Transcript

Shane Hastie: Hey folks, it’s Shane Hastie here. Before we start today’s podcast, I wanted to tell you about QCon London 2024, our flagship international software development conference that takes place in the heart of London, April eight to 10. Learn about senior practitioner experiences and explore their points of view on emerging trends and best practices across topics like software architecture, generative AI, platform engineering, observability and secure software supply chains. Discover what your peers have learned, explore the techniques they are using and learn about the pitfalls to avoid. Learn more at qconlondon.com. We hope to see you there.

Introductions [01:10]

Good day, folks. This is Shane Hastie for the InfoQ Engineering Culture Podcast. Today I have the absolute privilege of sitting down in a global conversation with my friends and colleagues from the InfoQ Engineering Culture team and a very, very special guest. So we’ll start, I’m just going to go around and say welcome and ask each person here to introduce themselves. So Susan, please, let’s start with you.

Susan McIntosh: Hello from Denver, Colorado. I am Susan McIntosh. I am a scrum master and coach with a background in web and software development. And ages ago I was a teacher. So sometimes that pedanticism comes out.

Shane Hastie: Welcome, and a regular contributor to the culture and method staff.

Susan McIntosh: Yes, that’s right. And I forgot that.

Shane Hastie: And Jutta, you are our special guest today. So thank you so much for joining us. Would you tell us a little bit about yourself?

Jutta Eckstein: Yes. First of all, thank you for inviting me. So my name is Jutta Eckstein and I am based in Germany and I’m in the agile field maybe forever, I don’t know, started in ’97 with XP or ’96, I don’t know. And I wrote a couple of books actually. And those books are also telling my journey from looking at how to scale agile and how to distribute agile and what to do on the organizational view. So help bring agile to the organizational level. And my current passion is really about what is Agile’s responsibility in well, understanding that the planet is on fire and that we also have to take our share and look for more sustainable products. So sustainability is my topic.

Shane Hastie: Welcome. Thank you. Craig.

Craig Smith: Hi Shane. I’m Craig coming to you from Brisbane, Australia on the land of the Quandamooka people and I am an agile coach and a delivery lead and been around the agile community almost as long as Shane.

Shane Hastie: And Ben Linders.

Ben Linders: Thank you, Shane. Ben Linders from the Netherlands. These days mostly doing workshops on Agile, occasionally some advice for organizations, also doing a lot of writing, picking up on my second book on retrospectives, making some more time available to dive into aspects like psychological safety and retrospective facilitation. I think those are still very important topics if you want to get real benefits out of your retrospectives. So that’s what I’m mostly doing these days. And writing for InfoQ, of course.

Shane Hastie: Writing for InfoQ like a machine, Ben. I think you’re still one of our most prolific authors. And last but certainly not least, Raf, welcome.

Raf Gemmail: Hey Shane. So I’m Raf Gemmail and I live out in the wilderness of New Zealand, the far north, near little town called Whangarei. I have been on a journey with computers since a little lad and I was an early developer booking.com and around that sort of time I started looking at XP books. I still have a little XP book, which I carry with me almost like a good luck charm, but it was a world where once I went from there into banking, we were doing waterfall and the transformation was still happening. And so I rode that transformation through banking. I’ve worked at scale for most of my career. So I was at BBC iPlayer where I came across Katherine Kirk who’s spoken at some of our events and has written for InfoQ at some stage. And she changed my life pretty much making me realize that an engineer has other skills beyond just building software.

And she introduced me to Kanban and she continues to be an inspiration. But that took me on a slightly different path. As I went through my career, I got more involved in tech coaching and some agile coaching, DevOps coaching. I’m very much on the side of let’s ship, let’s build, let’s learn, let’s educate, let’s build the right culture. And I still bounce from code through to delivery and making sure we’re building the right thing and working with stakeholders. So I’m a manager right now with a MarTech company working at massive scale and getting some really, really awesome engineers who want to be in the code. Also think about the customer as well, and it’s a fun ride and I write for InforQ since 2012.

Shane Hastie: And I’m Shane Hastie, I herd this group of cats, officially they call me the lead editor for culture and methods here at InfoQ and heavily focused on humanistic workplaces and the people elements of getting work done well, particularly in socio-technical systems. And for InfoQ’s case, obviously software teams are my areas of passion. So because I’m leading the conversation, let me go into my passion. What’s happening with software teams in the last year and what do you see happening going forward?

What’s happening with software teams [06:19]

Susan McIntosh: One of the things that I noticed, I don’t know, maybe it was because I was paying attention to the staff engineer track at QCon last fall. One of the things I noticed though is that we’re relying on the staff engineers to basically be a coordinator for teams to communicate with other teams to make sure that all of those little dependencies are handled and managed and tracked. And it doesn’t sound like that’s something that staff engineers really want to do. So I was very curious to see why they were left with that task. And maybe, I guess that’s more just a question and an observation than anything.

Raf Gemmail: Observation in my space is I’m looking very much at how to grow engineers and I think it’s been something for a while. I love Patrick Hill’s work around what it means to progress in the field and I have engineers who are growing on that path and I’m presenting with them with the option of a technical route forward towards becoming a principal or a staff engineer or someone who has a broader impact. And I don’t like the idea of individuals becoming proxies, but I do wonder what, there’s a journey usually as you stepping into those roles, you go through the journey of I will do all the things because I know engineering really well and I’ve had to personally step away from that quite a bit to, hey, I’m an enabler and as an enabler I’m connecting people. I’m not the proxy. I am able to make teams more performant, make them take a macro view and improve the processes they work with.

