The Future of Software Recap

[Pinja] (0:05 - 0:12)

The future of software development is AI-driven. So, what are these potential future scenarios here?

[Marc] (0:16 - 0:26)

Welcome to the DevOps Sauna, the podcast where we dive deep into the world where software, security, and platform engineering converge with the principles of DevOps.

[Darren] (0:26 - 0:55)

Enjoy some steam, and let's forge some new ideas together. Welcome back to the DevOps Sauna. Last week, we hosted The Future of Software Conference in Helsinki, and we have a couple of participants from the day to talk about how the event went.

We're sat here with Marc. Hello, everybody. And we have Pinja.

[Pinja] (0:55 - 0:56)

Hey there.

[Darren] (0:56 - 1:24)

So, it's kind of our mission at Eficode to support the future of software development. So this event has been kind of baked into the DNA of Eficode in the past, and it returned to the stage last week with about 300 people in attendance. And it kicked off by having Anne Lise Kjaer, the Futurist, talking about basically the future of software and the future of the human interface.

How did that go?

[Pinja] (1:24 - 1:39)

Anne Lise. She's the Futurist and her ideas and thoughts, how to combine technology with empathy that really speaks to everybody who listens to her and really spoke to the audience last Tuesday.

[Marc] (1:39 - 3:07)

It all comes from, she publishes this Trend Atlas, and the one that is current at the moment is the Trend Atlas of 2035+, and maintains this from time to time. And at the very top are two things that I like very much, quality of life and universal awareness. And, you know, one of the reasons I had originally come to Eficode was psychological safety was put as a very high importance.

I think it's never been more important than it is today. And connected with that, I think one of the most important things when we talk about why we do the work that we do, it's about situational awareness, or as she put it, this universal awareness, understanding that, you know, it all comes from, you know, not only how we work together and how art and science come together and how interconnected the world is, but also there's this neat kind of Buddhist principle, the power of one, which I think is really, really cool. And it was a wonderful experience to be in the same space with someone like this who really didn't have any barriers, and she really didn't have any boundaries either. But just to be able to kind of transcend everything, lifestyle, technology, environment, legislation, economics, politics, better world, consumption, you know, to name a few of the topics that she touched on.

It's astonishing that there are people that can can look at the world so holistically.

[Pinja] (3:08 - 3:36)

To me, what really resonated with me was the integration of the human-centric element and the sustainable principles and that into the strategic decision making of, I think she talks about the 4P model where people, planet, purpose and profit all come together. And she ended the keynote with the phrase, we're all futurists now. It was a perfect mitroph for that talk.

[Darren] (3:36 - 4:14)

That sounds interesting. It's kind of crucial with how AI keeps coming in and in a way, augmenting processes and in other directions, just completely derailing them and like causing these issues of reduced workloads and how that affects people. So it's kind of interesting to see that be brought onto the stage in the highlight.

I think AI was kind of a key theme for, as we would expect in 2024 for the future of software. I think one of the things that we had was about AI only being as smart as the data you give it.

[Pinja] (4:14 - 4:32)

Yes, we had Eero Jyske from Apheris talking about this topic and the data quality. So that the whole distribution of data also related to the successful implementation of the AI in this day and age, and of course, the coming years as well.

[Marc] (4:32 - 5:23)

And the neat thing here being if we acknowledge that our data has bias and if we understand that we're not really generating anything new with AI, but what we're doing is we're enabling humans to have better information and better access to that information so that they can make more informed decisions, preferably more accurately. That's a really, really important frame for all of these AI discussions that we're making today. And I still occasionally hear someone say that, well, you know, the AI is only this or the AI is only that.

But still, if you use it very mindfully and understand that the data that you are putting in has a great deal of value, but also needs to be qualified in its biases and the value of the input.

[Pinja] (5:23 - 5:37)

I've been very happy to see this change in the recent years now, people understanding the need to improve the quality of data to be used for this. It is not a magic machine, right? So we need to improve the other areas around it.

[Darren] (5:38 - 6:12)

I feel like that I haven't seen enough of that, to be honest. Everyone's always talking about what is coming out of AI and what AI can do for us and how AI will make everything better. And very few people are talking about what goes into AI and what AI needs and what AI does and creates as a byproduct.

Like everything about the inroads to me seems underrepresented compared to the influx of people saying AI is great. Look at all these cool things it can do. So, I think it's refreshing to have such a talk on the agenda.

