much time do coders actually code?

And I've seen it range from like 11% to 30, 31%, no more than that. And so, you know, and, and let's be clear, coders like to code, right? You'd like to see that number at 66%, I bet, or higher.

On the other hand, testers like to test security people like to secure. Are all of us being so inefficient that the, and I haven't seen surveys on this, how much time, you know, the testers spend testing. I, you know, but I wonder is the whole stack that distracted and really can AI help us up and down here?

I think that's the, that's the key, that's the benefit. That's the area where we can focus. Because today, again, we shifted that workload.

So why do we have this full stack developer that has a list that's longer than anything? Alan, I'm sure you would not hire your electrician to do your plumbing and your plumber to do your electrical. And yet we ask our software developers to manage our cloud, to add security to like literally every part and piece of the stack.

And it doesn't make sense. We've hyper specialized everywhere else in the world, and yet here we, we give freedom, but we don't actually do that. What we do is we shove workload on them.

And I think that's the biggest issue. And so, to your point, if we allow everyone to do what they're meaningful and, and what they're good at and what they love and they're passionate about, have you seen someone do what they're passionate about? They do it with, with, with every bit of energy.

If you force something on the they hate doing, you watch that. That's what we see the burnout for. And so all this is, is about empowering people to do what they're great at, and again, giving them the information at the right time, not the workload.

Of course, swim, we play devil's advocate here for a second. This isn't Star Trek, right? And we haven't quite done away with the monetary system and it's far from Nirvana or Utopia.

We don't have people in red shirt that disappear on every other episode. But, you know, that being said there, you know, not every job is perfect and there's always gonna be things that are toil and, you know, unpleasant, redundant, not interesting. Um, I, I think we do want to try to prevent burnout, obviously across the board as well.

It's burnout as, as you know, the world's health organization has recognized burnout as a real condition. It's not just, you know, some Karen moaning at work about how much work they have. Um, so it, it, you know, this is a real thing.

What makes us think AI's the answer though, Nick. So I think this is the way to actually go and start putting things in, in, in buckets and actually allowing people to focus on the areas that they love. And it's about also making it easy to do the right thing and making it hard to do the wrong thing.

And I know that's, it's a basic statement like, oh sure that, that makes sense. If that was true, the cloud wouldn't have existed, right? If it was easy to do the right things, people would've spun up VMs easily, but instead they went to the cloud.

And so I think by making it easy to do the right things, it's giving people guardrails. It's giving them the ability to have enough freedom to do what they need, but not too much freedom to where it kills them. And I think that's really the benefit is that by removing that worst part of the job, by making it easy to do the right thing and, and hard to do the wrong thing, like let's genuinely make it difficult for people to go do the wrong thing and to focus in the wrong times or, or spend their time in the wrong area.

Now we can actually start seeing it. And that's where then you can find everything. You know, one of the things we often talked about is you have to measure things.

How do we know what to even go fix and what use AI for if we're not even measuring it? And most organizations aren't even measuring what the entire software delivery life cycle looks at. That's one of the things that we, you know, come in and focus on, which is let's look at that entire engineering life cycle and figure out where should we spend the energy and the effort.

Where would AI be great? 'cause it's something nobody loves doing. Testing, right?

Writing rego for policies, doing all these things that people hate. Creating dashboards even, right? Versus where can we actually affect change with maybe it is process, right?

Well, okay, we're not gonna maybe fix that with ai, but you have to even start measuring in the first place. I think that's where you get to harmonies when we actually know what things look like and half the people don't even, don't even understand their metrics that they're even measuring or trying to fix. I, I don't disagree there at all, my friend.

It, it is. Um, you know, I just, so I do believe we're going to get to a future where AI 10 x is people A hundred percent Right? How soon?

I, I can't tell you for sure. Well, we have quantum computing by then, know who the hell no. But all that being said, you know, the promise is there and, and sometimes you just gotta keep your eye on the prize and fight through, you know, kind of the jungle to get to, to the promise land, right?

Um, I I do think it's gonna be an interesting couple years as we figure this stuff out. I, I just, you know, and there'll be voices like you who are calling, you know, for clarity, Nick, but I, I, I think we're in somewhat of an uncharted territory and we've gotta figure out kinda what makes sense all around here. If we keep the human at the forefront and the focus, right?

And I know it's on my screen, where's it at? Remember the human? But if we actually start thinking about people, right?

And we're using AI to actually, you know, better people's lives, I think that's a great focus. That's a, that's a great start. And, and by doing that and by again we have the people that work way too many hours, we have the people we lean on way too much.

If we can remove burden from them, why wouldn't we? Any intelligent company would do that. And I think that's if we actually keep the human at the forefront and that's where we win.

Absolutely. Hey Nick, we're probably over time. I, I have a good time talking with you though, so I let it run.

But um, hey, say hello to all our friends at Harness. I know you guys recently had some big corporate news out look for that probably. com.

Absolutely. And uh, we're working with Harness on our DevOps next event, uh, report as well. So I'm sure we'll be in touch on that.

But until then, Nick Durkin Field, CTO harness here on Tech Drunk tv. We're gonna take a break. We'll be back in a minute with more tech drunk tv.