And so yes, I don’t know about that specific trend, but I do know that there is a move towards how people are becoming proficient in those roles, can be successful using things like the application of observability to how we work. And that’s things like metrics around DevEx, metrics around friction. I personally measure waste in my teams, I’ve done that for years. Let’s see where the friction is, then we can do kaizen and make things better. If I have a spreadsheet, I can make a case now that we need to invest in local sandboxes or something because it has a cost. But being able to use those metrics really enables teams to put a mirror in front of themselves and say, actually these are the things we need to improve for a staff engineer or a principal to say actually strategically these are areas we need to improve.

Now I think that’s slightly divergent from what you’re saying, but the good principle, the good staff engineer in my humble opinion will then use that data to enable others and create an environment where people are more successful and can focus on their own value lines, in my opinion.

Jutta Eckstein: So if you’re asking about what happens to the teams, and maybe people think this is now boring what I’m saying, but I think that the biggest change that we have seen over the last years and also coming, is the working from home and what it does to collaboration. A colleague friend of mine Dasha, she’s researcher as well, she called said what we are seeing the flexibility and individualization paradox that on the one hand I guess we all love to be so flexible and work from home and we don’t have to commute and it just provides so many opportunities.

But on the other hand it just means we are all by ourselves. That means the ad hoc conversations are not happening anymore. And from what I heard, and I’m not a researcher so take it as such as, but what I heard and which is also my impression that as a result, on the one hand, the team cohesion is not so strong anymore, which also means the loyalty to the companies people are working for.

And what I even think is from my perspective more important is that there is less innovation happening because all the side conversations are missed and often it’s the side conversations that trigger a thought like oh yeah, well we have solved it like this here and maybe you can try that as well. Or, oh I see there’s this problem, maybe we should finally start doing something about that. And the way it is right now, we don’t know this anymore.

Raf Gemmail: I think that kind of points at the need to be more intentional in how we do remote and to make up for that. And my personal experience is all I can refer to here because I’ve seen the counter examples as well where it’s not as successful, but if you are intentional in things like making sure people communicate and that you have formal and social and non-social catchups, you have intentional bringing together of people across the business. So for instance, last quarter I took the whole concept of my big room planning, and I’m not a big quarterly planning fan, but sometimes you work within the persona of your organization and in this situation, brought together people from product, brought in engineers as well. So they were getting the insight and there were fruitful conversations.

There’d been some visits from senior staff, people getting together, intentional remote coffees, remote socials. You can be really crazy creative with remote socials and there are many things you can do to try and fill that gap. The side conversations around innovation, the coffee cooler ones may not happen as easily, but if you’re having random coffees, like I do informal conversations with people who are colleagues in other departments, people I don’t usually catch up with. And if you try and put that intentionality in, some of those ideas will trickle up, you can try and make up for that lost innovation in my opinion.

Ben Linders: I think a lot of organizations still have big challenges with teams and one of the reason might also be that they’re being too focused on the team as a kind of sole solution for the problem. Where the key aspect is that you want to have collaboration between people and they can be in the same team, but sometimes you don’t have all the people in that team and you have collaboration with people in other teams or other parts of the organization. So I think the focus should be much more on how can you establish the means and a culture where this kind of collaboration can happen and can flourish instead of trying to build strong team. Because actually some of the strong teams I’ve seen were so much inward focused that they had trouble fitting into the organization because they had difficulties communicating with the other parts of the organization because they were so strong and so good as a team altogether and they failed due to that.

So I think the focus should be much more on enabling collaboration in different ways, different formats, and yes, people being more remote makes this more challenging and I agree with you making explicit on what are you trying to communicate and how you’re trying to communicate stuff and how you want to work together is one of the key aspects to make this work, be more explicit in how you are communicating and how you’re working together to make it work because you need that because you don’t see each other on a daily basis, you don’t stumble into each other. So you have to be more explicit of what you want to do and what you expect for other people in there to make it work.

Susan McIntosh: It’s a really good point that now with the working from home option, which is becoming more standard, I’m looking at all of our backgrounds, we’re all working from home today. It just adds another dynamic that we have to pay attention to when we’re working in teams. Not only what is the team trying to deliver, but how are they working and what’s the environment that they’re working in, in order to make sure that they’re all as effective as they can be.

Raf Gemmail: I have a dedicated workspace and I think this is where I wish organizations would put their money, where their mouth is, there’s a return sometimes as you get people to work remotely, you may not need the physical space. I don’t know if that’s always reallocated, but I was looking at this survey around, like a state of remote work survey from last year and they said in there that 30% of people had dedicated workspaces, which means like 60% of people I think were working like couch-surfing with their computers and finding dining tables. And that is the thing that I think is the challenge, which worries me a lot.

You make allowances for people because there’s noise around them, there’s stuff happening and I don’t know what the specific solution to that is other than investing in giving people access to remote working spaces, the tools they need and make sure that these people are effective as you would in an office space. Also HR, physical responsibility around ergonomics and office equipment. Yeah, I’ve seen examples again in both directions, but there are definitely examples out there of people who are not necessarily working in the most effective environments.

Jutta Eckstein: I know at least in Europe that there are some companies who investigated in the workplace from their employees at home and so some of them, they just gave them a budget to set the stuff up the way they need it, which I think is a really great way. But it has also to do with trust of course, which is maybe a completely different topic.

However, I wanted to go back to what Ben was saying and there were two things for me. The one thing about, and I can’t agree more to that, the teams that are really stable are often then the ones who do the things like we are doing things here, there is nothing new and they’re not open anymore because they’re so, I don’t know, such a close group. And so I agree that we probably also need to look more at collaboration, maybe even across teams and so on.