[Marc] (6:12 - 7:01)

It reminds me, I just, I want to put a little plug in here that the Eficode Demo booth showing what we're doing with AI was, it was kind of a showstopper practically of our, of our own sponsored show, whereby taking all the information that you have from your contracts to your confluence or your specifications, your source code itself, and being able to present intelligent options to humans that can be used as communication up and down the chain or, you know, across the value stream, whichever direction you want to look at, to make more clear requirements for what people are going to do, which then also results in better code. That was something that a lot of people were really excited about.

And I'm really happy to see what Eficode is doing in this space.

[Darren] (7:01 - 7:26)

Another topic that I think came up and has been coming up over the past few years is this idea of a third-party landscape. So in-housing everything, in-housing development is occasionally more often not practical. So if we're looking at the inclusion of SaaS partners, I think, uh, Simo Haakana from Barona had some things to say on that particular subject.

[Pinja] (7:26 - 7:38)

Yeah, we have, if we think of the whole, whether the build or to buy, is it off the shell of the service? I think this is a very important topic at the moment to consider.

[Marc] (7:38 - 8:17)

The neat thing here to me as well as ERP is a dinosaur that many companies have the core of their operations, you know, fossilizing around. And what Barona did is they took the, the software pragmatist approach to their ERP and they tried to keep it moving and fluid. So they, if you're not updating your software, not only is it rotting, it's fossilizing folks.

So, being able to take the ERP and keep it moving and then realize that, wow, they could actually save a lot by going to a SaaS solution was really, really cool.

[Pinja] (8:17 - 8:35)

And, and particularly what Simo highlighted in his talk and was of particular interest was the basic building blocks, what they define a migrated, including the SSO in here. So I think that was the whole view of this is very interesting what he presented in his talk.

[Darren] (8:35 - 9:22)

And speaking of things that tend to like fossilize and stay in one place, we actually had this interesting presentation from the DVV, which, for those who don't know, is the like digital population register in Finland about digital by default because those who haven't lived outside of Finland might not know how bureaucratic and paper-based a lot of things still are. So, coming to Finland and using like bank credentials, strong authentication online is so immediately useful once it's available that it's so good to hear people taking these systems that were originally kind of static. And as you say, Marc, rotting and problematic and turning them into these digital by default systems.

[Marc] (9:23 - 10:17)

That's one of the interesting things. When I first came to Finland 19 winters ago now, I was astonished by the level of digital integration in things, you know, like, like the banking and the strong authentication and digital services everywhere. And everything seemed really, really kind of centralized.

And back then, even the trains ran on time. And then with the changes that we've had and a lot of things, it's not exactly where it used to be. And kind of looking at home and abroad, I do realize that, okay, it's not a unique to Finland problem that the state of I'll use towards customer service or user experience is, is really not that high in a lot of places at the moment, but seeing that, that Finland is pushing forward in a lot of these.

You know, if you think about the number of construction cranes around as an indication of your economy, then we're in pretty good shape here still.

[Pinja] (10:17 - 10:46)

And in this case, especially, it is the accessibility of the services. Yes, we do not have to go, to an agency physically and visit that, but then there is the accessibility for the elderly, the disabled, for example, how to use that and how that has been taken into consideration because DVV is a service that everybody has to have access in Finland. So how can we ensure that is, that has been really interesting to hear from, from this talk as well.

[Darren] (10:46 - 11:20)

But then, maybe moving on to something more modern, let's talk about Sensible 4. And I love Sensible 4, or the former Sensible 4 for doing this talk because it kind of takes a lot to come up on stage and talk about failing. Like Sensible 4, they were a great idea that failed in the funding, and to have Harri Santamala come up on stage and actually talk about everything about the, from inception to failure of that company, was kind of impressive, I would say.

[Pinja] (11:20 - 12:08)

It was. And Harri spoke very candidly about their learnings and, and what, what was his analysis of, of the whole journey? Because they, they tried a couple of things.

They had the financing rounds and everything was looking great. They tried, tried different kinds of organizational structures to overcome the silos as well. So again, these kinds of stories, when we hear also from the, from the, the companies who actually didn't make it, can we learn from that?

What can we take into our own world and own doing and own work? And if we think of automotive industry, it is a very traditional. One, right?

So there is bringing this in and trying to get the funding for a service like this and a product like this has been, of course, a struggle.

[Marc] (12:08 - 13:28)

I was the one that called Harri and asked if he would speak at this conference. And he reminded me of the day we met. Metropolia University in, in Finland, in Helsinki has a garage down by the Harbor, by the LOLU sauna, actually.

