[00:00:02:02 - 00:00:23:01]
Hi there, I'm Ben Pearce and welcome to the Tech World Human Skills Podcast. Every episode we talk through how to thrive in the tech world, not just survive. Now, if you want me to work with your team, just give me a shout. I love to help teams be more influential, memorable and successful with their stakeholders. Head over to www.techworldhumanskills.com to book a chat.
[00:00:23:01 - 00:00:32:02]
Tech World Human Skills Podcast.
[00:00:32:02 - 00:01:11:19]
It is brilliant to have you with us. Now today we are talking about the darker side of tech and sustainability. We hear a lot about how tech might be a tool to help the environment and sustainability goals, but we hear less about the impact it has on our environment. So today we're going to talk about that. Now, our guest today is an expert in this topic. He's a technology director at Scott Logic and also host of the Architect Tomorrow podcast and community. So please welcome to the show, Oliver Crunk. Oliver, it is great to have you with us.
[00:01:13:07 - 00:01:28:16]
Thanks Ben, that was a great intro. Really looking forward to this conversation. It's going to be great. Well, thank you so much for coming on the podcast. I'm really pleased that you've joined us. Now, for all our listeners that perhaps haven't crossed paths with you before, Oliver, could you tell us a little bit about your background?
[00:01:29:19 - 00:01:38:11]
Yeah, sure. So I started off my career as a hands-on developer engineer, got a classic computer science background as many working in tech do.
[00:01:39:15 - 00:03:00:11]
And then progressed into architect roles, chief architect roles. So now I've got to operate in a technology consultancy. You mentioned Scott Logic, but I've not always been in consultancy. I've flipped around from end user world vendor and consultancy a few times. And yeah, I'm working a lot, as you say, on sustainable technology. It's what I'm really passionate about. I did a stint actually at an environmental consultancy many years ago. And so that's what grounded me in a real interest and passion for the environment. So yeah, currently working a number of open source research projects, which people, which I'm sure we'll talk about a bit later on. And just sort of trying to help clients practically make changes to improve the impact that their technology has on the planet. Brilliant. Well, it's brilliant to have you here. And quite serendipitous. I think I said the word right because we met, we bumped into each other at the tech show London speakers lounge. And I remember we were like, we hadn't crossed paths at all. I sat down to have my lunch. You sat down at the same table. I can't remember who was there first. I just went, oh, hello, what's your name? And we had a great chat. And after we'd had a great chat, we went, do you know what, let's get you on the podcast. This would be a great conversation. And I think our listeners would love to hear from you. So very serendipitous and brilliant that we've crossed paths.
[00:03:01:16 - 00:03:27:22]
Yeah, that was a great conference. I was doing a talk about the nature impacts of AI actually was on a panel as part of the data center world part of Tech Show London. But yeah, it was great bump into you, Ben. Yeah. So with with all that said context, context set, let's dive into this. So can you talk to me? Yeah, we talked in the intro about the darker side, you know, what is the current impact of tech on the environment?
[00:03:29:17 - 00:03:45:22]
Yeah, I think it's interesting because I think we've been subjected to an awful lot of marketing messages, positive marketing messages in technology about how amazing technology is. And don't get me wrong, you know, I'm a massive fan of technology, right? It's been my whole career. But I think the last few years has been really quite eye opening.
[00:03:47:15 - 00:04:54:05]
Yeah, a lot of people think about the energy consumption of our devices or servers or data centers, things like that. And that, yeah, that's probably 30 to 50% of the problem. But the other large chunk of the problem is actually what goes into making devices, making hardware, making infrastructure. You know, data centers have lots of concrete and steel involved in them. So it's not even just the servers, right? It's the whole the whole lifecycle. So in sustainability world, there's a term called lifecycle analysis. And, you know, we're just as technologists, most of us are just blissfully unaware, really, of what it takes to get the handset that's in our hand or the laptop that we're typing on, or the cloud service that we know that's hosting our applications. We're pretty clueless, if I'm honest, about what the impact of all of that is, right? And there is the real real genuine impacts to planet and people. Yeah, well, so I would put myself in that, you know, fairly blissfully unaware. I mean, I was I have been involved in a lot of it was perhaps more the energy in the carbon side of the data center side, particularly with moving from on premise to in the cloud type scenarios.
[00:04:55:15 - 00:05:31:01]
And I'm aware of, you know, minerals used for my phone or whatever that are, you know, that there's rare minerals that are needed. And perhaps, well, not perhaps, they're not sourced in the most ethical and sustainable and fair way. But that's where I'd probably stop. So maybe should we take a couple of those key things? Maybe we could put data centers and then maybe devices later, because they're two of the things. And then can I have one in there? Then can I throw in AI at the because they seem to be three things that we're talking about. Could you expand the data center, the impact of data centers on environment?
[00:05:32:09 - 00:06:01:14]
So data centers are a really interesting one, like, and you talked about cloud and how a lot of people are moving to cloud. And it's sort of a blessing and a curse in a way, I think moving to cloud in that you can get economies of scale by concentrating things into one place, you get those efficiencies often. The trouble is, it's kind of outside out of mind. You know, when back in the day, I'm sure you're similar, Ben, you know, we have to go and physically rack servers in racks and data centers and stuff. And a lot of that now is taken away, right? So you don't really have a physical appreciation of what is it that my software is running on.