And building on this, which is another thing that I kind of heard between the lines of what Ben was saying, I think this is really a big trend because what we are seeing more and more is that it’s not like the supplier, meaning any company building a product, defining how the user is using the product or what the customer journey is of the product. It is more and more the user defining how to match various products together and build their own customer journey.

And it might be completely different what the one person does compared to the other, which also means in order to be more flexible with that, we cannot stick to the idea of building this one individual product, but we always have to be open to understand that the product is actually just a piece of a puzzle for our clients.

Shane Hastie: I’d like to dig into a topic that has been on our trends report for the last few years and has been highlighted in QCon events and others. Developer experience, what is DevEx today and where is it going?

The importance and value of Developer Experience [17:37]

Ben Linders: Well, what it is, I think you can get thousands of definitions from what it is and one trend that I’m seeing actually is that I see more and more organization that say whatever it is, it is important and it does matter. So they’re spending more time on this, both for their own productivity, also for retention of their employees. They also see that if you take care of developer experience, people are going to stay within your company instead of leaving. So it is getting more important. I think what it is also varies from developer to developer, so there’s no single definition. And one other thing that I’m noticing is that if you really want to make this work, you have to do a kind of shift right in here instead of providing solutions, providing to providing an environment and getting people to accept it. To us, getting your developers involved early on in what is the kind of tool that you want to build in your organization up to maybe even rotating developers into platform teams and taking care of the tools that they’re using or giving them space to develop their own tools.

So organizations are very much focused on, we made some stuff and now we have to find a way to get our developers to accept it. I think they are on the wrong track. It’s not about acceptance, it’s about making stuff where everybody says, yeah, I’ve been waiting for this for weeks already. I’m happy it’s here now, of course I want to use it.

Raf Gemmail: I think this ties to friction removal as well. When you shift right, what you’re starting to do is observe what is the impact. It’s that whole build, measure, learn loop, right? And so the earlier point about measuring and metrics, ties in very much with that DevEx. I mean there are methodologies out there, but all the DORA metrics, all the things that give us insight to how we’re doing, but it’s responding to that and then putting the loop in. So I talked about capturing friction. The things I get are things which slowed my team down or multiple teams down where they’re like common things like an environment that will never work. And so once the person’s in standup and one team going, oh, I lost about two hours yesterday. I hate that thing. The other guy in the other standup, the other person in the other standup, and it accrues.

It’s a huge amount of time that the org is losing. Many developers are frustrated, their days have been made worse. And so by addressing that and measuring it, you can start improving the pain points which are causing developer friction. And that may be tooling, it may be education, it may be environments, it may be processes. But to your point about shifting right, I really agree with that. Observe and then make the improvements. And for me it’s about removing waste, which as a manager I can make a direct case for because it has an impact on our ability to deliver features for the customer because back to the old sort of thing about tech debt is also another point of waste. We can start putting numbers on it if we can see where we’re slowed down.

Shane Hastie: Teams and teamwork. There’s been huge amounts of research that talks to the great value of diverse teams and yet we still don’t see the level of diversity in our industry that is in society. Why not?

Diversity is still a challenge and there are concrete things we can do to create more inclusive environments [20:46]

Jutta Eckstein: Well, once I said kind of like Conway’s Law, an organization produces a design that reflects their organization’s biases, which is on the other hand something that’s also known that we try to have friends around us who are kind of like us, which is I believe a natural human thing to do. And in an organization that leads to less diversity of course, and often at least I would think, hope, not on intention, but it’s just also how we are. And so the key thing around that for me is to continuously paying attention to it. And this is tough, it is, and you’re never done. That’s how it feels to me.

And I’m speaking more from maybe a bit different position than some other people who might listen to that, but I’m trying to bring more diversity on stage. So at conferences and at times I thought, okay, if I focus on that for a few years and then this will be more automatically because then I have all this network and people will come back and they will invite difference and so on. But it’s not happening, every time I’m starting from scratch. And I believe it’s the same thing in organizations. It really needs continuous attention.

Susan McIntosh: Yes, continuous attention and not just in the hiring funnel or chain, but once you hire somebody and there’s so much evidence that having one person of one ethnicity or diversity channel, that one person isn’t going to feel a part of the group. If there’s just the one of them, you need more. My husband works at a university and I’ve seen this at universities dealing with first generation students. So those students who have never experienced what it’s like to be at a university, as an example, they don’t know that office hours are when they are allowed to go to the professor’s office. One of them thought that was when you should not go. So how do you make those expectations and the things that we all assume are correct? How do you make sure that those assumptions are more explicit for the opportunity for hiring more diverse staff to make sure that they’re welcome and included?

Ben Linders: I think included is exactly the word that I was thinking about. It’s not that much about diversity, much more about inclusion, making sure that when you aiming for diversity that the people feel included in the team, that they feel safe enough to join in, that they feel safe enough to speak from their own values, to speak from their own ideas within the team. Certainly when those are different from the majority of the team because that’s when it’s going to be challenging there. So you want to have an inclusive environment for diversity to work, otherwise it’s just going to fail.

Jutta Eckstein: Which pays back to the developer experience by the way, I think, because then people will feel more safe and therefore also try something and learn and be open. Yeah.

Ben Linders: Yes, and feel more included.