And the day that I met Harri, he was standing over top of a Jaguar E-type and they were building electric car infrastructure into the Jaguar. And I was thinking to myself, Jaguar and electrics, oh boy. But if you rip them all out and replace them with something that's modern and from any country, any other country, then you're in a lot better shape.

But they were actually, you know, building electric cars. And this is more than a dozen years ago, but what a man to just always have a bright attitude and push forward on leading ideas over and over and the, you know, the failure, it's not an option. It's an outcome and it's not a bad outcome.

You don't learn from positive experiences as well, but astonishing story, astonishing company, and work. And, you know, all those people that worked on that benefited greatly from the opportunity that they had to do it.

[Darren] (13:29 - 14:12)

There was also a talk at the event, which to me was kind of an interesting clashing of future and past where we had Bruno Almeida from Fortum talking about bridging the gap between IT infrastructure, platform engineering, when like IOT infrastructure, this kind of stuff is often quite outdated, especially we know Fortam works in critical infrastructure and then platform engineering, which is basically the newest buzzword around how we build these tools in DevOps spaces and how we provide this platform. So that was, to me, kind of an interesting topic to see these kinds of clashing of worlds.

[Marc] (14:12 - 15:50)

A few takeaways here. First off, Bruno's an absolutely brilliant speaker and presenter and really shows how it's done. He was really smooth and comfortable on the stage.

He told great stories and, you know, showed a lot of experience and authority up there, and I thought it was really, really cool just the style of the thing. But, you know, platform engineering, self-service, thinking in terms of an internal product and, you know, keeping in mind one of the things that, you know, so many, you know, not DevOps teams or IT internal service providers or products, they always kind of forget about the internal customer because they're going to get paid whether the internal customer likes it or not, where in the real world, when we're looking at IT to product type of companies, they really need to have a good user experience. And in our case, platform engineering, we talk about this in terms of Developer Experience as well.

And then to see once again, you know, this large critical infrastructure company building, you know, Developer Experience, high-end platform engineering, it's really, really inspiring. We do a lot of work with customers in similar sectors. And when you look at a different company in the electric industry, for example, that's really struggling with their internal processes and their developer experience and all this kind of stuff, it's wonderful to see somebody do it really well, like Fortam is doing and be able to share that kind of thing and inspire others and basically what I really want people to understand is what is possible from a talk, you know, and that's what talks like this and what Bruno was able to do for our audience.

[Pinja] (15:51 - 16:09)

And this kind of a thing, actually looking into the future and looking at future proofing what they're doing is they're also in, they do service management. So how is that being done? So, and I feel this, these topics are the, one of the most interesting ones from my perspective as well.

[Darren] (16:09 - 16:29)

So, um, one of the talks at the conference kind of confused me, and it's, let's talk about tea, how we use tea strategy to revolutionize the company processes from a couple of speakers from Orange and this title confuses me. Is this some strategy I'm not familiar with, or are they talking about actual tea?

[Pinja] (16:30 - 16:34)

What happens when you put a teabag into hot water, Darren, what happens to the teabag?

[Darren] (16:35 - 16:36)

It gets wet.

[Pinja] (16:36 - 17:44)

But it also spreads the, all the flavors, right? So when it is immersed into the hot water, you get the nice quality tea, don't you? So what Emma and Gaëlle were talking about was the culture of the transformation they had at Orange.

So they started a large-scale transformation, but one of the things that they didn't have the buy-in from their management, because that was one of the questions that the audience had afterwards, after the talk, like, how did you manage to get the buy-in? And they're like, we started with, with no buy-in, we started on our own. So they were able to secure like-minded people, hundreds of them, and literally hundreds of them all across the world, because Orange is a global company and get them to make changes in their own environments and on parts of the organization.

So with those results, they were able to get the larger buy-in from the company management as well. So again, showing the results, showing to the management, like, this is what we did. May we please continue?

And can we get into the other parts of the organization as well?

[Darren] (17:45 - 17:50)

Okay. So it's about the spreading of knowledge and not the dipping of developers into hot water.

[Pinja] (17:51 - 18:03)

No, no, no. We have talked about the Developer Experience and that, again, that is not part of it, but it's the, the environment in which you get in, get the tea bag in, where you get the flavors out.

[Marc] (18:04 - 18:21)

I think in Finland, Vappu, you dump the students, including the engineering students into cold water. So I think that that's, that's a lot more Finnish. And then when, of course, we have, you know, things like polar bear pitch fest.

