[00:00.000 --> 00:11.240] Okay, shall I start folks? I'm afraid I have quite a few slides to go through folks, so [00:11.240 --> 00:17.440] if you, I guess I'll start now then. Hello folks, welcome to my talk, Getting to a Fossil [00:17.440 --> 00:23.680] Free Internet by 2030, before I start. Can you folks hear me all right? Excellent. Okay, [00:23.680 --> 00:28.520] all right. As you can see there's the title, a talk of the tech and policy changes to get [00:28.520 --> 00:34.040] us there. Hello, if you can't see my face here, because you're not in this room, that's [00:34.040 --> 00:39.040] what I look like, I suppose. My name is Chris, I work at an organisation called the Green [00:39.040 --> 00:45.000] Web Foundation. I've spent the last 10 to 15 years working in a series of wacky climate [00:45.000 --> 00:50.160] and data themed start-ups, as you can see here. Loco 2 is about locomotion, like trains, but [00:50.160 --> 00:56.280] also a low CO2, and going loco, going holiday, stuff like that, and it's just puns are a [00:56.280 --> 01:01.520] thing that I have been working with. But these days I do most of my time working with the [01:01.520 --> 01:06.320] Green Web Foundation, so we track the transition of the internet away from fossil fuels. I [01:06.320 --> 01:11.120] also run a podcast where I talk all about green software with other people, and I also [01:11.120 --> 01:15.560] work with a larger organisation called the Green Software Foundation, where we do work [01:15.560 --> 01:19.800] on policy specifically. During the time I have with you today, we're going to cover these [01:19.800 --> 01:25.560] three things here. I'm going to tell you about why I care about a fossil free internet by [01:25.560 --> 01:30.360] 2030. I'm going to share a framework for helping you think about sustainability in the digital [01:30.360 --> 01:36.880] realm, and then I'm going to use this framework to look ahead at some policy and tech changes. [01:36.880 --> 01:44.640] So let's start. Why a fossil free internet by 2030? So the main one is, let's be clear [01:44.640 --> 01:49.920] about this. We're in a climate crisis because we keep burning fossil fuels, and every single [01:49.920 --> 01:54.280] time over the last 30 years we've had a chance to get off them, we haven't. If you want to [01:54.280 --> 01:59.600] actually speak to your friends or your boss about why a fossil free internet is important, [01:59.600 --> 02:05.280] these are the ways I think you might explain it to someone. So first of all, it's achievable. [02:05.280 --> 02:09.640] There are big firms already doing this, but we'll see that small firms can do this, like [02:09.640 --> 02:14.680] climate emergency, carbon savings, that's another one. It saves lives because all the [02:14.680 --> 02:19.600] kind of stuff that gets put into the sky that it doesn't heat the world up ends up poisoning [02:19.600 --> 02:24.560] lots and lots of people. It saves money because, as we know, fossil fuels are not cheap if [02:24.560 --> 02:28.800] you've tried heating your house recently. And yeah, people like being in green firms, [02:28.800 --> 02:32.840] and yeah, there's a whole geopolitics thing that I think people in Ukraine will be very [02:32.840 --> 02:38.840] aware of, and as Germans, we were also, or as Europeans, we're aware of too. So that's [02:38.840 --> 02:44.160] why I'd stay there. Now a framework to think about this is what we use where I work. We [02:44.160 --> 02:49.720] have a model that we call CID, consumption, intensity and direction, where consumption [02:49.720 --> 02:55.200] is about, can I change how much we need? Intensity is, can I change how much harm is [02:55.200 --> 02:59.480] done by the things that we have to do? And direction is about changing, where do we want [02:59.480 --> 03:06.080] to go? What kind of future do we want to build? So the first of these is the lens is consumption. [03:06.080 --> 03:10.880] Because our society is largely still based on fossil fuels, using pretty much resources [03:10.880 --> 03:16.320] in any way, or create some form of carbon pollution and waste. And to start off with, [03:16.320 --> 03:20.240] if there are things that we want to do, it's a good idea to try to minimise the amount [03:20.240 --> 03:25.920] of resources we need to do that, no matter what. So with that in mind, we know that we [03:25.920 --> 03:30.320] should be reducing this, reducing emissions. What do the trends look like for the last [03:30.320 --> 03:40.200] year, right? This is a chart of historical CO2 emissions created by Dr. Robert Rode at [03:40.200 --> 03:45.560] Berkeley Earth, a kind of environmental data science non-profit. You can see the direction [03:45.560 --> 03:51.880] we're heading in. And the good news is that in 2015, almost every single country on Earth [03:51.880 --> 03:57.600] agreed at COP21, a climate conference, that we really should aim to reduce emissions and [03:57.600 --> 04:02.600] we should go for well below two degrees of warming for this century. Since then, we've [04:02.600 --> 04:06.920] actually found out that we need to be aiming for not just well below two degrees, but more [04:06.920 --> 04:13.040] like well below 1.5 degrees, which is a much more aggressive target. And that's what you [04:13.040 --> 04:17.880] can see. These are the potential pathways we should be on. And what does that mean for [04:17.880 --> 04:22.760] our sector, the tech sector? Well, the Paris Agreement will require for the information [04:22.760 --> 04:29.560] and communication technology industry to reduce greenhouse gases by 45% by 2030. So that's [04:29.560 --> 04:36.600] about 7% year on year if we started in 2020. And I'm not sure it went down the last three [04:36.600 --> 04:40.920] years, all right? So this was actually something that was put together by the Science Based [04:40.920 --> 04:46.360] Targets Institute, the ITU, which is the International Telecoms Union, I think, and [04:46.360 --> 04:50.280] a number of other kind of official organizations who spent a decent amount of time looking [04:50.280 --> 04:55.640] at this. And this is just for two degrees. So we need to be doing more than 45% really. [04:55.640 --> 05:01.520] So how are we doing so far? Well, there's good news and there's bad news, I suppose. [05:01.520 --> 05:05.600] So the good news is that like two billion people who have been connected to the internet [05:05.600 --> 05:09.520] in the last, say, five or six years, and we've been able to use the internet a lot more, [05:09.520 --> 05:14.120] which during the pandemic was very, very useful. I really appreciated having access [05:14.120 --> 05:20.160] to the internet during the pandemic. And what I want to draw your attention to here [05:20.160 --> 05:25.160] is that while we've increased usage of digital services, we haven't necessarily increased [05:25.160 --> 05:32.280] the usage of energy at the same kind of proportion, except for cryptocurrencies, which to be honest, [05:32.280 --> 05:35.680] the less said about that, the better. So I'm not going to really be talking about that [05:35.680 --> 05:40.960] too much if that's okay with you, folks. However, and I should ask, why have the figures, [05:40.960 --> 05:45.600] why have the figures not been higher? Well, the simplest answer is that Moore's law has [05:45.600 --> 05:50.360] been helping us. And Moore's law is basically where computer chips have been getting steadily [05:50.360 --> 05:55.460] faster and more power efficient over time. And as the sector has grown so fast, the average [05:55.460 --> 06:02.040] efficiency has increased. So the thing is Moore's law is not infinite. So going forward [06:02.040 --> 06:06.800] improvements in efficiency will likely have to come from new places, not just the chips [06:06.800 --> 06:12.000] themselves. And this is what this diagram from a paper in Science magazine illustrates. [06:12.000 --> 06:17.120] This paper is called There's Plenty of Room at the Top. What will drive computer performance [06:17.120 --> 06:22.240] after Moore's law? It's a fun read. And the general argument is that for efficiency gains [06:22.240 --> 06:26.920] to keep up with, but will keep pace with growth and really need to be going faster than growth. [06:26.920 --> 06:32.360] So we don't just stay flat, but actually go down. We'll need to start applying all these [06:32.360 --> 06:36.840] ideas at the top, all right? Which basically means it's going to land on us as developers [06:36.840 --> 06:41.440] and designers of digital services a lot more than it otherwise would have. We can't just [06:41.440 --> 06:48.320] rely on Moore's law doing the work for us. So that's consumption. And as Saul Griffiths, [06:48.320 --> 06:54.720] the author of Electrify says, you can't efficiency your way to zero. If we don't want to be cold [06:54.720 --> 06:59.320] and wet and without Wi-Fi, at some point there will be a minimal amount of resource usage [06:59.320 --> 07:04.880] that takes place. Faced with that, we need to think about the harm that occurs as a byproduct [07:04.880 --> 07:11.080] for every unit of usage that we have to make. So in our framework, we talk about this in [07:11.080 --> 07:17.120] terms of intensity because there are multiple dimensions of intensity from water usage to [07:17.120 --> 07:22.680] resource depletion to soil toxicity, air quality. But for the purpose of this talk, I'm going [07:22.680 --> 07:29.080] to focus primarily on carbon pollution. Now, in Germany, where I live, the carbon intensity [07:29.080 --> 07:35.440] of electricity is influenced by two main things. There's a lot of wind and solar for sure, [07:35.440 --> 07:40.880] but the majority of our power comes from burning coal. And these machines, that's what's powering [07:40.880 --> 07:45.520] our data centers. These machines, you can see, they destroy forests, they destroy people's [07:45.520 --> 07:51.800] homes just to dig up coal, which we then burn to heat up water to create steam. The steam [07:51.800 --> 07:56.600] turns a turbine, which generates electricity, which is what we end up using. Along the way, [07:56.600 --> 08:01.280] we lose about two-thirds of all the energy through cooling towers, and we emit a lot [08:01.280 --> 08:08.360] of carbon. So a kilowatt-hour of power is around, like, a kilogram of CO2. This isn't [08:08.360 --> 08:14.200] the only way. An alternative approach would be to directly harness the energy around us, [08:14.200 --> 08:19.680] using either solar or wind turbines, geothermals, and so on. Now, we're not burning fuels in [08:19.680 --> 08:23.400] this case. So we're no longer wasting lots of energy in the form of waste heat that we [08:23.400 --> 08:28.120] need to put into the sky. And it means that there's no need for the kinds of cooling towers [08:28.120 --> 08:32.960] that you frequently see, so it looks somewhat different as well. Also, the carbon emissions [08:32.960 --> 08:37.680] are much, much, much lower. Even when you factor in the actual making of the solar panels [08:37.680 --> 08:43.000] and disposing of them, for example, you can see the figure here, but it's something like [08:43.000 --> 08:48.480] 17 times lower. So this is why it makes sense to not be running on fossil fuels. Now, globally [08:48.480 --> 08:53.640] speaking, about 60% of the electricity we use right now comes from burning fossil fuels. [08:53.640 --> 08:57.960] But this is the global average in certain parts of the world. Energy is much cleaner. [08:57.960 --> 09:02.880] In other parts, it's much dirtier. Now, the numbers look tiny on this part. What you can [09:02.880 --> 09:08.320] see is that wind and solar, in the top right, are growing, and they've really been plummeting [09:08.320 --> 09:14.120] in cost over the last 10 years. Solar energy is nearly 10 times cheaper than it was 10 [09:14.120 --> 09:19.400] years ago. And something like 90% of all the new energy capacity that comes on stream these [09:19.400 --> 09:23.680] days is coming from renewables. So there is some hope. It's not just going to be that [09:23.680 --> 09:30.600] awful black line forever. So because the carbon intensity of compute is not uniform, like [09:30.600 --> 09:35.800] I said, all around the world, if you want to reduce the harm done by carbon pollution [09:35.800 --> 09:40.600] from the energy usage that powers your servers or powers any digital services, you have two [09:40.600 --> 09:46.400] main options. You can move work geographically to run things where the energy is cleaner. [09:46.400 --> 09:52.200] So this is like geographic migration. That's like the Iceland, the Nordics, and so on. Or [09:52.200 --> 09:56.600] because the energy, the carbon intensity of energy changes based on how much where the [09:56.600 --> 10:01.880] generation is coming from, you can basically move things through time. So you can wait [10:01.880 --> 10:06.560] to run things at a different time of day, for example. So you can wait till it's windy [10:06.560 --> 10:12.680] or sunny, or send it to where it's windy or sunny. So that's consumption, intensity, and [10:12.680 --> 10:16.680] direction. Finally, we have direction. If you care about open source technology and [10:16.680 --> 10:20.800] free software, you know, they're very important ideas that are kind of re-embed into free [10:20.800 --> 10:25.480] software. And this is like you choose into work on one project over another, you're making [10:25.480 --> 10:29.600] one version of the future more likely to actually happen than another. And this is the quote [10:29.600 --> 10:34.240] that I find really useful. It's from a guy called Cade DM. He says, technology is a social, [10:34.240 --> 10:38.440] political, and environmental accelerant. I think it's useful because it really kind [10:38.440 --> 10:43.680] of gets across the idea that we're making a deliberate decision about what we're supporting. [10:43.680 --> 10:50.120] And to give you some idea, this is a graph showing the direct emissions from a number [10:50.120 --> 10:56.240] of companies you've heard of in 2019, plus a single contract from Microsoft with Exxon [10:56.240 --> 11:02.320] Mobile, one of the oil companies that recently, like in the last month, announced $150 billion [11:02.320 --> 11:08.320] in profit last year. All right? Basically, the consequence of using AI to help Exxon [11:08.320 --> 11:13.440] Mobile basically drill for more oil and gas is basically around 6 million tonnes. So that's [11:13.440 --> 11:18.960] about a Facebook, all right? And this is the same figure. And this is actually something [11:18.960 --> 11:23.600] why I talk about actually thinking about what you use it for, not just can I make my stuff [11:23.600 --> 11:29.760] efficient basically. It's not all bad again. So Microsoft, they do this, but they've recently [11:29.760 --> 11:34.520] been doing more work about actually finding ways to actually get coal-fired power plants [11:34.520 --> 11:38.800] off the grid and replacing them with other forms of power, like geothermal or nuclear [11:38.800 --> 11:43.520] sometimes, all right? And every single time, that's like more than a Facebook, more than [11:43.520 --> 11:47.280] a Facebook, more than a Google's worth every single year coming down. So this is why it's [11:47.280 --> 11:51.000] important to actually be thinking about what you use it for. So let's use this framework [11:51.000 --> 11:59.280] to talk about some of the policy changes ahead of us. So remember that chart that I showed [11:59.280 --> 12:03.960] you before about emissions going up when they need to be going down, all right? Okay. For [12:03.960 --> 12:08.640] better or worse, this idea of reducing emissions to within safe limits has started to be referred [12:08.640 --> 12:13.120] to as net zero. And we're finally starting to see some consensus about what it actually [12:13.120 --> 12:18.600] means. Now, this is the ISO, the people who set standards. They've created some guidance [12:18.600 --> 12:24.280] about what net zero should really mean. And the actual guidance is pretty strident. They [12:24.280 --> 12:29.120] basically say net zero claims are no longer considered credible if they're not halving [12:29.120 --> 12:33.680] emissions by 2030 for an organization. So if your organization doesn't have a net zero [12:33.680 --> 12:37.560] target and it's not doing that, not really credible. They also say it needs to include [12:37.560 --> 12:42.440] the entire supply chain, not just like your own stuff, basically. And also you need to [12:42.440 --> 12:49.120] have interim targets. So targets for 2025 or 2026. And this is because seven years is [12:49.120 --> 12:54.400] roughly the typical kind of tenure of a CEO that you can say, we'll be net zero by 2030, [12:54.400 --> 12:58.120] leave and then leave it for someone else's job to actually fix. And this is why it's [12:58.120 --> 13:01.360] actually quite important to actually do this, because you need to have something which actually [13:01.360 --> 13:07.760] brings action early. Also in Europe, the, it's a bit of a mouthful, the European Corporate [13:07.760 --> 13:13.760] Sustainability Reporting Directive. In 2024, that's what organizations above 250 employees [13:13.760 --> 13:18.640] now need to be reporting, which means you need to have started reporting two months [13:18.640 --> 13:23.520] ago for 2023, if you haven't already started reporting. So this is a thing that's important. [13:23.520 --> 13:28.920] And then finally, the IFRS, they're a bit like the kind of Pope of accounting standards. [13:28.920 --> 13:32.960] They basically set a decree that everyone ends up following. And they've said that no, [13:32.960 --> 13:36.760] if you're going to talk about the carbon footprint of your organization, you need to talk about [13:36.760 --> 13:41.120] not just your organization, you need to talk about your supply chain as well. And that's [13:41.120 --> 13:46.440] actually really important because around 90% plus is makes up most people's emissions in [13:46.440 --> 13:51.240] the supply chain, not really your own emissions. So let's use our framework we spoke about, [13:51.240 --> 13:56.280] consumption and all that. Okay, can I change how much we need? Joseph did a really good [13:56.280 --> 14:00.680] job talking about things like the Blau angle and things like that. There are now emerging [14:00.680 --> 14:06.400] standards to actually try to label sustainability of digital services. And as he mentioned, [14:06.400 --> 14:11.600] yeah, there is Ocula was the first ever product that got some decent, got actually certified [14:11.600 --> 14:16.360] for this. This is important because people who spend money on digital software, like [14:16.360 --> 14:21.480] say, public sector or large, large companies, they can they can write these into contract [14:21.480 --> 14:26.280] and say you need to have a blue angle certificate in order to actually for us to spend tens [14:26.280 --> 14:30.720] of millions of euros with you, for example. If you want to like work with this stuff yourself [14:30.720 --> 14:34.880] or work out the environmental footprint of this, we mentioned before, the guy called [14:34.880 --> 14:39.480] Arna Tarara, he actually demonstrated a really, really cool piece of technology that does [14:39.480 --> 14:43.320] this stuff. If you have a digital service, it will show you, it will basically create [14:43.320 --> 14:46.960] a system. So you can see this and then it will come up with all these cool charts. So [14:46.960 --> 14:51.280] you can see this is like a chart, this is the output showing me doing various things [14:51.280 --> 14:56.240] and seeing how it creates spikes in usage. So that's testing something from the outside. [14:56.240 --> 15:01.080] You can also look at something from the inside. So the Firefox Profiler, that will actually [15:01.080 --> 15:05.560] self report its own information. So it will tell you I'm using, I'm using this much power [15:05.560 --> 15:09.520] and this part of my browser is using this much. That's really, really, really cool. And there's [15:09.520 --> 15:13.880] a whole talk about that later. We did some work with them to actually turn some of these [15:13.880 --> 15:18.440] into carbon metrics. So you get an idea of what these figures are from. And this is the [15:18.440 --> 15:22.920] kind of stuff that my organization tends to do. If you know this, then you can actually [15:22.920 --> 15:27.440] figure out, okay, now that I understand there are hotspots in what I'm doing, what do I [15:27.440 --> 15:33.040] do? And this chart has been shared, or this table has been shared quite a few places. [15:33.040 --> 15:37.960] Basically different languages have different kind of memory footprints. And this might [15:37.960 --> 15:42.640] give the impression that you should just rewrite everything in C or something like that. That's [15:42.640 --> 15:48.200] not the end of the message you should take away. But in many cases, it may be that there [15:48.200 --> 15:53.080] are certain parts that are already being rewritten for you. Or it's worth remembering that you [15:53.080 --> 15:58.480] might be using a language for other reasons than just efficiency. And like if you are building [15:58.480 --> 16:02.400] something super efficient that help people dig for more oil and gas, probably not that [16:02.400 --> 16:08.560] useful. So intensity, can I change how much, how much harm is being done? So as you might [16:08.560 --> 16:13.080] be aware, renewable energy from wind turbines and solar panels depends on the sun and the [16:13.080 --> 16:17.920] wind and they don't always blow. Data sensors are normally on 24 seven though. So how do [16:17.920 --> 16:22.720] you power something that's on with 24 seven with energy like that? The truth is that most [16:22.720 --> 16:27.840] of the time we rely on averages. So when Google says and said that they already run on green [16:27.840 --> 16:33.080] energy, they're basically saying they've used the same amount of energy as they view as they [16:33.080 --> 16:37.120] have they've generated. So they've generated the same amount of energy from green sources [16:37.120 --> 16:43.120] as they've used over the year, right? That sounds fair. Yeah, it's not quite that simple. [16:43.120 --> 16:48.560] Because you know how, say people say you should get eight hours of sleep a day, more or less. [16:48.560 --> 16:53.280] So over a year, which is maybe 9000 hours, that means you should probably get 3000 hours [16:53.280 --> 16:58.320] of sleep, all right? Now if I got 3000 hours of sleep by a beat by sleeping from January [16:58.320 --> 17:02.840] the 1st to April the 1st, then survived on Red Bull and chocolate bars for the rest of [17:02.840 --> 17:09.000] the year, I mean, it's better than no sleep. But I'm not sure it's 100% sustainable, right? [17:09.000 --> 17:12.440] And like this is actually one of the things that we'll be talking about later. So this [17:12.440 --> 17:16.760] is the difference between annual and hourly and why it matters. And as I mentioned before, [17:16.760 --> 17:21.000] if you don't have massive solar farms and you can't invest on all this stuff yourself, [17:21.000 --> 17:24.680] some things you might need to do is think about how and when you run computing jobs [17:24.680 --> 17:30.840] or how you do stuff like this. Now there was a talk all about this. And if you've heard [17:30.840 --> 17:36.680] of Kubernetes, right, there's basically, you can basically say please run this computing [17:36.680 --> 17:41.240] job where it's green, where it's sunny, or please run it when it's when there's lots [17:41.240 --> 17:44.240] of energy on the grid, all right? There's a really good talk about it. So I'm not going [17:44.240 --> 17:49.640] to talk too much about it. Okay. So what I will talk about though is just going a bit [17:49.640 --> 17:54.840] more detail about what our computers run on really, I suppose. So when we think of computers [17:54.840 --> 17:59.920] just using energy right now, what we're really using is a mixture of electricity from different [17:59.920 --> 18:05.200] sources, each with slightly different properties. So if you have some onsite renewables connected [18:05.200 --> 18:09.920] to say a building, that's one source, it'll be very low carbon. But you won't have much [18:09.920 --> 18:13.680] control over when the power comes on because it's basically the wind and the sun that's [18:13.680 --> 18:18.200] deciding this. And you're dependent on the outside environment. If you've ever used a [18:18.200 --> 18:22.000] laptop, you've used battery storage here. And like that means it comes on when you want [18:22.000 --> 18:27.080] it to be. I mean, a laptop that only turns on when it's sunny is would not be that useful. [18:27.080 --> 18:30.840] So this is like what you refer to as dispatchability or things like whether they're dispatchable [18:30.840 --> 18:36.080] or not, all right? But they have limited, they have limited amounts of power, or sometimes, [18:36.080 --> 18:40.960] but it's getting a lot better. Finally, there's the grid energy. Sometimes it's clean, sometimes [18:40.960 --> 18:46.200] it's dirty, and you can normally rely on it. But when like awful geopolitical political [18:46.200 --> 18:51.800] events happen, they kind of change how much it costs, both in human terms and in actual [18:51.800 --> 18:55.840] kind of financial terms. Now, I apologize for going super nerdy on some of this stuff, [18:55.840 --> 19:00.040] but if you've ever used a virtual machine, you'll understand that what we do with like [19:00.040 --> 19:05.800] virtual machines is we take a big server, and then we use virtualization to turn it [19:05.800 --> 19:12.040] into kind of abstract set of resources like compute, like computing power, network, RAM [19:12.040 --> 19:17.560] and storage. And once we've got this kind of broken down way of thinking about it, we [19:17.560 --> 19:21.600] can then turn it into a series of like smaller machines. So there might be one virtual machine [19:21.600 --> 19:26.440] that has lots of RAM, but only a little bit of storage. And by doing that, I can basically [19:26.440 --> 19:31.340] make much, much more efficient use if I have a set of computers. And this idea of like breaking [19:31.340 --> 19:35.920] things down and breaking them up again, I think is one thing that you can use when you [19:35.920 --> 19:39.480] think about energy as well. And there was a really cool paper that was published last [19:39.480 --> 19:46.280] year about this idea of ecovisors. And the general idea is that you can do the same with [19:46.280 --> 19:51.960] power as what we've been doing with like computing resources. So if you virtualize the power [19:51.960 --> 19:57.000] as a kind of grid made up of three different things, that you can allocate maybe a significant [19:57.000 --> 20:01.760] amount of battery for one program that needs to be on all the time, and maybe less to another [20:01.760 --> 20:06.400] which doesn't need to be on all the time. So you can basically allocate it in these [20:06.400 --> 20:11.160] different ways. And this allows you to then do something like this. You could basically [20:11.160 --> 20:16.400] create virtual machines with certain amounts of virtualized power that you can actually [20:16.400 --> 20:23.520] work with. And this is actually something that you can then share visibility of to the machines [20:23.520 --> 20:27.920] because computers generally just aware of them having power. They don't really know [20:27.920 --> 20:31.800] about what the kind of power is like, whether it's green power, whether it's battery or [20:31.800 --> 20:37.440] also one. But you can make some of this available to people to see. And there are ways that [20:37.440 --> 20:42.880] you can implement like an API. So a program can be aware of its own power that it's using. [20:42.880 --> 20:48.080] And that allows you to then actually basically design a system so it's aware of this and [20:48.080 --> 20:51.680] can make much, much better use of this or turn things off when it doesn't need or actually [20:51.680 --> 20:55.920] use more when there's an abundance of green energy, for example. So that's kind of one [20:55.920 --> 21:00.520] of the ideas which I think is super exciting that haven't seen anywhere else. You can also [21:00.520 --> 21:03.560] do some cool stuff with networking, carbon away networking. And I've got a few more slides [21:03.560 --> 21:07.440] so I'm going to have to run through this quickly. But the short version is, if I want to visit [21:07.440 --> 21:12.040] a website in England, sorry, and the website and the servers in Poland, that's to go all [21:12.040 --> 21:16.160] the way through here. Now, every single time there's going to be different kinds of energy [21:16.160 --> 21:21.680] that I use. But if the internet was aware of the network, I could go, I could take different [21:21.680 --> 21:26.520] routes which are cleaner and greener and some kind of like low carbon trick shot thing. [21:26.520 --> 21:31.680] So this was actually was A, this exists right now. There's a network program called Sion [21:31.680 --> 21:36.720] that does something like this already. There's also a way to extend IPv6 like this. We shared [21:36.720 --> 21:42.840] a paper specifically to this IETF internet architecture board workshop all about this [21:42.840 --> 21:47.720] stuff. There's 26 papers all about this and it's super exciting and interesting in my [21:47.720 --> 21:54.040] book. And a funny direction, can I change where we're headed? So we work on this idea, [21:54.040 --> 21:57.800] a fossil free incident by 2030 because we think the internet should be a global public [21:57.800 --> 22:02.320] good and it's healthy for the planet and you should be healthy for people who use it. And [22:02.320 --> 22:06.880] do you remember I spoke about the idea of hourly figures like annual versus hourly? Google [22:06.880 --> 22:11.400] as one company, they've actually put a lot of time and money behind the idea of saying, [22:11.400 --> 22:15.800] well, we're going to run 24 seven carbon free energy, none of that red bull and chocolate [22:15.800 --> 22:19.840] stuff. We're going to do it properly, which is good. And this is something that we didn't [22:19.840 --> 22:23.400] realize was possible. Microsoft is now doing the same thing as again, but they're coming [22:23.400 --> 22:28.440] up with their own words because they can't possibly have the same words for some reason. [22:28.440 --> 22:32.280] And I thought, wow, this is really, really cool. And then I found out literally like [22:32.280 --> 22:39.160] a month ago that this company here, Peninsula Clean Energy, they're a small nonprofit organization. [22:39.160 --> 22:44.680] They've said, oh yeah, we're already at 99% 24 seven matched and we reckon we'll do it [22:44.680 --> 22:52.000] by 2025. So this is a small company, maybe that is reaching say 100,000 people in California. [22:52.000 --> 22:56.240] They've basically built this into their governance to say, no, this is a priority over share buy [22:56.240 --> 23:00.640] backs, dividends to sharehold is everything like that. It's about intention. And they've [23:00.640 --> 23:05.040] basically got there. But what's super cool and for the FOSDEM group is that they've open [23:05.040 --> 23:08.840] sourced how they're doing it. They've actually shared the model they use to actually figure [23:08.840 --> 23:13.040] out where they buy the power from and the ideas behind this. And like, I think this is [23:13.040 --> 23:16.640] pretty freaking cool. And it's like, basically why am I excited about open source and everything [23:16.640 --> 23:23.080] like that? Finally, I've only used American examples so far. But now I can't say that this [23:23.080 --> 23:27.720] was entirely Europe's plan because Nord Stream was kind of switched off by like Putin and [23:27.720 --> 23:31.200] everything like that. But they've span it this way. They've basically said, well, we've [23:31.200 --> 23:35.600] completely got rid of our dependence on Russian fossil fuels now. It went much faster than [23:35.600 --> 23:41.360] we expected. So we now have the possibility to redirect the money we would have spent [23:41.360 --> 23:47.920] on gas to repower EU, 250 billion euros to net zero industries. So this is basically [23:47.920 --> 23:53.600] a quarter of a trillion euros. So basically to talk to deploy the kind of cool green stuff [23:53.600 --> 23:58.240] rather than the bad brown stuff that I think we should be getting away from. So that's [23:58.240 --> 24:03.040] the my recap. I wanted to share a model with you, but I call CID consumption, intensity [24:03.040 --> 24:07.360] and direction. Can I change how much we need? Can I change how much harm is done? Can I [24:07.360 --> 24:11.800] change where we're headed? And I want to tell you about a fossil free internet by 2030. [24:11.800 --> 24:16.200] It's achievable. We've seen that big firms do this, but small firms do it and faster [24:16.200 --> 24:22.000] and better. We've seen how it basically saves carbon. We know that doing this saves lives. [24:22.000 --> 24:26.680] It saves money because we know that fossil fuels are actually really, really expensive. [24:26.680 --> 24:29.480] And I've already done this slide once. I'm just going to skip it because I'm running [24:29.480 --> 24:34.560] out of time. But folks, I think that's the time I've used. So if you want to know more, [24:34.560 --> 24:38.240] our organization, the Greenbelt Foundation, we publish open source code, we publish open [24:38.240 --> 24:42.120] data and we work with other organizations who are trying to do the same. We also offer [24:42.120 --> 24:47.360] training and consulting. If I've been speaking too quickly, there is a transcript of this [24:47.360 --> 24:54.640] entire talk at that website. And I'm also online as Mr. Chris Adams on Twitter, but [24:54.640 --> 25:00.200] less so these days. And on Mastodon, social was there. And also if you use email, I use [25:00.200 --> 25:07.200] email too. Yeah. Thanks, folks.