Raf Gemmail: One thing to watch out for from personal experience with this as well is as we are trying to learn and enable people, I talked about measuring waste. One observation I’ve had is across an org you can sometimes predict the people who are going to be more proactive and capture these things or make the loudest noise and there’s always going to be a bias in your data. So you need to counter that with make sure you get inclusion from the non-obvious. And so when you’re trying to optimize even for DevEx to make things better for bringing new perspectives, I think that again, back to the point earlier, you need to be intentional in how you bring in that diverse set of views.

Ben Linders: This is actually where working from home can be an advantage, by the way, if you have somebody who’s more introvert, if they feel safe at home, if they take time and space there to think about how to do stuff and to write it down and contribute in that way into a team, they can feel much more included into the team instead of being in a meeting in the organization as an introvert where other people are making a lot of noise and bringing forward their ideas. So actually working from home can create a more inclusive environment in that way.

Jutta Eckstein: I would think it’s creating more an inclusive environment in terms of, well, if you have health issues, if you have a familiar situation that makes it just hard to be away from home for long hours and so on. For the other stuff you mentioned, I’m not so sure if it just relates to facilitation techniques because also if you’re in a room you can ensure that people do individual work or they work in pairs and then we have the whole plenary and so on and this way inviting various voices.

Overall, I think we all need to understand that more diversity in a team is not only will perhaps attractive, trendy, anything like that, but it also makes the product better because we all also see better what different kinds of users might be out there and therefore we create the products for a bigger audience. And I think this is something we all need to understand and therefore it also pays off. And of course there are the various studies which I don’t have handy, but where they say more diversity also means more productivity, maybe yes, maybe no. Definitely it opens up perspectives for the product.

Raf Gemmail: I’ve shared this in a previous one of these years ago, but I was working previously with a startup, which was sadly a casualty of recent economic climate. But in that startup we were an educational institution and in that educational institution we tried to bring in sort of people from diverse backgrounds, people who were not in technology because they were from an underprivileged background, people who’d not come in. And we counter gender bias, we counter ethnicity bias by being very, very analytical about how we brought people in. We were looking for people who had certain ways of thinking, who displayed an attitude to be able to be effective as software engineers. And there’s some machine learning stuff behind it as well to identify that. But these people have gone off and been successful. They’re people who completely changed their industries from admin or retail work or bar work to working in software.

There’s a person out there who I recently saw became a scrum master and she’d been a BA for a little bit and she was an effective communicator. She was great at bringing people together and she’s in the industry now. And so I guess where I’m going with this is that the recruitment face, if we are being meritocratic, we quit asking how we challenge those biases, we can get a more diverse set of people in. I’m kind of trying to do that at the moment, or at least a colleague of mine is in terms of looking at recruitment and how we can be more objective about it, create feedback loops, calibrate on how we assess someone who’s coming in so that it is based on their ability. And there are a lot of people with great ability, good learning mindsets, who you can bring into an org and who will be successful.

And then there’s the other side of it which you talked to, which is like creating the culture where once they’re in, they can continue to grow and learn and develop. And that again goes back to the principles. It goes back to others who can support them on that journey, make sure that they’re continuing to learn. Managers are taking them on that path, creating educational opportunities, pairing opportunities so that once they’re in, the pot of gold is at the end and it’s up to them to, they’re supporting trying to make the weather. I have my own neurodiversity issues. Another part of inclusion here is inclusion of people with different neuro diversities, which can really bring new insights into a team and challenge the way we build software or we build solutions for the customer.

Jutta Eckstein: What you pointed out, reminded me of something else as well. So it’s not only what’s happening in the team, but also how we look for example at outsourcing. And there are these two different kinds of, at least there are maybe more, but I know about the two ones, which is on the one hand impact outsourcing, which is about training, hiring marginalized people. So this is kind of what I heard what you said Raf, as well, but more like how bringing them into the team but you can also think of a constellation with outsourcing here. And the other one is ethical outsourcing where you just (“just” in quotes) ensure that your own values are also adhered to in the outsource company. So your own social or maybe also environmental ethical standards are true for the outsource company.

Shane Hastie: Craig, you’re being quiet. I saw you try to say something.

Adopting modern leadership approaches is going slowly [29:56]

Craig Smith: So I think the thread from listening to all of those great topics for me is there is an element to how we modernize, I’m going to say facilitation and leadership because it’s a little bit of both and we can’t put the blame squarely somewhere else, but when we talk about things like working from home and the whole remote thing is that we need better ways to be able to allow people to bring their best to the workplace. And the same goes for the inclusion topic. And the same goes for the DevEx topic is that unfortunately still in many organizations, we’re still running them like an industrial complex. The amount of time in this calendar year as we record this, we’re a couple of months in, I’ve been astounded by the amount of managers that have said to me that I have to bring everyone back to the office so I know what they’re doing.

And it’s because we haven’t educated our leaders and facilitators on how to manage that. But I think also we have a challenge as a community about how we also better support that. And I think it also comes down to things like tooling as well. Unfortunately, tooling like Microsoft Teams and Zoom as good as they are, are not ways to have remote collaborative teams working effectively. But also what was also mentioned there was things like investing in technology. And Shane, you and I know with the organization we work with, the amount of people who physically can’t turn on a camera because it doesn’t work or has been disabled at their end or their bandwidth typically throttled by their organization, that is so hard. So we’ve got a lot of work to do to enable this, but I think just saying, let’s go back to the old paradigm is an extremely backward way of thinking.

And as we start to think about all the challenges we have as a world, how do we actually help leaders, but also how do we help just those people who are facilitating, those people that are self-starting and self-organizing also, I think there’s a real challenge there for us, and I don’t think we have all the answers in the toolkit that we have right now.