You can, you can talk as long as you want, but if you're in the cold water, but.

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

It's a very different concept in that sense. But engineering students do it differently than the tea strategy.

[Darren] (18:28 - 18:50)

Okay. Moving on from the traditional tea, let's talk a bit more about the future. There were, I think a couple of talks regarding the development with generative AI, and one of them we had Peter Schneider from QT, I believe.

And he was basically talking about the cross-platform UI development using Generative AI.

[Marc] (18:50 - 21:52)

So, I have to tell the backstory here. I didn't bore the audience with this story, but you're still here listening to this podcast, right? So here we go.

So on February 11th of 2011, I was standing in the Nokia Research Center hall, the big atrium down there in Ruoholahti, where I had made my home since 2006, my work home anyway. And we had made seven or eight products in the platform that was then rebranded as MeeGo and was shared with Intel. And we were waiting for Stephen Elop to speak that morning.

We knew he had 200 Nokia managers in, I think it was London for some time before. And this is after the burning platform memo, which actually I first heard in Nokia context in 2004 by Sorin Yandri-Peterson, who was the North America Director, but now we're really digressing. So Peter showed me on his phone, the press release.

It said that MeeGo, the department I was in, was canceled, and I felt sick to my stomach, and that same day, Tampere Nokia made news because they had all walked out. And I was thinking what wonderful news it would have been. Had I gotten sick in front of a thousand people standing in a hall, waiting for someone to tell them that their life was about to change and how many other people might have gotten sick at the same time on the side of me, but Peter Schneider was the one who showed me the news that day, but let's get to the present and the future.

So QT is an astonishing organization. They have a language of their own, QML. They are a mainstay in automotive among other industries.

They make beautiful UI capabilities. 100 million developers use QML and QT tools on a regular basis. And what they did and what Peter talked about was they, they've got their own language, right?

So they set up a bunch of tasks, and they benchmarked various AI coding assistants and then did some fine-tuning on their models, and ran the same set of tasks. And they looked at how different performance levels were between what many people thought were actually competitive models, some of which were, but also what we noticed was that Copilot from GitHub, from Microsoft, actually uses the Claude 3.5 Sonnet as a preview mode today. That was one of the ones that scored highly.

So things are changing so fast that when you do these types of empirical analysis, as scientific as you can be, it might be changing tomorrow. Having their own language and wanting to know what the effect was of various tools. They set up these tasks and did this benchmarking, and then it's already kind of been changing, I think, since the presentation they did there.

But a lot of this comes back to a couple of principles, keeping ahead on the changes that are going on and understanding how your specific use cases can properly be measured in the context of the world and how it's changing at the moment. Good stuff.

[Pinja] (21:53 - 22:10)

In this context, we very briefly touched upon with the previous talker, we talked about the Developer Experience. So what happens with the developer experience in this case, when the developers are able to do with their own language, their thing and focus on coding, writing the code and being innovative.

[Darren] (22:11 - 22:42)

Kind of interested at the accidental point they've made here of the, well, the difficulty in how rapidly these things change. Like you're saying that the models were changing and the efficiencies were changing, and I think people have a tendency to lean on one AI tool. And it's kind of interesting that the best strategy they then may end up being leaning on multiple or moving in various directions, depending on the models.

So I don't think I've seen research like that before.

[Marc] (22:43 - 22:55)

And for avoidance of doubt, that was my observation based upon Peter's and QT's exceptional work that some of those things are changing already, uh, since, since doing that, but still your point is great.

[Darren] (22:56 - 23:20)

And it kind of leads us into, we had a talk by Richard Crampton from Atlassian about the opportunities in AI adoption. And I think one of the key things he talked about was the AI being a complementary tool for humans. And it's something we've heard before, but I think it can't be overstated.

It's something that has to be repeated until people understand.

[Pinja] (23:21 - 23:59)

It is the use cases that Richard showed us about how to get some of, the simplest tasks supported by AI. So basically, with Atlassian things at the moment and what they promote is that now and in the future, in the near future, actually, you would have a team of people, actual real-life people, but you would have, the AI-powered teammates as well the virtual ones. Maybe the ratio might vary a little bit, but how do you actually use them, they're talking about the Rovo Agents, that Atlassian has introduced this year, for example.

