When code is cheap, what becomes valuable? The new human advantage in the AI era

As AI makes writing code faster, cheaper, and more accessible than ever, the real question is no longer how quickly we can build, but what truly creates value. In this episode, we explore how software development is changing when code becomes abundant, why human judgment matters more than ever, and which skills will separate great technologists from the crowd in the age of AI.

[Henri] (0:03 - 0:07)

Then at some point you start to think like, is it valuable enough to pay for someone else?

[Pinja] (0:12 - 1:10)

Welcome to the DevOps Sauna, the podcast where we deep dive into the world of DevOps, platform engineering, security, and more as we explore the future of development. Join us as we dive into the heart of DevOps, one story at a time. Whether you're a seasoned practitioner or only starting your DevOps journey, we're happy to welcome you into the DevOps Sauna.

Hello and welcome back to the DevOps Sauna. With AI now being a commodity in software creation and agentic solutions, we really need to think about how we create the value and where we are creating the value because the implementation is progressing very much in that area. But organizations actually haven't been able to get that AI to transform how the work is actually done.

So here today with me to talk about where we actually create the value now that AI is changing so much is Henri Hämäläinen, the Chief Product and Partner Officer from Eficode. Moi Henkka.

[Henri] (1:10 - 1:12)

Hello, moi, and nice to be here, Pinja.

[Pinja] (1:13 - 1:37)

Good to have you here. This is not the first time Henkka and I are talking about how value is being created. I think we've done some webinars previously, so it's very natural that I continue this discussion with you specifically on this, because the code creation with AI assistants actually coming into play, that is not the scarcity anymore.

That's not the scarce resource anymore and the bottleneck. I think we've changed that into somewhere else. Is that correct?

[Henri] (1:38 - 2:18)

Yeah. Yeah, I think that's, of course, there's some time for all of the AI coding and the AI in DevOps to come to reality, but let's kind of really envision the future a bit more here now that we get forward on the thinking. But the whole ideology basically is that when coding and the whole software creation becomes much, much, much, much faster, then it kind of challenges us to think about what really is the value and how the value is created.

And I think this is such an interesting topic that it's nice to kind of dig deeper on this one overall.

[Pinja] (2:19 - 2:57)

I think we're going to get a little bit philosophical as well on how software is actually created and what matters and what is the competitive edge nowadays, because I dare to say that whenever there is a technological shift, whatever it is, it changes what the scarce resource now is. And when we were planning this, I took the example of basically getting Spotify and how we play music, because now getting the music was no longer the bottleneck, but it's actually our own attention span on what we're using that and how we consume the music. So the value creation part is shifted totally with that technological shift when Spotify was actually implemented and rolled out to people.

[Henri] (2:57 - 3:51)

Yeah. We've already seen this with the first signs out there that the coding is getting so much faster that the constraints start to be either on testing or then what we've seen already is that security and compliance become the constraints on the delivery side. And there's many others, like with one customer, it was a test environment capacity, which sounds funny, but just the thing that we don't want to invest so much on the test environments that we would actually be able to test all of the things at the same time that the software development can already kind of create the features.

And I think this kind of brings me all the time or brings us all the time closer to what is the value of the whole software and product development. I think that's the super interesting discussion. I think we've had it for many, many years, but this changes the discussion a bit more now.

[Pinja] (3:51 - 4:16)

It does. And because we have previously measured efficiency through what is the bottleneck, and that has been the delivery part previously. And maybe previously we also had to think about more on what is the impact and the effect on the next solutions with the code that we create right now, because this has also had an impact on how fast we can build the code.

But I'm actually questioning this. Is this still so? What is changing with this one?

[Henri] (4:16 - 5:29)

Yeah. So maybe I'll go a bit philosophical soon, but to answer first to your question is that I do believe so, that if we look at the speed of development on the LLMs and overall on the AI tooling, what was happening within two years, we've made such an enormous leaps on the code creation overall. And if, and when we continue the same speed of development, it really very quickly means that coding won't be a bottleneck at all.

And we are seeing this already now in many cases, that if we talk about coding being actually the lines of code creation, then of course, if we extend it to a larger development chain, including the requirement definition and talking about the architecture and thinking about the architecture and documenting and testing and doing the code reviews, of course, then the question is a bit different. But in any case, we are already seeing that the bottlenecks are very much changing now on where the constraints really are.

[Pinja] (5:30 - 6:00)

Yeah. And if we think about the value creation, I'm going to go to the value creation and what is the equation basically where the value is created? Because I know that Henkka you've been thinking about this for many years now and now with AI being into play.