The opportunities and challenges with LLMs [31:59]

Shane Hastie: There’s a few big topics that I want to make sure we get to. One leads very much from this conversation about remote teams, in person teams and also weaving into this, the application of large language models and AI. I’ve heard and seen and made the statement myself that leveraging the tools like copilot, like Chat GPT, like Perplexity, the large language models that are out there now and there’s so many more of them. Last I heard there were 63,000 new AI companies in the last 12 months, which is phenomenal and 80 plus percent of them I’m sure will disappear at some point. But I’ve said, and I’ve seen it said that using these tools will make good programmers, great and bad programmers, redundant, the same for scrum masters, the same for business analysts. The thing that I see and worries me is how do we bring new people in with the base skills if we’re giving them these tools or am I being naive?

Susan McIntosh: There’s a danger, right? AI and large language models and generative AI, they can create so much content, but how good is it really? And Ben, you’re laughing, but it’s true, right?

Ben Linders: Yes, it’s true.

Susan McIntosh: It’s so easy to see the articles that have been written solely off a generative AI program. And I admit I use it, I try it because it’s so much easier to start with a draft than it is to start with a blank page when you’re writing, but it’s not the best of editors. And so it’s a good way to start instead of starting from scratch, but then we need to learn a new skill of editing and fine-tuning the information that we get out of AI and generative AI. It’s a challenge.

Jutta Eckstein: Definitely it is. So my first thought was especially the way you were stating in a chain, isn’t it the same? And of course it’s different as what we have seen for many years, which is a fool with a tool stays to be a fool. And that’s really, I would think still true here. And yes, I see that risk of well, you can get up to speed, but then to really use it and leverage it, you have to probably also cross the chasm. And the question is if you’re seeing that you have to do that or if you just then say, I lean back because I see it so easy and I’m having all that progress and therefore I’m not even seeing the chasm I have to cross. And again, so I think it is the same as we have seen with other tools and yet again, it’s very different because it’s so much more powerful and it’s just everywhere.

Ben Linders: And being much more powerful makes it even more dangerously. I would say that with AI, I think 90 or 95% of us should consider ourselves to be the fool and not the expert enough to work with these tools. And I use them myself in some situations and I was just comparing on what Chat GPT was telling me as a solution for a certain problem and what I could find out with Google and I got more or less the same stuff I did with one important difference, with Google I could see the source, I could see where the information was coming from. And for one topic, actually 95% of the information was coming from commercial parties who were trying to sell something and that wasn’t Chat GPT telling me. Google, I could see that.

And that made a huge difference on, okay, how reliable is this information? Is this something that’s proven or is this your sales talk? And it was sales talk, but Chat GPT didn’t present it like that. So I think we have to be even more careful with using AI tools right now because stuff may look very nice and may sound reasonable, but you have to be really, really well-educated and really well first into the topic to judge whether the information that you’re getting there is valid or not valid and what is the stuff you can use and cannot use.

Raf Gemmail: I think that’s going to evolve though, Ben, a bit. There’s a thing called drag, this retrieval augmented model thing where you can have an L and M and you can get extra data in and you can go off and query stuff. I’ve started playing with Bing’s chat copilot thing, which is a GPT-4 and that thing will show you some citations as well. Now the Bing garbage ones as well. I’ve gone in and it’s not exactly what I think it is. I think I had one which redirected me to the police when I was asking something about cyber bullying and it didn’t actually show me the source. The source and the link were different, but this is an evolution. We’re in the early stages of a new product and so I think it is going to improve. A friend of mine is in a startup with some of the people from Hugging Face and he’s in the University of Auckland down the road. New Zealand is a great place.

He was working on stuff to do a specialization of LLMs for your local domain and your local context using open source technology. And their thing is, let’s go to your data, pull stuff out and then you can specialize the model to give you better answers. Now because we’re in the early stages, I think we’re going to get a lot of garbage out of it and the buyer beware or user beware of like we need to self-regulate, review what’s coming out of it, also review what we put in. There are lots of people using Open AI’s model, Bing’s model, and these are things which come with clauses saying, hey, we can use your data for learning. So look at the fine print regardless of which one you are using. Also open source models, I have a machine here, people are paying fortunes for these massive models.

I have an AI rig here at home which has consumer grade GPU in it, and I can do chats, type things. I can get some information out of it, image gen.  So my hope is that people will invest a lot more in some of the open technologies, in building models which are able to look at their local domains a lot better. And because we’re in the early stage, it’s like a few big players have given us some fancy toys and we’re like, wow. But I do think the sophistication will improve. It kind of ties in with years ago, I remember, I think it was, who’s the hackers and painters guy? Paul Graham, I think it was him. There was a piece I remember seeing long ago when IDEs were kicking off and I was like an Emax user at the time and it was like, hey, IDEs make developers stupid or something like that.

There was some sort of catchphrase and as you went through it, it was like, hey, you’re going to rely on this IDE. And we’ve come to a point now where the features of our IDEs, like templates and things, they don’t take away from our need to think. They give us tools to do autocompletion and we use that to build the right solution to understand the customer’s need and figure out how our architecture plays with the piece of software we’re building, at least from a software perspective. I do think it looks like magic at the moment, but I think it’ll become another tool in our tool belt, at least for a while.

Sustainability and climate impact of technology [39:11]

Jutta Eckstein: Well, with my passion on sustainability, I cannot talk about AI without bringing up that topic as well. I’m sorry, I just have to, and we still too often look at, oh, what are all the great options or also what are the risks? But really one thing we all have to know that it really comes with a higher cost to the planet. And I know those estimates, they are also varying, but one says at the moment, training a large language model is about the same demand as my country in Germany. And again, I have not done this research so it might be wrong, but definitely the energy needed is extremely high and just thinking this is all there for free. It’s not.