[00:06:02:15 - 00:06:27:14]
But and the other challenge we have with cloud providers is there's a lot of secrecy around running a cloud provider, because there's quite a lot of commercial sensitivity there, right? If you're running a cloud in a certain way, it's a bit like F1 teams, you know, there's obviously lots of competitions and secret things going on to make their car faster than the other one. Yeah. In data center world, it's like, well, you know, we're doing this clever thing with cooling, or we're doing this clever thing with how we're using our servers or whatever. So there's not an awful lot of transparency.
[00:06:28:15 - 00:06:37:00]
And the transparency there is tends to be around carbon. But what people are starting to become aware of, and we'll talk about this a bit more when we talk about AI, is the water consumption.
[00:06:38:09 - 00:06:49:14]
One of the things people don't realize is because we've obsessed on carbon, and don't get me wrong, it's important to track carbon. But because we have looked at energy consumption and carbon so intensively, organizations have over-optimized on just pure energy.
[00:06:50:15 - 00:07:14:14]
And the side effect of that is often water. So there are huge amounts of water used for cooling, often in places with water shortages or water constraints like Spain and parts of the states and other parts of the world where there's real problems with water supply. And we're just evaporating that water into the sky, because that's the cheapest way of cooling a data center. And it uses the least electricity to just waste a load of water to cool your data center.
[00:07:15:15 - 00:08:27:14]
So that's one example. There are a whole host of other impacts, just like nature and biodiversity by sticking a massive great big bit of concrete in the ground. There's often sort of side effects from doing that, right? But the big one is water. We're just not measuring enough and we don't have an awful lot of transparency around. Yeah. And it's also interesting because, like you say, we're taking all of these resources and pulling them in one place. So let's say around London, for example, right? Because that's where lots of people are. And that is where we want low latency to get to all the businesses, to all the people, all of that kind of stuff. But around the London area, which is where I live, and you're on the M4 corridor as well, there's not a great set of reusable renewable energy. Like, like if we were building data centers up in Scotland, you know, we could be using hydro, we could be using wind farms, we could be using whatever. But around here, that's not, and yet that's where the data centers are. And the actual electricity they're using, regardless of what the marketing teams at tech vendors might want to tell you, they are using electricity from the local grid. And that local grid is coming from non-renewable sources, for example.
[00:08:28:15 - 00:08:39:15]
You've touched on something that I will dive into just briefly. There's some terminology here that's important. One is it's called location based versus...
[00:08:41:16 - 00:08:49:14]
Oh, come on. I'm not ready to go in my head. Sorry. A first bit of editing. If I'm not going to just leave this in and make it funny gag.
[00:08:50:15 - 00:10:50:09]
Sorry, I've remembered location based versus market based. Okay. And so you're you're dead right, Ben, right? Like a lot of marketing is based on what's known as market based, like what's the market's average COT or you've bought something on the market, you've bought renewable energy on the market. But the reality is your location is actually consuming maybe a gas fired power electricity. And your point about London is interesting. I mean, like we've made massive investments in the North Sea offshore winds. There's actually a huge amount of wind power coming into the London area. But it's it's getting that power to where it needs to go. So, yeah, I actually think data centers and cloud are a really interesting spot, right? Because they are centralized and consolidated. The problem is we're moving to a world where power is actually quite diffuse. And your point about actually lots of wind being available in Scotland will actually answer probably is to put edge compute closer to wind farms rather than have lots of big centralized data. So I think we are going to move into an interesting future world to meet the demands of AI, which obviously we'll talk about a bit more in a bit. We're going to need to get more creative about how we connect the power to the compute. And some of that is going to be probably using edge devices and using end end user devices a lot more for AI than perhaps we are at the moment. Yeah, so really fascinating. And I love the point that you made earlier. And it's a bit like cash and cash cards. And now, you know, tap tap a go on your on your on your phones. You're so disconnected from money that you just spend, spend, spend. Whereas if you had a stack, if you had a stack of pound coins, and you're like, right, I'm going to take these pound coins. And these are the pound coins I'm going to pay the mortgage with. You suddenly go, right? And now these are the coins that I'm going to use for my monthly spend at coffee shops, you know, suddenly you'd start going, Oh, actually, I understand. Whereas because it's obscured now and digitized, we're perhaps less familiar with it. Say same with compute and like, well, it's cheap, we'll throw it over there. And therefore, I don't need to think about the energy or the resource that's being used behind it.
[00:10:51:22 - 00:11:37:18]
Yeah, absolutely. And to move on to sort of hardware, I've got a question for you, Ben, I know this is a bit controversial, as you're supposed to be asking me the question. But, but how much how much materials, how much raw materials do you think are required to produce a two kilogram laptop? How many kilos of material? Right. So to produce a tube, so to produce my laptop, the average laptops about two kilos, right? Okay, your MacBook or your Dell, you got on your desk is probably about two kilograms, right? How many kilograms of material do you think you're required to make that? Well, I'm gonna I'm gonna use my Minecraft. So I'm imagining you've got ore, and then you got to turn that ore into metals or whatever. And then there's gonna be a process to do that. So I'm gonna say it's a lot more. I don't know that so I'm going to make up I am gonna say 50 kilograms.