[Darren] (24:00 - 24:17)

Yeah, I believe Atlassian was mentioning their virtual team members. Also at The DEVOPS Conference, they were talking on the same subject. It seems to be something that they're pushing now and it will be interesting to see how they start shaping Team Topologies over the next couple of years.

[Pinja] (24:17 - 24:38)

Yeah, this is Rovo is, is a new thing. And in The DEVOPS Conference where we had Chris Davidson from Atlassian, he also mentioned it a little bit from a different perspective, but also the same ideology that the combination of humans and the virtual team members would actually combine and bring the value through the Atlassian stack.

[Darren] (24:39 - 24:46)

And it all leads into that idea of the combination, not one or the other. You need both in the mix to make it work properly.

[Pinja] (24:47 - 24:47)

Exactly.

[Darren] (24:47 - 25:18)

And actually, I think that's something that AI demonstration at the events has also leaned into that. It may be a pivot from coding to checking, but it's still a very human requirement. Fully agree.

And then I think that the last thing on the event, it was wrapped up as I think they have been before with Marko and Henri Henri Hämäläinen from our side at Eficode, they were talking about some scenarios of the future. What did they come up with?

[Pinja] (25:18 - 25:57)

We had live coding on stage. When showcasing what can be done with, with the power of AI right now, and it's not, perhaps not even tool specific at the moment, but what are the business effects of, of using these tools? So for example, if you get this much faster code, or you can come, you can connect to APIs this much, much faster.

So it's not just that your developers can do work faster, but what are the implications in a much larger scale? Cause the, well, their main message was that the future of software development is AI-driven. So what are these potential future scenarios here?

[Darren] (25:58 - 26:01)

Yeah, it does feel like a tool we're not going to be able to ignore going forward.

[Marc] (26:01 - 28:10)

And I think there's so many kinds of aspects here, but I'll just, I'll applaud the live demos at a big conference type of thing, especially when you have the variability of AI's willingness to play along sometimes. So Modico has been almost a hundred percent in the live coding demonstrations in front of an audience. And it's, it's super exciting, and you just, you can feel the tension in the room when you're waiting to see, is the AI going to deliver working, immediately working software?

And then how is the presenter going to react if it doesn't work the first time? You know, I've, I've seen some, you know, not from Modico, but I've seen like some cooking show style stuff before where, well, had it not worked, this is what it would have looked like. And it's like, ha, we want to see you sweat, and the audience wants you to succeed, but it's also just a great deal of fun to build up the tension there.

And, you know, some of the things that are, that are covered here, you know, building up templates using AI to then have the AI utilize those templates in order to make applications that work like these are React applications working in real-time. We actually saw in The DEVOPS Conference in Scandinavia he was actually using this, in The Future of Software as well. But building an app that allows the audience to scan the QR code and, in real-time, be able to vote and see it on the screen in the browser that he's showing, you know, and building that type of functionality in real-time in front of an audience, really, really cool stuff.

And I would, I would urge people to have a look at this one. This is available on our website as well as the other talks and kind of understand that one of the things demonstrated here is if you set up the right kind of scaffolds or templates for your employees with your developers, then it makes a lot more predictability in how you're going to deliver business value using the AI tools that are available instead of just depending upon them to spit out whatever they do.

[Darren] (28:10 - 28:34)

And as Marc mentioned, these recordings are available on The Future of Software website. You can find it through Google. I think some of the recordings won't be available at the request of the speaker, but for the most part, the recording should be available, and the uploads for them have already started.

So you can go and watch them now. But that's all for the Sauna. I'd like to thank Marc and Pinja for being with me today.

Thank you, Darren.

[Pinja] (28:34 - 28:35)

Thank you.

[Darren] (28:35 - 28:36)

And we will see you next time. Bye.

[Marc] (28:36 - 28:37)

Bye.

[Pinja] (28:38 - 28:39)

Bye.

[Marc] (28:43 - 28:49)

We'll now give our guests an opportunity to introduce themselves and tell you a little bit about who we are.

[Pinja] (28:49 - 28:54)

I'm Pinja Kujala. I specialize in Portfolio management topics at Eficode.

[Marc] (28:55 - 29:04)

Hi, I'm Marc Dillon. I am a Principal Consultant at Eficode, specializing in platform engineering and Enterprise transformations.

[Darren] (29:04 - 29:10)

Hey, I'm Darren Richardson, Security Architect at Eficode, and I work to ensure the security of our managed services offerings.

[Marc] (29:11 - 29:18)

If you like what you hear, please like, rate, and subscribe on your favorite podcast platform. It means the world to us.