And the other thing which also is for me connected to sustainability is probably also not news, but of course the models affect with the data that is there and the data is biased. And so speaking of social sustainability and inclusion and all of that, also what we talked about earlier, diversity and you name it, same problem. And maybe both of the problems are even getting worse with seeing that it’s used even more and I guess we really need to solve that.

Shane Hastie: Let’s dig into sustainability, climate impact. Jutta, this was the key reason for inviting you on is the work that you’re doing there. Do you want to tell us more about the work that you’ve been doing and what are the implications for our industry?

Jutta Eckstein: So I do various kinds of stuff, but the one thing that I am doing, so I’ve created an initiative with the agile alliance, so the Agile Sustainability Initiative, and it started in September, 2022 and we just tried to increase the awareness on the topic and therefore also hopefully increasing the responsibility of people that we actually can do something about that. So it’s not the topic of other people, but we also have our part here. The way I look at it is if we are looking at agile and sustainability in particular, I see on the one hand that what we can do is using agile to increase sustainability, which is then called sustainability by agile, meaning we use all our practices, values or maybe, I don’t know, mindset principles, you name it and help others to get more sustainable. And there are a lot of people doing that in various NGOs, mostly.

It can be something like heck, your future is one of the things that I kind of like, and this is an NGO working with refugees, helping them to on the one hand practice their developer skills and also show their developer skills so it’s easier for them to get a job and I know some people who’ve helped them as a scrum master for example, to do exactly that. So this is using whatever skills you’re having and they’re really various ones. Another example is also what I did together with Steve Holier. We worked for NGO in the climate sector and we just use what we know, which is open space, events storming or story mapping or whatever we felt was needed for them to make their next step. So it’s using all that stuff that we know and tell. That’s the one part. The other part, and I sometimes refer to this as our homework, the other part is sustainability in agile, which is looking what we are actually doing when we develop, well at least my clients mostly develop software.

Carbon footprint as a quality attribute of software [43:30]

So when we develop software and can we do this better? So one basic thing, well, which is basic but maybe not so much so, is to monitor the carbon footprint as part of a quality attribute in the Definition of Done so that it’s really embedded in the work we are doing. Maybe one last notice for me it’s important also to understand that sustainability really is a thing that looks at various different things. So the one thing is the environment and we hear stuff like green computing, green software and things like that. So we look at the carbon footprint of the system we are using. There’s also a lot we can do. Another part is a thing we have talked about already. So it is the social, all the people aspect. So it has to do, if I translate it directly into software, it has to do with diversity, inclusion, safety, security, then me being German, privacy is really to our heart. I guess it’s in our DNA.

And the third one is economic or prosperity, where we think more of that the whole product is responsible or that we are talking about sustainable economies. So the reason why I think it’s so important to look from these three perspectives is that it’s so easy to focus on one at the cost of the others. We can do it really in a great way here, speaking of Germany, at the cost of that, well this is actually what we are seeing right now. So very often I see these waste monitors and then Germany is really great. Well why? Because we ship our waste to other countries and then they have the problem and we are fine. It’s unfortunately, but this is what’s happening. I think I leave it there, could go on of course.

Shane Hastie: And we’ll make sure that there’s links in the show notes to that work that you’ve been doing on the Agile Alliance. Anyone want to pick up on that? Yes and to this.

Raf Gemmail: I love this, just throwing it out there, the idea of the carbon footprint in your DOD just made me feel like that’s awesome.

Jutta Eckstein: You know what the interesting thing is and well I discover so much on my way of working on that is that a lot of the things, if you speak about the environmental aspect goes actually back to really good principles and coding habits that well, people who are maybe my age have learned that at one point, maybe we forgot it and the reason was because we didn’t have the memory, we didn’t have the bandwidth and all of that. And if we rediscover all these patterns, principles, that would really help also the software to be more environmental friendly. So it could be just perhaps one thing is if we think about jobs that can run asynchronously, to give a real example, what we can do now is we look at when is the time where the energy mix in our area or where the job is running is renewable and then have that job run asynchronously at that time versus in the past we did also something like that.

We just looked at, well actually it’s similar because we looked at are the servers really under power like crazy and therefore we cannot run that job so therefore we postponed it that it’s run during the night. Well now maybe the night is not so good, well depending, well yeah, but if you are more looking for solar power, then maybe the night is not good. But it depends what kind of energy is used. But yeah, so again, I think that that’s really interesting for me to understand that the good old principles, patterns, they should be revived because they help us.

Raf Gemmail: I wonder how you measure things like the carbon footprint there as well because I love this because it ties in also with, we talked about cost savings earlier and having observability of your fleet and does an EC2 instance need to be up at all, right? At a certain point in time. Can we spin things down, use spot fleets, can we use containers or serverless or be more clever with that? And to your point, just going back to engineering, like we’re optimizing our software to be efficient in a way that has an impact, but how do you compute the carbon footprint?

Jutta Eckstein: Very good question. I don’t have a good answer. And the reason is that the tools are changing all the time, but there are tools available and I would point people to the Green Software Foundation who has really developed and makes that available. A lot of open source software which you can use to actually find out about that. Maybe one other thing, because you were talking about the cloud, what you now can do as well, for example, if you use Google as your cloud, the region picker allows you not only to look at latency or cost or anything like that, but also at the carbon footprint. And so you can also make this one of your criteria for where are you going in the cloud. And again, what I’m trying to do with the work that I’m doing is just to increase the awareness, what’s possible. And on the other hand, I also have to admit it’s really hard to keep up because there is so much going on, which is excellent, which is great. So it’s really a completely new field where stuff is happening.