[00:11:39:16 - 00:11:50:14]
Oh, okay. So would you be shocked to hear it somewhere between 500 and 700 kilograms? Wow. So almost a ton. Yeah, get towards a half a ton to a ton for just my laptop.
[00:11:51:18 - 00:12:06:14]
Just one laptop requires that much. And then water is crazy as well to like refine all the metals and you know, yeah, so like the the the stuff that's outside out of mind is is wild. And on the other side, the flip side of buying new devices is that our old devices have to go somewhere.
[00:12:07:15 - 00:12:30:09]
So e-waste is a really, really big problem. And unfortunately, the easiest way for you to recover some of the metals in a motherboard, there's a lot of gold actually in electronics, which is obviously quite valuable. What you end up having is people burning motherboards and circuits to release the gold from that board. But in the process, they're they're obviously breathing in toxic fumes.
[00:12:31:15 - 00:13:37:02]
They're creating further waste in the landfill site. So there's all kinds of wild things going on in the wider tech, you know, could a supply chain that if people knew about they would be absolutely horrified. So where so you know, the key thing to do with your hardware when you get when you get rid of it is absolutely do not throw it in the bin. Firstly, see if you can find someone else who can make use of it. Yeah. Can you sell it or can you give it to a charity or a school or something? And if you can't, then absolutely take it to the waste electronics disposal. So take it to your local council or wherever you live around the world, take it to the place that will actually process it properly and dispose of it properly. Because the worst thing is for those things to end up in the in the in the in the wrong place and have toxic materials potentially going into water courses and things like that, because you've not put your laptop or your phone in the right place. Absolutely do not put your phone in a black bin is the one take one first take away perhaps for your for your listeners. Right. Okay. Yeah. Yeah. Do it properly. And I guess is there lots of e-waste coming back from data centers as well because they must have stuff going on with life.
[00:13:38:05 - 00:14:37:20]
100% and it's a pretty big scale. And again, we don't really know how much because there's a lack of transparency. But it's going to be significant. And you think about the GPUs that are going on at the moment and we're kind of bleeding into a we've already bled into the conversation. We'll get into it properly in a second. Sure. The turnover of new types of GPU, these these processes that are used for gaming and for AI and for lots of mathematical calculation that every time we go for a new generation of these devices, hopefully there's someone else who needs a slightly lower grade one. But if there isn't a market for that, guess where it goes. It goes on the scrap. So, you know, it's it's pretty crazy like the yeah, we need to start rethinking how we do compute. It needs to be far more modular, far more repairable and upgradeable. I think we've we've gone for a phase of like wanting really shiny, slick clicking devices, which is great. And they're waterproof and they're amazing. The downside is if your if your phone's battery dies, Ben, do you replace that yourself?
[00:14:39:15 - 00:16:47:09]
So I normally do. I normally go. Yeah. So my model of phones is maybe a bit different to everybody else's. OK, fair enough. Then maybe it is. Maybe it's not. So what I normally do is wait for the next generation to come out and then I then look at what the generation before was and go and get a refurbed one off of eBay or Amazon or something like that. Well, that's very good. Well done you. And then I try and then use that for a couple of years because it just ultimately looks like a black screen with it with a case that I buy for five or six pounds on it. And it's got a camera and the cameras have been pretty good for a long time. Yeah. And then I try and normally rinse that. Now, the battery is normally the first thing to fail after maybe years. And I maybe do one battery upgrade that the replacement batteries never last as long as the original batteries. They may may maybe make it last another year or whatever. And then that's normally the point where right batteries fail twice. We're a couple of generations. So that's how I tend to do my phones, which is maybe a little bit. But they're really hard to upgrade, right? So it used to be I remember an old Samsung phone I had a few phones ago. You could just unclip the back of the case and pop in another battery. Yeah. Super easy. Yeah. Can't do that now. Because if you want a device that's waterproof and looks really cool and is super, super thin. Yeah. Like you have to start getting heat guns out. Yeah. Melting adhesives and it's messy. It's really quite a specialized process. So yeah, I kind of feel like this fast fashion and fashion for like really nice looking devices is also has a part to play in how disposable a lot of stuff's become. So the planned obsolescence is another thing with these devices. We have so many cycles of device like, you know, there's a new iPhone. It feels like every year, 18 months. So you almost feel like you're pressurized into getting a new device. So yeah, this whole fast fashion and planned obsolescence and lack of repairability devices is a bit of a problem when it comes to environmental impact. Yeah, it is. So we thought about data centers. And what I've liked thinking about there is the outside out of mind. Just don't need to consider it anymore. Then we thought about those devices and the fast fashionness of them and whether you can recycle them, how you should recycle that kind of stuff.
[00:16:48:10 - 00:17:23:14]
Can we have a little talk about AI? Because we've referred to it a couple of times. What's your take on AI? Let's start with the negative effects on the environment. And then maybe we can swap around to the positive. So we do put a bit of balance in there. Let's start off with the perhaps negative effects on the environment. Yeah, okay. I guess AI is a really, really broad term, right? And I think at the moment, a lot of people when they say AI mean generative AI and the challenge, I think with generative AI, if we just sort of focus on that for a moment is it's incredibly brute forced.
[00:17:24:15 - 00:17:42:14]
So, you know, these models are incredibly power hungry because the kind of current AI approach to a lot of complex problems is just throw huge amounts of data and computer problems and sort of crunch away and rely on very high end GPUs or fast processing to almost do magical output.
[00:17:43:15 - 00:17:56:22]
Of course, the flip side of that is it requires a huge amount of hardware. We've already talked about the impacts of hardware, right? It requires a huge amount of electricity. It requires a huge amount of water for cooling. So it's like traditional compute, but it's on steroids. To give you an idea,
[00:17:58:20 - 00:18:10:14]
we're seeing figures like, you know, the average Google search maybe uses 0.3 watt hours of electricity. And people are projecting that to use chat GPT for a similar sort of query is a 10 times, you know, 10 times lift.
[00:18:11:15 - 00:18:33:14]
So it's about three watt hours of electricity. The reality is I go back to the point we made earlier, we don't really know because there's a lack of transparency, but we can tell. There's a great article actually for your show notes for those that want to dig into this a bit deeper. MIT just did a really, really good, quite long researched, you know, article on the impacts of AI as much of the data as they could gather.
[00:18:34:15 - 00:18:43:16]
It's a brilliant series they've got on AI's sort of energy and environmental impact. So I'll ping that to you to let the listeners sort of dig into that in more detail.
[00:18:44:17 - 00:18:55:14]
Yeah. And it's fascinating because, you know, what you've started to talk about there and also because I'm such a consummate professional, I do my research and I was reading one of your blog articles on conscientious computing.
[00:18:56:18 - 00:19:33:14]
Now, but thinking about layers of abstraction and the impact that has. So I remember at uni, when I did uni, we started off, we coded in binary and then everybody went flipping heck. That was a nightmare. Then we went the next level up and we started with assembler and then everybody went flipping heck, loading stuff in and out of memory. No way. That's horrible. Then we then went to just, you know, normal like code. Then we went to object oriented and then that gets obstructed and obstructed to the levels where you are now to the point where you are asking an AI to generate code for you.
[00:19:34:15 - 00:19:56:15]
The energy efficiency of me typing one naught, one naught, one naught and putting specific bytes in registers versus me getting an AI to generate the code from me, the amount of energy and then just the finished product, how less efficient it is versus a binary thing that you've written that's very slick and very specific.
[00:19:57:16 - 00:21:39:06]
Wow. You know, the difference is huge. It's wild. It's wild. And like we've traded off developer efficiency for machine efficiency effectively. And I'm hopeful. And I think, as I say in some of that blog series, you just referenced that we can find ways to kind of actually allow those two things to happen together. And it kind of is possible, right? Actually, if you can use your AI to write C or Rust code rather than Python, then in theory, it should be, it should consume less energy. But it's, but I think the question we're not asking ourselves is, is that your use of AI to do that justified? Will it pay back? We're not asking these payback questions in a lot of cases. We're just doing stuff because it's exciting. It's cool. And like, yeah, let's just, let's just build stuff. Sometimes I think we just need to take a bit of a pause and go, do we really need this? Or actually, am I doing something that's more efficient or yeah. I think we're not even asking those questions. In our case, we're just getting giddy with excitement of what I can build a thing and a computer can now talk to me in language like me. And yeah, it's really exciting, right? I was a community event of presales leaders, so lots of leaders in presales coming together and we were talking about AI and how we can use it in our business. And one of the anecdotes, one of the guys there was just said, oh, he was like, whatever the novelty AI thing, you know, that pops up and everyone goes, oh, have you done this? Have you done this? And he said to his daughter, oh, go on, ask me this. And his daughter, I think was a teenager. Go on, do this. And I'll show you an AI. And so she said to him, why would you want to do that? Like you're burning three trees just to show me a reindeer on a toboggan or whatever the novelty thing that the AI had generated. And I just thought it was quite refreshing, actually.
[00:21:40:15 - 00:21:54:14]
And certainly from somebody from a future generation going, actually, should we? Yes, you can show me this AR novelty that will entertain me for a few minutes. But actually, should you? Because there are these unseen costs that we're not thinking about.
[00:21:55:15 - 00:22:55:19]
100%. I mean, my tech mentor, Garrosa Sol, he said he's no one with us, he used to have this great phrase, which he used to be quite a lot, which was, Oliver, just because you can doesn't mean you should. And I think it applies to AI. And don't get me wrong, I think AI can be incredible and amazing. I just think at the moment, there is so much sort of hype around it. It's like you should be using it for everything. And I think some people are falling for that hype. I think others are thinking a bit more considerably about where do I use it? You know, where does it actually really help me? And it saves a lot of work versus actually, it's something frivolous that, you know, do I really need that? And I think hopefully some of that frivolous novelty will wear off. And I think if I'm honest, Ben, I think some of it will be forced to because we're living in what I'm calling a bit of a free trial period right now, right? The true costs, even financial tree costs of these technologies aren't really being passed on to us as consumers or as businesses. And the burn rates are crazy, you know, we're talking billions a month for some of these AI companies, right?
