AI’s Impact on Platform Engineering and Developer Roles with Akmal Chaudhri

This is Textron tv. Hey guys, thanks for the throw. We're here with Akmal Chaudhri, who's technical evangelist for single store, and we're talking about everybody's favorite topic these days, developer productivity.

Akmal, welcome to the show. Thank you very much, Mike. Thank you for inviting me back.

I feel like this conversation keeps coming up on a semi-regular basis, and these days it's hot and heavy. 'cause everybody's talking about AI and platform engineering. But ultimately, what is the root cause of the issue?

I mean, are developers fundamentally unproductive or they spend a lot of time on things they shouldn't be doing? Um, Gimme some background on how we got to where we are. Yeah, that's a great question, Mike.

Uh, I, I'll draw from some of my own experiences. Uh, I mean, I've been in the IT space now for a long time. Um, and I think what I've noticed laterally is that productivity and, you know, the ability to use tools has become a lot easier.

It's become a lot better, uh, with the integration of AI as well. That makes many tasks much simpler. But I would say that it's taken us a while.

You know, it's been a journey, uh, and it's like a marathon rather than a sprint. So where we are today has taken quite a bit of time to get to. I think AI has really just speeded up that process.

So I, I hear about lots of great innovations recently, even within the last couple of days. There's some cool stuff that's come out, which I thought to myself, wow, I mean, that's gonna save me an awful lot of time. It's gonna, uh, take away a lot of that drudgery, that low level plumbing kind of work that I might have to do.

And now I can think much, uh, you know, at a higher level, business problems, those kind of things. Let let these software, you know, take care of the, uh, uh, of the coding aspects. The ai, you know, will do all of that for me, and I can now focus on sort of more interesting things.

It's funny, you know, if I go back in time, I kind of think about it this way. We started DevOps to get out from underneath the thumb of a centralized it, there was ticket base, and if I wanted something, I'd generate a ticket and, you know, and then I come back in three weeks and it might be redick. Um, and yet when we shifted everything left, it seems like developers are now buried under the weight of doing all that stuff themselves.

And I feel like we're trying to just maybe find a middle here between the developers having to do everything and, um, and an IT organization that historically was not as agile as we wanted it to be. So, you know, is this, are we really talking about prove? I would say so, yes, Mike.

Uh, in, in, in my opinion, uh, there may be, I mean, it's, it's like that the old saying, you know, you're two steps forward, one step back kind of thing. So there, there will be progress and I, I still feel we are moving forward rather than it moving sideways kind of things or moving backwards. You are absolutely right in your assertion there that, that, you know, there's a time when, uh, the challenges, I guess in terms of what we had to do as developers, significantly more processes took hold.

You know, things were a lot slower. Uh, uh, if you use your example, you know, raise a ticket, you have to wait a while before someone actually fixes it. Um, now I think with a lot of the automation that's in place, uh, a lot of the cool stuff that's coming out and with the advent of, uh, uh, you know, these developer tools that are coming out, I really see, I mean, great potential.

I mean, we're not there yet. I mean, it, again, it's a journey and we're at the start of that journey. I mean, it's only been, what, two years since chat GBT came out.

And I think OpenAI was working on it a long time before that, but that was really groundbreaking. Um, and I, and I think in the, in the last call we had, I told you I ignored it for about four or five months. 'cause I thought, that is a waste of time, not, you're not gonna be able to use.

And I was persuaded to look at it. And then once I started looking at it and I thought, Hmm, there's potential, you know, there's things that I can now do that I couldn't do before. Um, and I tend to use it on a regular basis now to help me.

I mean, it, it, it helps with a lot of these kind of boilerplate type of code. You know, it gives me some solutions. They don't always work, but it gives me ideas and it helps me, it makes my processes and my ability to deliver much faster as well now because, uh, I can now do things and work in programming languages and tools that I knew nothing about before.

The technology kind of helps me and it, and again, it puts me, if not give me the solution that lead shows me the direction that I need to go in. And then of course, I have to do my own thinking as well and, and, and address the kind of issues that I need. But I feel it, we are getting better.

Uh, uh, that's my kind of, uh, uh, take on it. So what does the future look like to you? 'cause you hear people describing these, um, AgTech AI approaches where we're gonna have an AI agent that is specifically trained to handle various tasks, and then we may have a, an Uber agent that kind of sits on top of that, that becomes the mechanism provoking all these things.

But is the future or software development, Uh, A mixed team of AI agents and humans that are trying to orchestrate something together? I think so, yes. Um, and I, and I'll again, let me give you a quick example.

So agents is a recent, uh, kind of innovation or maybe it's been around for a little while. Uh, and again, I took the opportunity to take a look at this 'cause I thought, well, you know, everyone's talking about this. I, I need to go and find out someone.

My boss might ask me to do a webinar or, you know, I need to know about this. So I took the time to do a bit of research, and then I, I just by chance, OpenAI release something called Swarm, uh, which is for educational purposes only. I mean, they say don't use it for production environments, but it's on their GitHub repo.

And I had a look at it and I thought to myself, okay, let me think of a problem that I can use this for. Uh, and so I know nothing about blockchain, so I thought, thought, let me pick blockchain. I'll write two agents, one that is an expert on blockchain, the other one knows about investment strategies in the stock market.

So I wrote a little bit piece of code, didn't take me very long to do this, just using their examples. And some as data, they scraped off the internet, and then these agents kind of talked to each other. So one agent goes away, you know, does this analysis and comes back with some ideas about blockchain.

And the other one takes that information and says, okay, based upon this, here's where I think you should invest your money. These are the things that are gonna be hot, uh, coming up. And I thought, that's pretty cool.

You know, you've got now these bits of software that kind of kind of talk to each other and having that human in the loop as well. Now, obviously I looked at their solutions that, like I said, I know nothing about blockchain, but I was impressed with the results that it came back with. And what I see there is great potential, you know, having, uh, these agents, um, acting in a kind of semi-autonomous, autonomous mo autonomous way.

Maybe a human in the loop as well. That's important. I think some situations, guardrails I think are important.

You know, we make sure that we don't go off and start looking at things. We, we don't want to safety. All of these things are there being addressed.

So I I, again, great potential there. And I was very impressed with it, with a little bit of time that I spent just building something that previously I had no experience with, but even getting it to work, coming up with some nice solutions that I thought, you know, this looks good. It, it looks genuine.

It looks as though it would save me a lot of time. And then I can see kind of potential for this in lots of other areas as well. Uh, but again, uh, I think we're just at the beginning of that journey.

Uh, and again, lots of things to think about in terms of safety and guardrails and, and those kind of things, which I think the vendors are kind of addressing. Uh, and we'll get there, but we're at the early stages still. I think Some people say, then we're gonna build more software in the next two years than we have in the last decade.

Does that sound like hyperbole to you, or do you think we might actually do that? I, I think we might actually do that. My, I think for this reason, um, I'm guessing that, you know, with the development of these cool tools that are coming up and the integration of ai with that again, uh, perhaps the role of the developer and DevOps itself will change significantly.

Maybe we, you know, the boilerplate code, the plumbing, the low level stuff will be handled by the ai. It will give us enough of an infrastructure and outline of what we need to do. And then we as developers, humans in the loop add the finishing touches to that and improve upon what's been delivered.

Ultimately, at some point in the future, probably long after I've retired, uh, I'm guessing, you know, you'll have software that is good enough, you just described what you want to it, it'll go away and build it for you. And it will probably be, you know, 95% or better, uh, in terms of delivering exactly what you're looking for at some point in the far future, probably a hundred percent, but, you know, probably long after I've retired. But I think, so, yeah, I, I agree with that.

Assert. I think we are gonna see a lot more code simply because I think the software will generate most of that for us. And you know, what software is like, I mean, it will tend to build everything for us.

Um, and then people are gonna get very adventurous. They're gonna want more and more, you know, and if the, if the technology can deliver, we are gonna want more capability and more features, more functions. And that's where I think that the bulk of the code is gonna come.

So how often Do you think historically developers might have had some sort of idea or notion and then they thought about it for a minute and they're like, no, by the time I pick up all the development environments and get all the things in place to do that, I'll just go back to sitting on the couch. Um, great point. I, I, I, if I look back at my career in the early days, I, I think it was very challenging, you know, as a, as a kind of a fresh graduate in my first IT job, oh, uh, you know, it was the worst eight months of my life, you know, first eight months.

But, uh, you know, having to pick up and learn stuff was the key to success, simply because no one is going to come to you and help you. It's, you have to be a self-starter. So if you wanna advance in progress in life, you want to learn and improve your skills.

And ultimately, you know, if you want to get that better job, get a higher pay, whatever it is that motivates you, you've got to learn stuff. And so I think there comes a point in time where all developers will feel that, yes, you know, there, there's a probably half a dozen or maybe a dozen tools that they should focus on simply because that's gonna help them progress. Uh, maybe find that next great opportunity in terms of the role they're looking for, or, you know, simply upskilling to find future, you know, opportunities that come along.

It, it's challenging though, because if you look at the breadth of tools available today, it's mind boggling. No one's got the skill or the time to learn everything, but there is always a bunch of things that you think to yourself, well, you know, learn a good idea. For example, learn a good programming language that's popular today.

Uh, learn some other sort of, you know, AI tools, uh, and so on. You know, you could, uh, put together a reasonable list of these and then, you know, keep your eyes in ears open, read the press, uh, attend shows like yours, for example. You have awesome people that come along and tell you that you, this is what you should focus on because this is hot.

Um, so I, I think it, it, there, there's always been that case, and again, if I put my developer hat on, I used to be lazy, right? So I used to say, well, look, you know, I, there's stuff I don't wanna learn, but I thought to myself, well, look, no one's gonna do it for me. I've gotta go out and pick this up myself simply because it's gonna, you know, it's gonna use useful for my job, I need to know it.

And if it's gonna save me time, that's very useful for me as well. So I, I think developers, you know, psychologically, mentally, there, there is a certain way that you tend to think, but ultimately, if you want to progress in life, I think most developers will go down that route. They'll, they'll pick a bunch of tools and they'll learn them and they'll be very skillful at them.

There's a debate out there that says, is AI gonna help more of the senior developers who therefore won't need as many entry level folks? And maybe the bar for entry will be higher versus another argument that says, well, the entry level folks will get smarter faster, and we won't need as many of the senior level specialists. So which end of that spectrum are you on?

Yeah, that's a great point, Mike. Uh, I, I, I would kind of sit in the middle on, on the fence, if you can put it that way, simply because think of it this way, look, the senior developers, right? You're kind of level 10 or whatever they are in Google.

These guys are wizards. They know their stuff inside out, when they look at a problem, they know exactly where to go and the solution to that, and then they've got time to help other people as well. On the other hand, if you look at the junior developers, it takes them a little bit of time to get up to speed, you know, maybe 12 to 18 months before they are really able to grasp and take hold of a problem.

With the advent of ai, that process, that time can be shrunk. They can now become much more productive, faster. Uh, and again, of course, you know, realizing that AI is not the be all in the end, or it's not the silver bullet, but it's gonna help short term, medium term, long term.

Of course, like I said, probably long after I've retired, maybe it's gonna solve all our problems. I don't know. Um, but I think at both ends of that spectrum, it's gonna be helpful, uh, simply because the, the wizards, those level 10 kind of coders, they are gonna focus much more and their skills are gonna be still needed.

On the other hand, the AI is really gonna, I think, upskill the junior developers and bring them up to speed much faster because it's gonna help them with all the mundane and, and the sort of boring tasks, if I can put it that way. So they again, get a chance to focus on the interesting kind of work Folks back in the day, um, mark Twain once observed that, um, he hated writing, but he loved having written. I'm pretty sure the same thing.

Imp applies to Cokes. Hey Achman, thanks for being on the show. My pleasure, mark.

Thank you for having me, and thank you. It was great to see you again. Thank you.

So, and Back to you guys and see.