The passion for making a difference [49:21]

Craig Smith: And I think this is a bit of a golden age. I liked how you talked about we have to relearn some of the original things. Like many of us on this call, we talked about, we’ve been around this since the start of agility, for example, a lot of the issues that we’re having in the community as people have forgotten why we did things the way we did them, and it’s been overloaded by lots of frameworks and processes and practices and things like that. I know sometimes you have to remind people and go, well, the reason we do this is actually you are new to this and you miss the 16 steps of things that we went through to get here. I feel the sustainability thing is now the same thing is that many organizations are talking about the fact that they want to be carbon-neutral, green, zero, all those type of things.

But we haven’t been very good at bringing that back to just the everyday person, the engineer who’s sitting there cutting code in a bank or something like that. And what my hope is through your work, Jutta, is that for people listening here is it’s a bit like agility, that didn’t start by a manager or a paradigm in an organization saying we need to be better. It started by the practitioners themselves going, we need to do better. And this to me feels like that rebirth of innovation that I felt when agility started and that’s why I’m so excited about the work that you’re doing. It’s going to take a while, but we had to start somewhere. And as you say, it’s awareness, but that awareness means that it’s not going to happen because your manager’s going to come down and say that, people find it hard to link their work to their organization’s targets. It’s up to us to start to have those conversations both as individuals but also as teams.

Ben Linders: One thing I’ve seen is that there’s a lot of people out there who are by nature environmentally aware, this is something that’s in their DNA, this is something that’s been important for them for many years already and they’re now trying to combine this with the technical work that they’re doing. So these are people, software engineers, maybe people working as architects and they’re looking for ways to do it in their work. And if you see somebody in your organization who’s having this awareness and is looking for ways to do it, give them space and try to find a way to leverage this and to help them to look for ways to do your software development better because they are driven from the inside, actually.

Jutta Eckstein: They have the passion, right?

Ben Linders: Exactly. They have to have the passion. They have some ideas, they don’t have the solutions, but they’re open to solutions that are out there. They’re looking for ways to apply that in their daily work and do that with their teams. So if you have somebody like that, please give them space, support them, make clear that you do find this important in your organization and support them in any way you can to be there and to do this work.

Jutta Eckstein: I like how you both, Craig and Ben, mentioned that it’s probably needs to be more bottom up and kind of more like agile came into place and I think that way as well. But I see in companies that, at least in Europe, it’s often also starting top-down, however, then it is stalling very soon because what’s done top-down leaves out our software development. They think about, well, let’s change our fleet and we have e-cars or no cars or bikes or anything like that, but then they’re done. Versus if you embed it in your daily work, then on the one hand you have more awareness throughout the organization. So everyone sees it as their own topic and it’s not a topic from somebody. And you can also make a difference on the long run and not only kind of the low hanging fruit. So yeah, bottom-up, I agree.

Awareness and greenwashing [53:16]

Raf Gemmail: This ties in the parallel with something I heard earlier in this week, which is slightly tangential, but there’s a lesson in it possibly for sustainability, which is that there was a talk by Laura Bell around DevSecOps and she was talking about how they were with DevOps and many of these other initiatives, it was a really good idea to have an outcome which was very humanistic, which involved collaboration. And then a term comes around, it gets taken into the org and it falls into the anti-pattern, which is the one it was trying to counter. You end up with the DevSecOps team, you end up with the DevOps team, you do not end up with the collaboration in the team. Agile, same thing. We have frameworks like Scrum that came in, we had values and principles. We are about ceremonies and the next sprint and the meaning is lost from it.

And in this space of environmentalism, we see a lot of greenwashing across products. People are going, “Hey, we’ll do the minimal thing. Look, we’ve got a picture of us on bicycles.” And I don’t know what the solution is there, but often well-meaning initiatives which start from the ground up can sometimes be taken out of context intentionally or unintentionally and then fall into some antipattern which doesn’t address the thing that you were trying to address in the first place. So I do wonder how we can try and avoid something like that for green technology because our world, we don’t have long, we’ve already gone over our emission targets. So it’s like it is imperative. And I agree with you, I don’t think we worry about this as much as we should or give it as much credence as we should.

Jutta Eckstein: I have a maybe very specific opinion about greenwashing, which is that, well, maybe I’m just an optimist. I think perhaps it is good because it says that the companies are really thinking this is an important topic, therefore they have to fake it. And my hope, and again I’m an optimist, is that this is their first baby step before they do the next one. So the fake it till you make it kind of approach. And yes, probably I’m wrong, but I’m just hopeful.

Ben Linders: Some make it, some will stay fake.

Susan McIntosh: Yes.

Jutta Eckstein: Yes, that’s true.

Susan McIntosh: But if organizations emphasize their desire to want to go green and they talk about it with the organization, with everybody in the company, then it’s possible that the software engineer might say, oh, well maybe I should pay attention to how much energy I’m taking to write this 16 line piece of code.

Ben Linders: That’s again given space.

Susan McIntosh: Yes.

Raf Gemmail: This is why I drew the analogy to security. Just because security is something that we know has a high risk. When it goes wrong, when our planet is past the point of no return, the stakes are very, very high. Yet security is something which is often skipped. Here are the ACs, what about security criteria? The non functionals, they’re not always the top and foremost. Sometimes it’s because of lack of knowledge. But the same thing with sustainability. The org may value it, but we need to then make a case that this story will take X times longer or this feature will take X times longer because we’re being environmentally aware.