[00:22:57:15 - 00:28:56:11]
Investors eventually will want to see a return on those figures. So the companies will either have to try and increase prices, they'll have to start monetizing your data, giving you adverts. We're already starting to hear chat GPT, a potential you can have like, you know, product sort of adverts based on what you're what you're asking, right? So we're living in a bit of a bubble at the moment, just because we aren't seeing the true costs. And this is why in a way, one of my more recent pieces, I'm a little bit more hopeful, because I think the financial pressures, the economic pressures will actually reduce the environmental impact. Because if something is really expensive, well, you'll use it less. And as a result, then the environmental impact, the demand will be less and therefore the environmental impact. Will also be less. So it's not the only thing we need to do to fix it. I'm not saying just put the prices up. But it's certainly something. And the other thing, and this was this is going back to the layers of destruction with, of course, deep seek coming on the scene, you know, everybody's saying everything needs to be done on these super duper Nvidia chips, China, perhaps not having the availability of these chips, and therefore, innovating and going, actually, we can't just throw all this stuff at it to solve it. We've got to think from a better architecture, a better approach, which actually ends up in something that's perhaps a more efficient way. 100% agree. I mean, like this, but I talked about this brute force approach at the moment. I think this is often I think the way things work, isn't it? It's like it becomes it's very expensive, very, you know, almost clunky, the first implementation of a technology. And I think that's what we're seeing. We're seeing sort of this raw potential of something, but you do have to throw a huge amount of energy and a huge amount of hardware at it. And I am hopeful that this global sort of AI arms race will actually result in more efficient models, more efficient hardware, because look at Bitcoin mining, right? Originally, people used to do that on GPUs on Nvidia graphics cards, right? But they came up with someone went, no, we can create dedicated silicon for mining overnight, the value of graphics cards plummeted. I don't know if you remember this when you know, when this happened. Sadly, it was COVID when they were still expensive, and I bought one during that period. So I know precisely how much they were whilst they were being used for mining as well. Yeah, so I think with that, that moment will come, we're really starting to see companies, innovative companies and new startups looking at how can we run AI differently? And I think, yeah, the Chinese stuff, you know, love or hate it, they are showing a different pathway. And I'm, yeah, I'm pretty hopeful that the architect to me is very hopeful that we can find more efficient architectures to power AI for sure. Now, we said we were going to look at a bit of a balance for the, you know, the good and the bad. So we've looked at perhaps the environmental impact of AI. On the flip side, how is AI going to help us with the environment? And how do you see that playing out? Yeah, yes, there's a couple of answers here. One is, I go back to the point about the AI term being really, really broad. There are non generative AI technologies out there that help look for drugs or materials and that sort of stuff. So already, scientists are using AI models, not things like chat GPT, but more specific AI models to look for new battery chemistries, for example, that will help us, you know, have electric cars with great ranges or, you know, design better wind turbines that generate more electricity or help balance the grid. So there's, there's a whole bunch of things around energy and materials and all that sort of stuff that AI will absolutely help us. There's other things like looking at deforestation and looking at Earth observation, so satellite imagery and looking at the planet and going, what's happening with the planet? Can we, you know, help sort of predict what's going to happen with rainforests or, or, you know, kind of different uses of land? And you can't do that manually. You know, there's just too much data, too many photographs for even massive teams of people to pour over. So then again, AI models are incredibly effective at helping, you know, kind of look at illegal deforestation or all that sort of good stuff. Right. So there's some amazing uses of AI. Now, they are relatively narrow uses. If we look at generative AI, I actually think for sustainability, what we can do with generative AI is actually process all the amazing expert material that's out there on sustainability so that on the internet, there will almost certainly be something that will tell you as a company or an individual how to be more sustainable, you know, how to, how to kind of make your business operations more efficient. The trouble is it's often in quite technical sustainability language or scientific language. I think the beautiful thing about generative AI is it can take this vast field of sustainability and it can help translate it into language that's, you know, language that you or I would understand or anyone would understand. So I've got a great example of this, actually, we got a bit of research funding at Scott Logic to look at generative AI and try and find greenwashing. So we actually built over three months, we built this tool that takes companies' ESG reports and yeah, analyzes them and looks for claims in them that are potentially false. And when we categorically call things out, it's designed to help a person identify the red flags or amber flags, go and look at these things because they look a bit suspicious. So yeah, and actually, we were blown away with how well it worked. It's not just generative AI power to use some other technology as well to ground it, stop it hallucinating as much. But the cool thing we also were able to do is detect green hushing. I don't know if you've heard about green. Tell me about that. Green hushing is when a company doesn't talk about the things that actually are great news. So they talk about things they've made progress on, but they keep quiet the stuff that they aren't, you know, they aren't doing such a good story on. So by loading up all the expert materials that say, you know, well, as this kind of company, they should be talking about this, this and this, we were able to then look at the ESG report from a company and go, does it have those things in it? And to what extent does it have those things in it? So we were, we were pretty blown away, actually, with how effective using generative AI and other technologies, I should stress, it's not just chat activity on its own, that sort of stuff, you need other things as well. But how well it called out, yeah, greenwashing, greenhushing, stuff like that. So I think for generative AI, for sure, the the power is going to be helping come up with sustainability plans, helping come up with, you know, effective action plans and roadmaps to reduce the impact and digest all this amazing scientific literature and put it into sort of, you know, more layman's terms so that people understand what they need to do.
[00:28:57:21 - 00:30:02:10]
Interesting, interesting. And, and it leads me on, I think that's a great segue into our next piece. Yeah. So, yeah, a couple of years. So I left Microsoft, let's say three years ago, yeah, something like that, not quite that, but but certainly when I left the, the focus on environment in carbon reduction was huge, we would talk about it at the company, our customers would talk about it, our partners were talking about it. It was almost like no brainer, everybody in the world is thinking about it. If you were going to put an investment buck in somewhere, if you put it in green technology, everybody's going to bite your arm off to do it. Now, three years later, it feels like that pendulum has swung a little bit. Certainly, when I look at the political situation, you know, and I hear world leaders saying, drill baby drill, you know, and that kind of stuff. And perhaps the impact that somebody like Donald Trump saying that, which then filters down to American tech corporation leaders, which filters down then to the tech, the tech community and the tech ecosystem.
[00:30:04:21 - 00:30:17:00]
What's it like? Are people still focused on environmental and sustainability goals with the same level of focus that they had before? Or are they, are they stopping now, given that sort of political world?
[00:30:18:07 - 00:30:21:20]
The world is fracturing a little bit on this topic. Okay.
[00:30:23:17 - 00:30:45:20]
So certainly think the US now. Yeah, so let's face it, some of the clients I was talking to about this topic have stopped talking about it. Okay, because they feel like it's too much of a political hot potato, for example, or it's just not as much of a priority as it was. However, in Europe, and in the UK, people are still very much on this topic. And I think the reason is, there's a couple of reasons, I think.
[00:30:47:00 - 00:30:47:04]
Firstly,
[00:30:48:13 - 00:33:55:18]
politics can change science doesn't, unfortunately, you know, we can, we can, we can do hope as much as possible that by saying something is no longer true. The reality is, you know, we're experiencing already changes in our in our in our client that you know, increasing pressures on our supply chains and there are even if you ignore the climate change piece, right, and you just look at card, hard, cold economic reality, it's getting more and more expensive to dig things out of the ground because there's, there's only a, there's only a finite amount out of them, right? So companies are already starting to think about how do we become more circular? How do we find uses for our waste products? But like we were talking about earlier, how do we recycle more and actually see those things as a value stream rather than a liability? Great example, this is heat. So I've had some amazing conversations recently with startup data center operators that are heating up elderly homes, care homes with the excess heat from a data center. So they're putting a smaller edge data center inside a residential building, be that an elderly home or apartment block. And they get free, they get a free location to put their compute for one thing. So that's a tick. And then they also get rid of their heat so that their waste heat actually becomes a sellable product. So they kind of trade heat for real estate, right? They get a space in this in this city, perhaps where they wouldn't otherwise put computer get compute in. Those buildings often really got fiber. So they've already got the infrastructure ready to go. And they and they provide heat to that building, which is amazing. So I think that's organizations that see this isn't just about regulatory compliance or reporting or tree hugging. This is actually cold, hard cash here. You can make your businesses more profitable, more efficient, leaner and be part of the sort of future. I think organizations that have their heads in the sand and think, oh, we can just continue to be extractive and just, you know, do the whole thing. You know, do the whole take make throw away type economy. I think in the next five to 10 years, we're going to see a lot of those businesses really struggle, because the environment just isn't going to be able to sustain them. You know, the amount of raw materials to put out on the ground just aren't going to be there. And also people's consumer brands live and die, I think, on their reputations. And increasingly, we're seeing brands like Patagonia and others who are showing the way like being very leading. And I think people will come to expect that from every brand. So I think you're right. There's certainly headwinds in one part of the world. I don't see that. I don't see that being a contagion elsewhere too much. But for sure, it's having an impact. So, and so I guess this then leads on to the next bit, which you've already started talking about. And we've already alluded to, which is this idea that actually if you build something sustainably and with efficiency in mind, not only does that have an effect on the environment in terms of your carbon, but actually, there's something there about unique business models, there's something about cost savings. Could you unpack for us what are the benefits for taking like an ESG style approach? What are the other selfish benefits aside from the planet and the environment and everybody else? What are the selfish benefits you get from adopting that same approach?
[00:33:57:10 - 00:34:53:04]
Yeah, I mean, let's just make it really, really simple. I mean, like the word sustainability is about an enduring organization and enduring business, right? If you want, if you want sort of long medium to long term value for your organization, and many investors do, right? Private equity firms do, pension funds and so on do, right? They take these longer term views. You need to make sure that your business is on a long term footing, not on a short term extractive footing. And whether that's how you treat people or how you treat the environment, that they're very similar. So yeah, there is the sort of viability of your business over a longer timeframe. And then that's stakeholder views of the brand image of your company, how investable are you as a business? Are you a short term, is it sort of organizational or do you take a longer term view? In which case your investors will look favorably. Employers, sorry, employees will start voting with their feet already are, in a lot of cases, particularly the younger generation we talked about earlier, who are more savvy and more conscious of planetary issues.
[00:34:54:05 - 00:37:19:09]
So there's this sort of customer and employer and investor brand reputation, which I think can't be ignored. You know, there are significant kind of pressures that come from that. So there's all that. And then there's your point about efficiency. And what's interesting is, if you ask people to save cost, they'll maybe find a little bit of cost saving. If you ask them to go and save the planet and save cost, they come back with far greater cost savings, because there's another reason to do it, right? It's been proven. There's been a couple of a couple of examples, more than a handful, sorry, I should say of examples of where, because actually in a corporate world, if you go and magically go and find cost savings, people go, why weren't you saving that money before? Right? It's an embarrassing, it's kind of an embarrassing conversation to have. If you're put on a cost saving exercise and you actually do really well, it can actually make you look quite bad. Like what Oliver, what Ben, why weren't you saving that money? I mean, you're saving this money for the business before. And so whereas the convenient thing about using sustainability is it's a bit of a smokescreen to go, well, you know, we looked at the problem a different way. We looked at it from a sustainability environment. So it's been some, you know, magically doing that, but I was defined 25% savings. The reality is those savings were probably always there if you just sort of thought about things a bit more efficiently, but it provides another reason to go and look. So, you know, some people just don't care as much about cash as perhaps we think they do, but you go and tell them to care about environment. And by the way, that will result in cost savings too. What do you know, you end up in a load more cost savings as well. Yeah. And like you said, it's thinking differently, isn't it? I mean, I remember a few years ago, a colleague of mine had built a little carbon calculator where a few, so if you were creating an app in, I think it was in, it would have been Azure at the time. So you're creating an app there. Basically you can figure out, right, well, how much energy is that using? Right? Like what sort of machine have you got it on? And then, and then he'd done the next thing, which was there's an app, I'm sure, you know, the ESO app, which is where you're going. Yeah, it's a carbon intensity. Yeah. So if, for example, you were to do able to do that in data center X versus data center Y, and you can see in the local grid of that area, what it's like, you could say, well, actually, if I run this in, let's say Scotland, for example, I know there's maybe not so many data centers in Scotland, but if I was going to run that in Scotland, then I would get this efficiency. If I was going to run that near London, I'm going to run that in the area. Yeah.
[00:37:33:05 - 00:38:02:04]
So, you know, I think that's one of the things that you could do that if you thought about it differently, massively reduced the app developers impact by designing the app well. Yeah, 100%. You've beautifully just described the sort of location shifting and demand sort of time shifting approaches. I mean, going back to AI, AI training isn't latency sensitive, like that could be running anywhere in the world. And so I'm really pleased to see some AI model companies running their workloads, training workloads in Scandinavia, or, you know, remote parts of the world,
[00:38:03:05 - 00:38:11:05]
that've got plenty of renewable energy, and also have air cooling, because the part of the world where they can just use free air to cool their data center. Yeah, that cold.
[00:38:12:15 - 00:39:44:04]
Yeah. And so, like, there are certain workloads that don't need to be run right next to London, right? It doesn't make sense. You don't, you know, yes, maybe inference, they're running a model, running AI, running applications, there's more latency sensitivity. So that doesn't need to perhaps run a bit closer to the user. So training sort of somewhere remote, and someone was playing a power or maybe overnight, you know, in batches makes a lot of sense, as does I think starting to run more and more of this AI on our devices. So I think it's really interesting to see what was announced recently around open AI and Johnny I've you know, the guy who designed the Apple devices, because I think the future of particular consumer AI is going to be a lot of it running on the local device. And so I do wonder if if that's a sign of open AI, maybe pivoting to training the models in big data center still, but actually getting them to a point where they can be more efficient and run on your bit of consumer hardware, because then, because if we're starting to put more and more of our lives into these things, like very sensitive situations, conversations, you know, our deepest thoughts, do we really want that going to a cloud data center, you know, with hacking scandals and various things that you know, we've seen, or actually, we're far more comfortable with that to sitting on a device that we physically got. on ourselves. So I know great thing there is potentially that's battery powered, it's a marginal additional cost to an existing device. As long as it doesn't create lots of throwing away existing devices to buy new AI powered ones, there's obviously a trade off there. I do think that's hopefully where the future is. It's sensible training of these models, and then it's running them on lower power hardware and less brute forcing this problem. Yeah.
[00:39:45:05 - 00:40:01:04]
Well, mate, time is escaping us. Yeah, where I'd like to finish just before we fully wrap up is what can people do? So we got lots of people listening to this people working in the tech sector leading teams in the tech sector. What can people do?
[00:40:02:05 - 00:40:35:02]
Yeah, so there's a great course by the Green Software Foundation. So if you're working in tech, and you've not been to learn dot Green Software Foundation, do you go check out it's a couple of hours, it talks to you about the things we've just talked about in more depth, like location shifting, demand shifting, all that sort of stuff, carbon intensity, gives you the basic terminology, and it gives you some practical things that you can go and do. So that'd be number one. The other thing is just talk about and ask questions like ask questions of your suppliers of your teams. We really need to raise awareness of this topic, like there's far little awareness. There's not enough awareness of this topic.
[00:40:37:00 - 00:42:55:04]
And then I guess the personal plug, we've been running an open source project for about over a year. It's called the tech carbon standard. It's designed to help people understand that broader lifecycle impact of your technology from a carbon emission standpoint. So there are great open source projects out there. Do you go and check those out and support them where you can? So yeah, I think it's awareness and then, you know, start to look at what you can do easily. And then yeah, there'll be harder decisions to make down the line. But I think until people are aware of this challenge, then, you know, we're in a really difficult spot. And then from a personal perspective and a devices perspective, there was something like e-waste. So make sure you're never putting your stuff in the bin, seeing if you can get other people to use it if you can pass it down the line. And also, I will say I am a huge advocate of refurbished phones. I've been doing it for probably 10 years now. So much cheaper because you buy it for a few hundred quid and it lasts last years. And it's a refurbished phone to begin with. And like I say, if the camera is good enough and it's the same size as it like what and you put a different case on it anyway, I could go and put a Thundercats case on or whatever. Nobody actually sees the phone and how thin it is. It's great to hear, Ben, that you're living and breathing that because that is a great tip. Like, do you really need to buy new and that applies to everything. I mean, I bought a laptop recently. It was refurbed rather than brand new. And because actually the prices of new hardware are pretty have gone up quite a bit. So, yeah, absolutely. Think about the hardware because the hardware has as much of an impact in many cases as the electricity consumption. For end-user devices, actually 80% of the impact is in the device itself rather than the electricity consumption, would you believe? And we do it with cars, right? We get second-hand cars. I'm into music. I get second-hand music stuff all the time because you can tell it's decent quality. It's going to last. So get into that culture of second-hand and refurb stuff with your devices. I'd love to see, Ben, I'd love to see a culture of like thriftiness and second-hand stuff being seen as cool. I think in some situations it is sort of vintage sort of stuff, retro sort of stuff. But yeah, I think that's what we need. We need to sort of value what we've got more. Yeah. Yeah. 100%. Brilliant. Well, let's wrap up. So for anybody walking their dog or out in the gym or wherever that's listening to this podcast and zoned out a little bit.
[00:42:56:05 - 00:43:00:07]
Key takeaways from your perspective, what would be key takeaways?
[00:43:02:01 - 00:43:16:12]
So yeah, I think be mindful of the impact just because it's outside out of mind doesn't mean it doesn't exist somewhere. Right. So when you're using a device using a cloud service that's running somewhere. So just be just be aware of that. What can you do about that?
[00:43:17:23 - 00:43:24:03]
We just talked about some resources, so Green Software Foundation, Tech Carbon Standard, things like that to help you navigate this world and understand it better.
[00:43:25:14 - 00:44:34:03]
And then yeah, if people want to talk to me, I'm always open to chat about this topic. Come and find me on LinkedIn. Come and check out some of the Architect Tomorrow stuff. We've talked about cloud compute. We've talked about a whole bunch of different topics on there. And we're thinking about how do we architect better tomorrow. So yeah, that would be my sort of summary. Yeah. Brilliant. And from my perspective, the thing that's really just resonated with me is this idea of being frivolous with your user. You know, and it's so easy to just sit at your desk or sit on your phone and just do stuff going, "Oh, it's free." And it's almost like I think it's Martin Lewis that talks to the money saving expert or whatever that talks about if you were to take your salary each month in pound coins and stack it in the corner of your room and then see how much goes on the mortgage, how much goes on the car, how much goes on coffees, you'd really get an appreciation of where you could save money and what you could do. I love the idea of that with, you know, using cloud services and using AI services and that kind of stuff. Somebody needs to think about how they're going to do that, how they implement that. But I just, just the mindset of doing that I think is brilliant.
[00:44:35:07 - 00:44:46:03]
Yeah. 100%. Brilliant. Wow. We're out of time. So I'm just going to say, Oliver, thank you so much for taking the time to join us. It has been an absolute pleasure to have you on the show.
[00:44:47:16 - 00:44:49:11]
Brilliant, Ben. Thank you. It's been a great conversation.
[00:44:51:17 - 00:45:11:20]
Don't go yet. Remember to subscribe to the podcast and rate the show. It really helps us grow and book new great guests. And remember, if the podcast isn't enough for you and you want weekly micro learning delivered straight to your inbox, sign up to the TechWorld Human Skills Weekly. Head over to www.techworldhumanskills.com to sign up.