So we need to think of the factors, right? So there's the product value, obviously, and we already talked about the effect on the next solutions that we're building, but previously we talked more about the effort and the work required and how much calendar time is actually needed. So is that where we get the value for a feature for a product?

[Henri] (6:00 - 6:48)

Yeah. The funny thing about this is that I did the first version of this equation of product development efficiency more than 10 years ago, almost 15 years ago. I defined this product value plus the effect on the next solution. So just to be very clear on this, why the effect on the next solution in the current world and the old world is so important is that whenever you did something, you had to make sure that the next time you add software on top of the existing stack, it wouldn't be a lot harder.

It always is a bit harder because you have to test more, there's more complexity. So that's the reason the effect on the next solution is at the top of the equation.

[Pinja] (6:48 - 6:52)

So it's basically building on not creating technical debt as you go along, pretty much.

[Henri] (6:53 - 7:38)

Exactly, exactly. Because there's always some technical debt you create because the complexity just increases when you add rows of code and new features. So it always adds technical complexity.

It's more about how much you add it. But then on the bottom of the equation, it's always about the effort, how much effort you've put in to create the product value and the less effect on the next solutions. And then that has been divided by the calendar time.

And I think we should go to the calendar time discussion a bit later because it's a very interesting part in any case. But what I see is that if you think about productivity, it's like value divided by effort. And now the effort starts to go towards zero.

[Pinja] (7:38 - 7:39)

It does.

[Henri] (7:40 - 7:57)

So then it kind of changes the equation a lot. Then it doesn't immediately mean that like effectiveness is just that you just put tons and tons of features out. I did say to you this, that like, would you like your phone to have like seven new features a day?

[Pinja] (7:57 - 8:00)

And you saw the horrified face that I made.

[Henri] (8:01 - 8:01)

Exactly.

[Pinja] (8:02 - 8:38)

And I still do. It's not just about that. It's not that simple that we just increase the volume, right?

And this is, we're not going to go into what value and what outcome-based work looks like so much, but it's the lines of code has been one of the measurements, unfortunately, still until very recently. And some even now go into tokens used as a measurement to go to the extremes. But we do need to think about what is the product value itself and the effect on the next evolution, not just the calendar time and the effort and work required being pushed down.

[Henri] (8:38 - 10:11)

And I think now it's a good time to be a bit philosophical here, because like the thing is that, if you think about any efficiency measures, it's always measured by the, what's the limiting factor. That's how it has always been. Like if you are like, whatever you do, you're in a fast food restaurant, you're normally just thinking about how many hamburgers you get out in an hour.

But what if it would be unlimited that like there wouldn't be any limitations. I don't think that would either be like what's happening now with the software world. Like then there would be another thing you would start to measure.

You would start to measure how many people you can actually serve at the same time, which kind of was happening in the, with the self-service in the fast food chains. But like what you measure in the efficiency starts to change also compared to your bottleneck. And I think this is the, by far the largest change on the, when you think about the efficiency in software development, that's going to happen now in the, in the very soon that you have to think about measuring different things.

It's not about how many features you get out, how many hours you put in, but about what real customer value you actually are able to bring. And I just want to say that, like, I think this is something that some people say that like, this is what we've always been measuring. But my answer to this is that that's not true.

All of the companies always measure how efficient they are by how busy their people are.

[Pinja] (10:11 - 11:24)

How busy their people are, how many new features they're shipping out. Fine. There are the NPS scores.

Yeah. Some people are measuring that, but it is always measured through that lens. Of course the resources have been scarce.

Like the couple of companies that I remember from a couple of past years that I've been consulting with, they were really thinking about how we can make our development part of the organization's R&D more effective. And then there's a product part of the organization's thinking, why aren't they faster? What is going on?

Many, many things, obviously, but that's just to illustrate where that bottleneck was and what the organization itself was thinking, where the value is being created and now the speed of the development was previously the advantage. Obviously, with the feedback loops, it still is. Obviously, we need the feedback loops, but now that we're actually moving into this much faster world, we need to figure out what does it take from an organization, from a human being, from everybody to actually be able to implement and take on these new things.

But are we solving the right problems? I think it would be one of the questions, because you said product value being at the top of the equation for the value creation. How do we get to that point if we're shifting towards that thinking?

[Henri] (11:24 - 12:38)

Yeah, I think that's a super good question. And then I'll just give a short sidetrack to answer you. It's kind of been a super good thing that there has been a limitation of how many features you can implement, because then the whole, the product management, the business people have had to choose.

There has always been with all the companies, we've worked together with some of the companies, which has like years and years of backlog. And now what if they wouldn't have the limitation, they would actually like all of the funny sounding ideas, they would actually get in immediately. So without really having to compare what is valuable to the customer, I think this goes back to exactly to your question about what then is valuable, because it really boils back down to where the customer and the company value actually comes.

And why I mean company value is that it shouldn't only be valuable to the customer, but the thing that is valuable to the customer should also bring something in for the company you are actually doing the work for.

[Pinja] (12:38 - 13:01)

There's so much value in organizations and product organizations actually doing the work and having a large idea backlog. It's not whether we're saying ,with all the respect and with all the love that we have, we say that it is a good thing that not all those ideas came to life. But at the same time, it is very important that the ideation process is there.

But in order to be able to prioritize, there needs to be a scarcity of something.

[Henri] (13:02 - 13:02)

Exactly.

[Pinja] (13:02 - 13:04)

So the backlog management is one thing.

[Henri] (13:04 - 14:40)

Yeah. And that kind of makes it even more important to be able to do this. I hate to say lean startup, because I think that was such a thing in the past that everything should go via the same methodology.

But I think there are some good learnings on those times that you actually kind of try with the very... And I think it somehow can now come much more to life. Some of the business people actually can nowadays use the AI tooling to actually create the lean startup type of new features to get the real feedback from the customer that would actually bring value in.

So I think that's going to be a super big thing in the future is that the whole experimentation process will become one of the most important things for any of the companies. And for a different reason than in the past, in the past it has been because there was the scarcity of development resources. But now the scarcity is in the ecosystem adoption rate, or the customer adoption rate, or even the company adoption rate overall.

But it's still the same, or it's even much worse now because there's no limitations. So you have to limit yourself. You are in a...

Now this is going to be a funny sounding thing, but you get to like a Las Vegas buffet, right? Where there's like anything you can eat, right? And if that's every day, there's anything you can eat, you have to kind of control yourself a bit more.

[Pinja] (14:41 - 15:46)

There's a funny shift, because now the new part of the organization is waiting for the part that was actually waiting for the others previously. But it's... If we think of the ecosystem for a bit, and you mentioned the customers, we have the organization's ability itself.

We have partners, for example. They might be partnered with some other companies, technology partners, etc, etc. But I talked to many, many customers about this, because they have the biggest fear of missing out.

They have the FOMO about not being able to do something that Frontier is doing right now with AI, because you go on LinkedIn and you see what the best of the best are doing, what the innovators are doing, because they're the first ones with the adoption curve going forward with these things. And then we need to think about the limitations. For example, there was a conversation a couple of months ago with a financial institution.

It's not like they, with their ecosystem, they cannot do what a startup can do, or somebody who's personally trying to invest their own time and develop new solutions and trying to figure out what they can do with agent AI and orchestration, for example, is not as simple because of the ecosystem ability to implement.

[Henri] (15:47 - 17:42)

Yeah. It's a difficult game in the sense that some part of the FOMO is real, that you have to be acting fast enough that you still can match most of the competition on speed. But then at some point, there will be the tipping point where speed isn't anymore an issue.

And for most organizations, we are still in the very early phases where speed is still an issue because there are companies who still have very bad basic CI/CD. They don't have test automation in place. So there's a lot to do with AI or just basic DevOps automation to actually get up to speed.

But then after a certain point, it kind of turns around. And then it's not anymore about the speed, but it's exactly about where really your value is. I've used this example many times.

If I would take our CTO and maybe some of our senior developers, I'm quite sure in two weeks, they would be able to do a really good banking solution, right? Yeah, I bet. They would be able to do a really good banking solution with current AI tools, but still they couldn't run a bank and they wouldn't get any customers and they wouldn't be any kind of a competition to the actual banks.

But then it might be another bank that actually could implement some of the ways to gain some market there. So that's the reason you have to keep up with your competition in some sense. But I think the difficulty of the whole question is missing out on what and whom.

[Pinja] (17:42 - 18:19)

Yeah, because there are these industry-specific limitations, obviously. Because if we think of, in theory, in an ideal world, there absolutely are no limitations to the speed, to the capacity. But if you think, I'm going back to your reference to the fast food restaurant and thinking about, in my mind, I have this picture of, in a cartoon, there is just a kitchen spewing out hamburgers and the waiting room, the restaurant itself actually filling up with those.

That's not what we want. It's absolutely not what we want and it's not going to get to the customers. And there are these limitations and the speed is not the competitive advantage in every single industry.

[Henri] (18:19 - 19:10)

Yeah. Or there is an element of speed that gives a competitive advantage. If you go to a fast food restaurant and you have to wait 15 minutes to get your meal, that's not good.

But immediately when you click done in the automatic machine, it would immediately appear the meal to you. That would almost be the opposite. You would almost feel that there's no way this actually was made for me and with good quality.

So there's an element of three to 10 minutes or three to seven minutes is the speed. And I'm quite sure McDonald's or someone has made studies on what this actually is. But I think everybody will get the point that there's a limitation that speed will bring value and then it's something else that starts to bring additional value.

[Pinja] (19:11 - 19:51)

And as you said before, you need to keep up with your competitors on this. And there's one customer that I'm talking to at the moment and we're building up their capabilities, especially around AI. And we're doing this by actually benchmarking them against their peers.

And we're looking into where it makes sense for this company to put in the effort and how far they should do the effort. Basically looking at where's the plateau for them in terms of adding more, this is the worst thing to say, but more AI into their workflows, but to make it faster. Because after a certain point, it's not going to give them the advantage anymore.

And they're just going to be spewing money on that and not getting the value back.

[Henri] (19:51 - 21:00)

Yeah. And I think we're kind of going around on this that the constraint will change and there will always be a constraint. And it somehow is a funny thing that I think we are approaching in a couple of years. The constraint will be, it sounds a bit theoretical, but the ecosystem's ability to take new features.

And I think the big thing is that when we reach the point that this ecosystem, which we mean the customers and the partner network, and maybe their own organization, when we reach that that will be the limiting factor, the competition changes totally around. Then it's all about if your feature actually makes sense, is the customer experience, is the user experience, is the network, I mean the people or that type of a network, is all of that in place. And I think that's going to be like a huge shift for many organizations.

[Pinja] (21:00 - 21:33)

And I want to tie this back to the value creation equation basically. Where do we get it? So for example, calendar time, it is going down.

Effort and work required, definitely there is an impact with AI solutions. Effect on the next solutions, as you said, there's always going to be some kind of technical debt that needs to be handled. But even with that, AI is going to be helping that along the way.

So the product value might be the thing. And as you said, how is the look and feel? What is the user experience?

What is the ecosystem like? Is it actually answering the customer's problems? Where do we get that understanding?

[Henri] (21:34 - 23:00)

Yeah. And I think it's going to be a much more brand name game also. It's not about who has the best solution, but who has the solution everybody else says it's the best.

It feels again funny, but it's about like, are your friends there? Who gets the people to be there? And then we go back to what we started to talk about 15 or 20 years ago about the whole influencer market.

And I think many companies have already realized that you have to win the influencer markets. And then when you are in a B2B, so okay, what's your influencer market on the B2B? But it's going to be a very, very funny game.

And I'm at least super happy to be in this role to actually be able to talk with so many different companies and they're thinking about it. But it's going to be at the same time, the whole internal side of the organization, meaning that how the machine works will dramatically change. And there's tons of transfer, the biggest transformation ever going to happen.

But in some sense, that's not enough because then you still have to still change the whole like, okay, so what is the business value we bring to the ecosystem? And how do we actually capture the market? And how do we get the customer?

How do we create the competitive advantage? So it's at the same time like a scary, but very exciting time.

[Pinja] (23:00 - 23:28)

So many things are changing so fast. But I think just to tie this all together a little bit, just if we think of these principles of understanding your ecosystem, trying to figure out what is it that we need to build and not building everything, we still do not encourage you to build your whole backlog. There must be amazing, great, wonderful ideas in there, but please do not build your whole backlog.

But I think this is going to revolutionize the cost point of software. And definitely, as we said, the competitive advantage is going to shift.

[Henri] (23:29 - 25:41)

Yeah, yeah, yeah, definitely. And that's going to like on top of everything we've discussed already, like it changed the whole like price points of software, like who's gonna pay for stuff that you can actually ask AI to do. So then like, of course, on the personal commercial side, I mean, like on the apps business, it's already visible that like, hey, for the very basic apps, no one is really not paying anything because you get an AI to do many of those things already for you.

I think that one of the things on the like B2C business is that this maybe it finally is time to the micro payments era where you actually someone has made the simple software with AI to actually do a task really well, you're willing to pay like a 50 cent a month for it. So maybe that but especially then in the B2B world, that's gonna be even more difficult. Like we're seeing it already, many of the SaaS companies have been trying to kind of raise the prices a lot because then I guess they know that it will go down. It has to go down because the cost of creating software is really gonna be a fraction.

It's not just like some percentages, but it's gonna be a fraction of things like one of my former colleagues, for example, I don't want to maybe go into the specifics. But like there was software we used internally. And he used a weekend to actually recreate the software just for fun with AI.

And within a weekend, he said that he believes he has the same level of feature set than in the software we were using internally. Of course, there's another value on the buying, like paying for that software, because someone takes care of the maintenance of it. And we already migrated all of the information there.

But for many things, this kind of a change is the thing that like if you're paying a million for something a year, and then at the same time, someone can recreate it within a week or two, like then at some point, you start to think like, is it valuable enough to pay for someone else?

[Pinja] (25:41 - 26:08)

Is it valuable enough? And going back to the ecosystem thinking, as I said before, if our CTO and a bunch of senior consultants could, for example, whip out a banking solution in a couple of weeks time that would have the exact features, the exact same look and feel. But would we get people to transfer into that?

Because it would not be a bank, it would not be a financial institution yet as it is, because we cannot deal with change very well. Can we?

[Henri] (26:09 - 27:35)

Exactly. Yeah. Yeah, this now we go very philosophical back again.

But like, it's the human's ability or almost like human need is to kind of seek for stability, right? It goes back to like, you don't want your mobile phone software to change seven times a day. You don't even want it to change once a week, most probably.

So there's some kind of stability you want there to be. It's the same as like, like with any software, you really want to be like, knowing your basics, your car should work the same way, like almost every day, right? And so like, now we go like, very, very meta level philosophical, but like the limiting factor for the digitalization will be the human's ability to tolerate change.

And this will be like we start to compete on the same scarcity, which will be human's ability to tolerate change. So it's, it's gonna be, I know some listeners are like, what the heck you guys are talking about, but like, very philosophically, this is that. And I think it's already happened.

Like, like, I think you used the example of Spotify. It's, it's still about like, do you use time for Spotify? Or do you use time for audio books?

Or do you use time for YouTube? Or do you use time to think?

[Pinja] (27:35 - 27:56)

Like, where do I put my attention? That's the thing because it has become so easy to just open my phone and open up Spotify or Apple Music, whatever I use. And then there's a competing app.

There is the, as I say, there is the, there's audio books. Do I go on YouTube? Or do I do something else completely?

Or go old school and take a book, for example.

[Henri] (27:56 - 27:59)

But it's, again, I thought you were saying old school and Facebook.

[Pinja] (28:00 - 28:16)

We go even further. But that's, that's exactly the thing that now the scarce thing is indeed our attention is our ability to take on these changes. So not much has changed in the human brain in the past dozens of thousands of years, to be honest, in this trade.

[Henri] (28:16 - 28:17)

But yeah, exactly.

[Pinja] (28:18 - 28:22)

Let's see how the AI is going to do with us. And not how we do with AI.

[Henri] (28:23 - 30:09)

Yeah, exactly. But I think we've gone quite round tour. But like, all in all, the point is that it will be, like going on, the largest change any organization has ever seen.

And the fundamentals of each organization will change. Of course, the change is a bit slower than everybody anticipates. I remember two years ago, I went to see some leadership teams of companies and they were immediately saying that like, hey, now we will do 70% faster coding.

And sure, but like, so what? So what changes? And I think that's what we've been kind of discussing today. Is that like the fundamental ideology of what's your measure?

What does efficiency actually mean? What's your competitive edge? And those are stuff that I would kind of encourage all of the organizations to kind of really think about.

What's the value you bring to the world? And I have to say one thing before I'll let you jump in is that like, I've gotten most calls from leaders, like C-level leaders in the past like year that I have for my career ever on the about like, these leaders are like calling me with the message of like, so what should we do? And that hasn't ever been in the past, it has always been we have a problem, could you help?

And I've been happy to help. But now the calls have been like that, so what should we do? Where should we focus?

And that's amazing. Like it's no one knows really. And I think it goes back to like understanding the value you bring to the world.

And I think that's where it starts from.

[Pinja] (30:09 - 30:28)

I think that's a good point to end our conversation with that note, because I think we do need to figure that out. Organizations need to, I'm not gonna be as harsh and say, take a good look in the mirror and think about what is that you're bringing in. But I think that's a good observation point and self reflection for everybody.

Thank you Henkka so much for joining me here today. I had a blast.

[Henri] (30:29 - 30:29)

Thank you.

[Pinja] (30:29 - 30:39)

And thank you everybody for tuning in. And we'll see you in the sauna next time. We'll now tell you a little bit about who we are.

[Henri] (30:39 - 30:44)

Hey, I'm Henri Hämäläinen, Chief Product and Partner Officer at Eficode.

[Pinja] (30:46 - 31:00)

I'm Pinja Kujala. I specialize in agile and portfolio management topics at Eficode. Thanks for tuning in.

We'll catch you next time. And remember, if you like what you hear, please like, rate and subscribe on your favorite podcast platform. It means the world to us.