Now, I don’t necessarily think that’s the case. I think we can get into a place where it’s the way we work. We can start using low energy servers, building for arm architectures. There are all sorts of things we can start doing to get there. We need to value it and we need to value it more than we value shipping the feature, which is where maybe I have a little cynicism about the way we work in general, but I think if we keep championing it, we can make a difference there.

Jutta Eckstein: Absolutely. And I think security is a great example because if I think about security 10 years ago, not much has happened at the time, people were talking about it but not really paying attention. But then over the years, quite some stuff has happened and now it seems at least in my world that people are considering it and it’s part of development and it’s not like a thing that’s postponed to Neverland. And so this is also what makes me hopeful that this might happen with sustainability as well. So that the more we bring it to our awareness, and of course also the more we learn about it, what we can actually do, the more likely it is we will actually do something. And talking about the doing, I feel like I want to share at least one more concrete example that people can do, which is based on the fact that very often it is actually software that is creating the e-waste, the electronic waste.

So if you think about your phone, whenever you substitute it with a new one, probably it’s because the OS is not updated anymore or the apps are not running smoothly anymore. Well, there might be some exceptions that I don’t know, you have smashed it too often or anything, but most often it’s really the software, the reason why people buy a new phone. And the same is actually true if we think about the products we are creating in our companies for our clients, that very often we expect the clients to have the latest hardware. And if we are just starting there to come up with a responsive design, having, I call them sustainability toggles, which are kind of feature toggles where we discover what client is using the product and then therefore offering these features and deactivating others because they are not supported on that client, that would already save a great bunch.

So that’s also probably an example of going back to the old principles, patterns and so on. But it just felt like, I want to give at least one more concrete example so people can have a takeaway here. This is something I can do. Yes, you can.

Shane Hastie: Concrete advice for engineers working on real world products. Thank you so much. Folks, our time is almost up.  As always, it’s a wonderful opportunity to get together around the screen with this team. We don’t often get together in person, but virtually we get a lot done together. We’ll make sure that all of the links, the references from this conversation are in the show notes, and of course, all of our contact information, our info queue profiles and Jutta, your contact information will be there as well. I’d like to end with one point from each. What is your hope, wish, or desire for software engineering in 2024?

Hopes and wishes for software engineering in 2024 [60:03]

Jutta Eckstein: I think we should not only always look at the right solution to whatever problem we have at hand, but instead look for better questions and let those guide us.

Susan McIntosh: I like that, Jutta. I keep thinking of the multitude of things that we’re trying to juggle these days, work from home and artificial intelligence and sustainability and all of these different mindsets that we have to think about. Perhaps with all of these different ideas floating around in our heads, we’ll come up with a very simple solution that will solve everything, that would be a dream. But maybe with so many very smart, talented engineers and other people in our industry thinking about it, we can come up with something.

Craig Smith: I think we’ve spoken a lot about the tools and the things, and my hope is that we can help people find better ways of working to support these tools, whether it be sustainability or AI or other things. We have to keep moving forward, and I think sometimes we,  we’ve seen this for a long time, people still resist. We’re still trying to apply sometimes, if not 20-year-old solutions, 120 year old solutions to the problems. So we have to move both in at the same pace, and I don’t think that some of our ways of working are keeping up, and I hope that’s a challenge for some of the people listening, that we can have those conversations and talk about them here on InfoQ as they start to emerge.

Ben Linders: My wish is that we take time to stop occasionally, to go slower, to reflect, to listen to each other and use that in the end to go better and to go fast, but it is something that we all need to do naturally. Things are going faster and faster and everybody’s going with the flow, with the current, so unless somebody occasionally interrupts in there and whether that’s a reflection or retrospective or somebody from lean says like, okay, let’s stop the line because something go wrong here. Whatever term or mechanism that we want to call it, it doesn’t really matter, but take the time sometimes to just stop a moment think and then continue.

Raf Gemmail: For 2024, I’d like us to continue building on the feedback loops we have everywhere. I think we’ve talked about frameworks and things, which can become dogmatic, but I think we’ve got to a stage where we can be really evidence driven. We can build learnings, take those learnings and fit them into how we work, our processes, our tools, our ability to deliver customer value. We were already doing that. We’re measuring so much more than we were, and this has gone early majority and beyond, the whole build, measure, learn thing, experimenting with customers, communicating with them hopefully is something we’re doing a lot more, but let’s apply that everywhere, right? We’ve got the DevOps life cycle. Same thing, right? We’re learning, we’re optimizing. I think there’s a lot more data we can collect and we can hold a mirror to ourselves and look at it, and that’s something that’s become part of my journey a lot more is look at the data, try and understand what’s there, talk to the people, understand the context of the data, but at the same time, we have a lot more ability to make decisions that are informed.

I think we should use that. On another note, there’s a lot going on in the world at the moment. I was a child who grew up during the Cold War and my fears with ones I thought my kids would never have, but now there’s a lot of stuff going on and you’ve got younger developers in the industry, people working across the industry who may be dealing with new things, new threats, new notions that I’ve, maybe I’ve reverted back to an old model, but I think there are younger generation out there who are like, whoa, what’s all of this stuff going on? Because the environment wars and that instability on the planet, there’s all sorts of stuff going on at the moment, which are new threats and new fears for people, and let’s be compassionate about that.

Shane Hastie: For me, it’s all about the people. We build software that changes people’s lives. Let’s make sure we do it in a way that is kind to the people and kind to the planet.

Thank you all so much. It’s been fun.

Mentioned: