[00:00.000 --> 00:11.440] Who can hear me? If you can hear me at the back froming through the sound, can you, okay, [00:11.440 --> 00:17.600] it's, I think, it sounds like you can hear me at the back. Excellent, all right. Okay, [00:17.600 --> 00:21.840] well, it turns out I'm a little bit early, but I'm going to use those two minutes for [00:21.840 --> 00:27.440] questions at the end of the talk, if possible. All right, folks. Welcome to Responsible Clouds [00:27.440 --> 00:34.320] and the Green Web Triangle, how to make the climate case for a diverse cloud ecosystem. [00:34.320 --> 00:39.720] I work for the Green Web Foundation.org and the URL for that non-profit is written across [00:39.720 --> 00:47.000] the bottom of this screen, as you can see. In case you can't see me, that's my face. [00:47.000 --> 00:50.440] This is a bit of a hangover from when you're doing remote talks because, yeah, it's weird [00:50.440 --> 00:56.200] being in a room with other people again. I have a background working in wacky environmental [00:56.200 --> 01:01.160] climate, climate tech startups, like as you can see, Loco 2, which is a triple layer pun [01:01.160 --> 01:07.440] around locomotion, low carbon, and like going a local on a crazy holiday. I also worked at [01:07.440 --> 01:11.280] a company called Amy, which stands for Avoid Mass Extinction Engine. We burned through [01:11.280 --> 01:16.600] something in the region of 20 million US dollars of VC money trying to put a carbon footprint [01:16.600 --> 01:21.600] API on anything we could find. These days, I work at the Green Web Foundation, where [01:21.600 --> 01:25.920] we track the transition of the internet away from fossil fuels to something greener. I [01:25.920 --> 01:30.560] also work with the Green Software Foundation, which is a larger industry body, where I'm [01:30.560 --> 01:36.560] the policy chair there. I also run a podcast. I co-host a podcast called Environment Variables. [01:36.560 --> 01:43.720] Get it? Yeah, which is all about green software. I also am a contributing editor to a magazine [01:43.720 --> 01:47.200] called Branch Magazine, which is all about the stuff, this kind of intersection between [01:47.200 --> 01:52.000] climate and tech. This is what we're going to cover today. I'm going to frame a kind [01:52.000 --> 01:56.600] of problem that you might think about when it comes to sovereign tech, sovereign cloud [01:56.600 --> 02:00.920] and sustainability. I'm going to talk a little bit about the drivers for sustainability at [02:00.920 --> 02:05.640] kind of regulatory level that happen in Europe specifically. Then I'm going to kind of talk [02:05.640 --> 02:11.880] about the idea of competing on transparency as a basis for a more diverse cloud ecosystem. [02:11.880 --> 02:18.720] So off we go. Framing the cloud, framing the problem with the Green Web Triangle. Most [02:18.720 --> 02:23.960] of you know that we basically build services by compiling them these days, lots of different [02:23.960 --> 02:28.400] services. So it's a bit like a kind of graph where things people see at the top is actually [02:28.400 --> 02:33.680] comprised of all these things that we might build all the way through that. When you're [02:33.680 --> 02:39.880] trying to actually use a service, you kind of end up with this kind of problem. If you [02:39.880 --> 02:44.400] want to make a service available to other people, or if you want to use a service, there [02:44.400 --> 02:48.880] are kind of like three things you want in my opinion. You want something to be convenient, [02:48.880 --> 02:52.240] like ideally hosted by someone else a lot of the time or at least so you don't have [02:52.240 --> 02:56.080] to be doing all yourself if you're a kind of medium to small size company. You probably [02:56.080 --> 02:59.440] want it to be fossil free because there's a whole kind of climate emergency thing going [02:59.440 --> 03:04.560] on. And ideally you'd like it to be kind of diverse so you're not dependent on only one [03:04.560 --> 03:09.800] provider who can jack up their prices or do all kinds of things that bad things happen [03:09.800 --> 03:15.240] when you're only relying on one provider. And you basically have, I think there are [03:15.240 --> 03:19.120] different failure modes when you think about this triangle because of most of the time [03:19.120 --> 03:24.720] you can only pick two. So for example, let's say you wanted to go for convenient and diverse. [03:24.720 --> 03:28.880] This is like the common thing that a lot of us kind of default to. Like it's basically [03:28.880 --> 03:34.800] the default, it's fossil fuel powered and it's a bit like being in a spaceship that [03:34.800 --> 03:39.520] you know is on fire and then you go to work and you type in a keyboard and you know things [03:39.520 --> 03:43.680] were on fire. Then you go home and you realize things were on fire. Then you feel kind of [03:43.680 --> 03:47.800] a little bit anxious. Unlike people call that climate anxiety. Like that's one of the things [03:47.800 --> 03:51.840] that some of you might feel when you think about this. But on the kind of like global [03:51.840 --> 03:56.120] scale there's an issue there. So that's the thing that we probably are in right now that [03:56.120 --> 03:59.960] we don't want to do all the time. The other one is diverse and fossil free. Maybe you [03:59.960 --> 04:04.600] do care about this and you're going to look for say green providers. But you also care [04:04.600 --> 04:09.360] about having diverse ecosystems and building or everything yourself. This is cool. But [04:09.360 --> 04:13.600] it kind of has another failure mode in that you spend loads and loads of time on undifferentiated [04:13.600 --> 04:18.720] heavy lifting which isn't necessarily creating the kind of value that you're basically employed [04:18.720 --> 04:23.960] for a lot of the time. Or alternatively there is you can go for convenient and fossil free [04:23.960 --> 04:28.680] because lots and lots of large companies now like say Microsoft and Amazon and Google [04:28.680 --> 04:33.880] are telling a really really like loud story about how much they're really kind of like [04:33.880 --> 04:38.840] running on green energy and doing a load load for the climate. Now there are kind of all [04:38.840 --> 04:42.960] kinds of problems which I'm in the middle. I'm in a sovereign cloud dev room where I [04:42.960 --> 04:47.480] don't need to kind of talk too much about. But there is also issues because these folks [04:47.480 --> 04:52.640] tend to be the people who are often the market leaders in helping people get oil and gas out [04:52.640 --> 04:57.840] of the ground and burn it. So even if you're running on say green software like they may [04:57.840 --> 05:02.000] be doing things which you might not be aligned with as an organization or you might feel [05:02.000 --> 05:07.440] somewhat uncomfortable about. So these are the three ways. These are the triangle and [05:07.440 --> 05:11.400] these are the kind of common failure modes that I think that are useful to talk to other [05:11.400 --> 05:17.160] people about and how I explain the problem with basically sustainable digital diverse [05:17.160 --> 05:20.720] services to people who want to actually make a switch and start using something which is [05:20.720 --> 05:25.960] not necessarily a number of really large providers but don't understand why you can't [05:25.960 --> 05:31.240] just move to like next cloud for everything. So that's like the problem that's the triangle. [05:31.240 --> 05:36.240] Now let's talk about the drivers for sustainability in tech at a regulatory level that might kind [05:36.240 --> 05:42.960] of help for some of this. So this graph is from Dr. Robert Rode from Berkeley Earth which [05:42.960 --> 05:47.840] is an environmental data science nonprofit and they're using data from the global carbon [05:47.840 --> 05:52.720] project and what you can see is historically like we've been growing in emissions however [05:52.720 --> 05:58.560] long and you can see this trajectory which people might refer to as business as usual. [05:58.560 --> 06:04.920] There's lots of different scenarios and it's actually titled SSP2. There's a lot of these. [06:04.920 --> 06:10.000] Can you see the kind of changes we would need to actually make to get down to like 1.5 degrees [06:10.000 --> 06:14.960] of global warming which is considered like the least worst option that you will typically [06:14.960 --> 06:22.560] see a lot of the time and this is you might have heard like 1.5 versus or 2 degrees for [06:22.560 --> 06:27.720] example like in 2015 pretty much every country on earth agreed that we should probably do [06:27.720 --> 06:32.640] something about climate and we should go for not more than 2 degrees of warming and then [06:32.640 --> 06:37.160] they actually you had a bunch of people say well actually 2 degrees is still pretty bad [06:37.160 --> 06:41.560] and then we did a bunch of research and found out that yes 1.5 is what we should be thinking [06:41.560 --> 06:48.360] about naturally and despite that we've actually had people thinking okay well what does that [06:48.360 --> 06:56.640] mean for us as a sector as a technology sector and in 2020 you had a bunch of digital and [06:56.640 --> 07:00.760] sustainability focused organizations trying to figure out what that would mean for us [07:00.760 --> 07:05.480] as a sector like how would how would we need to kind of drop our emissions for example [07:05.480 --> 07:09.080] and as you can see this quote that came out when it was issued in this from this press [07:09.080 --> 07:13.880] release basically said yeah to actually have the Paris Agreement which is less ambitious [07:13.880 --> 07:19.480] than that 1.5 thing they basically said you need to halve emissions by 20 you need to [07:19.480 --> 07:24.800] reduce emissions by 45% by 2030 in seven years time so that's basically making reduction [07:24.800 --> 07:31.640] of around 7% each year year on year and well we haven't really been reducing things by [07:31.640 --> 07:36.720] 7% each year so it's actually going to be larger so this is basically what the science [07:36.720 --> 07:42.000] is spelling out for us the thing that's quite interesting this year is that you've had people [07:42.000 --> 07:45.760] talking about net zero for a while but now people are starting to actually say well what [07:45.760 --> 07:50.480] does net zero really mean in the context of this science actually and has anyone heard [07:50.480 --> 07:55.480] of the ISO here ISO with standards folks okay they come up with something really interesting [07:55.480 --> 07:59.880] that they published in November they basically said if you want to call something net zero [07:59.880 --> 08:03.800] you actually need to have some teeth behind it and needs to actually mean something now [08:03.800 --> 08:07.760] so while there's been lots and lots of kind of greenwashing net zero targets and things [08:07.760 --> 08:11.840] from those organizations they've basically come out and said actually this is the kind [08:11.840 --> 08:16.520] of stuff you need for net zero now and they've basically said if you have net zero claims [08:16.520 --> 08:22.680] by 2050 without having about anything related to harbing missions by 2030 they're not credible [08:22.680 --> 08:26.400] and if you don't include your supply chain to just only think about your own things rather [08:26.400 --> 08:30.880] than like who made a computer or anything like that not credible and also if you don't [08:30.880 --> 08:35.880] have any interim targets in the next few years also not credible because this is quite important [08:35.880 --> 08:41.320] because seven years is about the same time as the average tenure for a CEO which basically [08:41.320 --> 08:45.320] means you can say oh we're going to be net zero leave then it's someone else's problem [08:45.320 --> 08:48.680] to solve and this is them basically saying actually you need to have something now so [08:48.680 --> 08:53.920] you can't do that kind of stuff. What's also interesting in Europe specifically is the [08:53.920 --> 08:59.800] catchly titled European Corporate Sustainability Reporting Directive. Basically if you were [08:59.800 --> 09:05.440] in a company which has more than 250 employees you need to be recording and reporting some [09:05.440 --> 09:10.000] information about what your carbon emissions look like and you need to start reporting [09:10.000 --> 09:16.800] in 2024 which means you need to be reporting two months ago to start doing this. So that [09:16.800 --> 09:22.160] is interesting because lots of lots of organizations don't even have a sustainability team and [09:22.160 --> 09:27.520] don't even know how to start reporting this stuff and then finally the IFRS sorry about [09:27.520 --> 09:33.320] all the acronyms these folks are you can think of them as like the pope of accounting right [09:33.320 --> 09:37.680] they're the people who just basically set a decree which everyone then follows in accounting [09:37.680 --> 09:42.240] land. They've basically said that supply chain thing that people talk about yeah actually [09:42.240 --> 09:46.960] we should be doing that and you can't just say about your own organization rather than [09:46.960 --> 09:51.880] all the things you buy from other people for example. So that's what people have been saying [09:51.880 --> 09:56.120] and how are we doing so far like as a tech sector right. So there's good news and there's [09:56.120 --> 10:02.200] bad news all right the good news is is that well we have seen lots and lots of use and [10:02.200 --> 10:06.400] like more people are connected to the internet which I think is in broadly speaking a good [10:06.400 --> 10:10.120] thing and I don't know about you folks but I quite liked having access to the internet [10:10.120 --> 10:13.840] during the pandemic. I think that's been more useful and it'll be a lot more kind of lonely [10:13.840 --> 10:20.320] and sad and cold and everything if I wasn't connected to that and you generally see that [10:20.320 --> 10:23.960] we've basically been using much much more in the way of digital services over the last [10:23.960 --> 10:30.040] say 10 years or so but the growth hasn't been that large really it hasn't been proportional [10:30.040 --> 10:35.520] and there's a number of reasons for this. Let's okay we can just can we just not talk [10:35.520 --> 10:41.520] about cryptocurrency I'm sorry like it's not helping billions of people right the two billion [10:41.520 --> 10:45.560] people who joined the internet pretty sure didn't help two billion people yes using like [10:45.560 --> 10:50.560] about half the impact half the energy of all the data centers on earth all right. So that's [10:50.560 --> 10:54.280] the thing so we're basically going in the direction which is not ideal at the moment [10:54.280 --> 10:59.920] and we need to kind of be reducing this and I guess there are now some companies which [10:59.920 --> 11:02.760] are trying to step up on this and they're trying to basically say well we want to keep [11:02.760 --> 11:07.480] growing which has issues of its own but maybe one thing we can do is actually make the energy [11:07.480 --> 11:13.200] we use greener. So you now see very very large companies like say Google say well 2030 is [11:13.200 --> 11:17.880] an important date we're going to make sure that all of our energy is running on or the [11:17.880 --> 11:24.840] energy we use is coming from green sources 24-7 and 24-7 is important because okay you [11:24.840 --> 11:28.920] know how like you bill for cloud like an hourly basis and there's about maybe 9000 hours [11:28.920 --> 11:33.000] in a year yeah all right. Normally when you look at green energy and people say I'm running [11:33.000 --> 11:38.800] on green energy it's like an average basically saying for all the power I've used if I've [11:38.800 --> 11:46.560] used maybe actually the way I describe it to people is like this we all know that we [11:46.560 --> 11:51.320] should probably get about eight hours of sleep right yeah like a or around that in a 24 hour [11:51.320 --> 11:57.080] period and if I wanted when you normally have green energy right now you can basically say [11:57.080 --> 12:02.440] well I've it's a bit like me saying I'm no only to get eight hours of sleep every night [12:02.440 --> 12:07.480] and I've got a year so what I'm going to do is I'm going to have 3000 hours of sleep at [12:07.480 --> 12:11.840] the beginning of the year all right and then I'm just going to run on red bull and chocolate [12:11.840 --> 12:15.800] bars for the rest of the year and that's going to be sustainable that's kind of like annual [12:15.800 --> 12:19.440] energy balancing that's kind of what the accounting is going to be like and there's more to it [12:19.440 --> 12:23.080] than that but like that's what we currently have at the moment which is why 24-7 is actually [12:23.080 --> 12:27.920] allowable and quite an important goal Microsoft is also doing this they've said well they're [12:27.920 --> 12:33.800] going to run a hundred percent a hundred percent zero coming with zero carbon purchases which [12:33.800 --> 12:38.640] again sounds kind of cool and like these are companies with almost infinite money I say [12:38.640 --> 12:43.960] almost because they use as a basis for firing lots lots of people despite being really really [12:43.960 --> 12:50.640] profitable still but there it turns out that actually you don't need to be a ginormous [12:50.640 --> 12:53.720] company with infinite money to actually kind of set this kind of target and actually be [12:53.720 --> 12:59.320] moving this quickly this is peninsula clean energy they're based in California they're [12:59.320 --> 13:03.720] a small community choice aggregator which is basically a bit like a non-profit energy [13:03.720 --> 13:08.280] company and they've said well actually we think we think we can do it by 2025 because [13:08.280 --> 13:15.680] we built the idea of avoiding an absolute climate kind of crisis into our governance [13:15.680 --> 13:20.120] board basically and we if it makes us slightly less profitable that's okay because we're [13:20.120 --> 13:23.440] not answering to shareholders so they show that they can do that and they've also actually [13:23.440 --> 13:27.680] shared all the actual underlying model in python to show how they actually bought all [13:27.680 --> 13:33.400] their stuff as well so you can actually do this and there's something interesting in [13:33.400 --> 13:37.240] Europe that's happened this in the last week actually do you know how like Elon Musk did [13:37.240 --> 13:40.920] that whole thing where he just cut off API access to Twitter and then everyone's like [13:40.920 --> 13:44.760] oh now this thing I was going to do one like that's a massive enough start using now because [13:44.760 --> 13:48.400] I should be using and I meant to do it anyway right there's this kind of my idea that you [13:48.400 --> 13:53.120] would get around to it kind of the same thing happened with fossil fuels in Europe right [13:53.120 --> 13:57.280] so we basically had this idea where Europe has been using loads of fossil fuels from [13:57.280 --> 14:02.720] Russia and then Russia basically switched off the pipe and now you basically see press [14:02.720 --> 14:06.680] releases like this saying well we've got off Russia we've got off Russian gas folks and [14:06.680 --> 14:09.920] now we've got all this money we can spend on that we would have otherwise spent on on [14:09.920 --> 14:15.040] fossil fuels so it's a bit like lemon and making lemonade but basically the upshot is [14:15.040 --> 14:19.280] that you basically you basically got the European Commission president and the people who essentially [14:19.280 --> 14:23.920] run Europe saying well we've got a quarter of a trillion euros that we would have spent on gas [14:23.920 --> 14:28.160] and now we're planning to spend it on kind of more cool things like net zero technology and like [14:28.160 --> 14:34.480] wind and solar and stuff which I think is actually well I'll take it I suppose all right so that's [14:34.480 --> 14:37.920] kind of what's happening at kind of regulatory level but I think it's quite interesting and [14:37.920 --> 14:43.840] anyway let's talk about competing on transparency for a diverse cloud ecosystem because the thing [14:43.840 --> 14:48.320] is we basically have the science telling us we need to reduce emissions and there are ways [14:48.320 --> 14:51.760] of there are some organizations which start saying yeah we're making all this we're doing [14:51.760 --> 14:55.760] all this kind of big splashy stuff and it's turning up there's more in the way of resources to do [14:55.760 --> 15:02.080] this stuff and there is this kind of narrative in policy circles right now where all the innovation [15:02.080 --> 15:07.280] has to come from very very large companies and I think it's not really true particularly in like [15:07.280 --> 15:11.680] well basic cloud for example there's low all kinds of cool stuff that you see from small providers [15:11.680 --> 15:17.360] in Europe for example so OVH cloud pioneers in say liquid cooling and servers as you can see here [15:17.360 --> 15:22.240] and then we're doing it for ages scale way for like a french cloud provider using arm service [15:22.240 --> 15:25.840] for ages as well like this is the thing that people make a big thing of but it's been like just [15:25.840 --> 15:31.040] standard as well as okay adiabatic cooling is basically using some ideas first developed in [15:31.040 --> 15:35.920] Iran to basically reduce the use of water but still make things cool they basically use it inside [15:35.920 --> 15:40.160] their servers rather than drawing loads and loads of water out of like the local kind of aquifers [15:40.160 --> 15:44.960] this is important because in some parts of the world in there are google data centers which use [15:44.960 --> 15:50.080] the same amount of water as the city they're in basically so like you should be caring about this [15:50.080 --> 15:53.440] kind of stuff and anyway there's loads of other examples that I've listed here I'm sure you can [15:53.440 --> 15:59.280] read them as well and there's also an interest at a kind of regulatory or kind of public sector [15:59.280 --> 16:03.600] level now where people are trying to actually figure out what sustainable software really looks [16:03.600 --> 16:09.520] like and how do you standardize it and how do you procure for this so ocular which was released by [16:09.520 --> 16:15.360] which was kind of celebrated by kde recently they have the first kind of blau angle blue angel kind [16:15.360 --> 16:21.200] of sustainable software kind of certificate which was issued to ocular pdf reader around [16:21.200 --> 16:25.600] september last year and this is interesting because this is the kind of stuff that people [16:25.600 --> 16:30.240] who spend money on trying to actually procure green and actually buy green and digital services [16:30.240 --> 16:36.240] are actually getting now and if you work in a company and you're using some of the large providers [16:36.240 --> 16:40.960] this is important because it turns out to be quite difficult to get the metrics from the very [16:40.960 --> 16:46.320] very big hyperscalers this is a this is actually a table from a person called mark butcher at [16:46.320 --> 16:50.240] positive cloud he shared some of this with me because he was trying to figure out well okay if [16:50.240 --> 16:54.960] you're using these big providers what information can you get from them who wait when they're in [16:54.960 --> 17:00.000] your supply chain that you need to report against and well basically I won't go into all the details [17:00.000 --> 17:04.960] here but the red stuff is bad right so if you're using the stuff on like those two providers [17:04.960 --> 17:10.160] not really a good one even on the left hand side you do see some information so microsoft is probably [17:10.160 --> 17:15.200] one of the better ones right now but they're still there's still some questions about what they report [17:15.200 --> 17:22.480] amazon the the level of reporting they give you is so so so low resolution that you can't tell if [17:22.480 --> 17:26.560] the emissions are coming from like one data center or another data center which is on the other side [17:26.560 --> 17:31.680] of a country thousands of miles away so you have some issues here because you might have one data [17:31.680 --> 17:35.680] center running essentially on coal and you have another data center running on hydropower which [17:35.680 --> 17:39.760] is definitely cleaner than coal all right and you have all this stuff here and this is one thing [17:39.760 --> 17:43.360] that you don't really have access to that if you're going to try and if you need to report you're [17:43.360 --> 17:47.680] going to have to be able to figure out how to actually get these numbers so you can see if you're [17:47.680 --> 17:52.320] making if we're going up or if you're going down like we saw before there's also loads of really [17:52.320 --> 17:57.280] cool open source tooling which makes it possible to get these numbers out so the most well-known one [17:57.280 --> 18:03.520] is cloud carbon footprint this is maintained by ThoughtWorks it plugs into billing APIs from say [18:03.520 --> 18:08.480] amazon from microsoft and google then they do some guesstimates based on available data to tell you [18:08.480 --> 18:13.920] what your figures might be because the various cloud providers have their own native ones which [18:13.920 --> 18:18.640] have very different assumptions to the open ones that you can challenge and you can ask questions [18:18.640 --> 18:24.080] about green coding metrics they did a talk today they basically put if you have a digital service [18:24.080 --> 18:29.200] they'll put every single service in a kind of docker container then track the energy used and [18:29.200 --> 18:33.360] i've got a really nerdy kind of part of a talk that five o'clock later today where i talk about [18:33.360 --> 18:39.600] some of that scaffold i'm not going to pronounce that properly it's a french word for diving helmet [18:39.600 --> 18:45.360] and or diving suit it basically gives you process level energy usage metrics from any machine you [18:45.360 --> 18:49.520] run it on if you have access to the to the bare metal and it spits it out that you can kind of [18:49.520 --> 18:54.400] put into like dashboards and grafana kepler does more or less the same thing but for kubernetes [18:54.400 --> 18:59.680] using Berkeley you know extended Berkeley packet filters that's what they're called right yeah [18:59.680 --> 19:02.400] so that's another thing they do and there's another talk again about that kind of stuff [19:03.760 --> 19:07.280] so there is all this stuff which you can use and plug into your tooling where you can ask [19:07.280 --> 19:11.360] people to use so you can actually get the numbs to show that you're kind of putting us in the [19:11.360 --> 19:15.200] right direction if you're interested in and i think that's my time i'm just going to wrap up [19:15.200 --> 19:19.760] that was a green web triangle there are three things you can have convenient diverse fossil [19:19.760 --> 19:24.480] free most of the time you can only really pick two there's pressure right now to really push [19:24.480 --> 19:29.120] towards fossil free being a key thing which then really makes you decide am i going to be diverse [19:29.120 --> 19:34.800] am i going to be fossil free i mean the other one convenient you can compete on transparency and i [19:34.800 --> 19:40.480] think this is some area that you could actually kind of really kind of lead on and i think this is [19:40.480 --> 19:46.080] actually a way that this is this is probably underexplored right now and finally diverse [19:46.080 --> 19:50.080] ecosystems are healthy ecosystems there's policy support and there's money and there's funding [19:50.080 --> 19:55.200] to make this stuff possible and i think you should do it and i think that there's actually like [19:55.200 --> 19:59.440] even a policy idea around this the kind of word you need to use is twin transitions that's the kind [19:59.440 --> 20:04.240] of pseudo password you need to use when you speak into a policy maker and i think that's it thanks [20:04.240 --> 20:09.040] folks for my time there's a talk about a fossil free internet this talk is online with links to [20:09.040 --> 20:14.960] everything i showed you i'm on mastodon.social and okay there's probably mr chris helms on twitter [20:14.960 --> 20:18.640] but i don't really use it very much but you're welcome to tweet there i suppose and yeah that's [20:18.640 --> 20:29.280] my email address thank you everyone thank you chris um open for questions yes [20:32.320 --> 20:35.840] i'm also in like the element did matrix thing oh so if you've got a question there i'll try and [20:35.840 --> 20:46.560] answer it as well well hello i suppose i waited for some serious questions but i have one maybe [20:46.560 --> 20:53.120] a weird one what do you think about low earth orbit data centers you've talked about data [20:53.120 --> 21:06.080] centers in space that beam things down exactly uh i think they're if there was a way that you could [21:06.080 --> 21:10.320] get them into space without the rockets and you kind of there are actually some ways some wacky [21:10.320 --> 21:16.080] ways to do that i think it'd be cool i think astronomers have really strong opinions about [21:16.080 --> 21:21.360] that and also the whole idea of like privatization of space makes me feel a little bit uncomfortable [21:21.360 --> 21:25.040] because we're not great at dealing with our waste right now and if you have like all the waste that [21:25.040 --> 21:31.600] stops other people trans then i feel like that's maybe not the ideal solution basically also a [21:31.600 --> 21:37.120] little bit worry about having the guy in charge of twitter in charge of the future of space maybe [21:37.120 --> 21:40.640] like let's look at like how it's how he's handling twitter and think we want to kind of [21:40.640 --> 21:45.360] see where that goes i'm not sure about that that's that's my answer i suppose but probably not related [21:45.360 --> 21:53.600] to climate anyone else [21:59.840 --> 22:06.960] hi i just wanted to ask what can you do as a small company 10 to 15 people you are developing [22:06.960 --> 22:15.360] software probably for bigger clients i think what you can do you can write into your contract [22:15.360 --> 22:18.960] actually there's a website called the chance relaying project which has all these clauses [22:18.960 --> 22:24.800] for working with larger companies you can also allocate some of your revenue 1.5 percent of my [22:24.800 --> 22:31.120] revenue for 1.5 percent of us being alive and then use that to allocate that to fund things [22:31.120 --> 22:35.680] either internally making your own sustain your own organization greener or by funding things like [22:35.680 --> 22:39.920] non-profits or groups who are actually pushing for quality change stuff like that i think those are [22:39.920 --> 22:44.160] the things you could do which are easy to like write into a contract which would actually show [22:44.160 --> 22:48.720] that you were demonstrating leadership compared to other people this is also what we do in terms of [22:48.720 --> 22:52.320] training and consulting i'm very happy to chat about that because oh yeah i probably should have [22:52.320 --> 22:56.000] told told you everyone that yeah like this is what we do as well but yeah i'm happy to chat about [22:56.000 --> 23:06.000] that in more in more detail or answer questions on the element and stuff from that okay [23:10.000 --> 23:17.760] thank you hi thank you um you showed a few list of tools that help you get um energy yeah this one [23:17.760 --> 23:27.040] yes um how how much are how precise are they and how much do they go around like how how many [23:27.040 --> 23:32.480] assumptions do they make to get the numbers and what would be needed to don't make assumptions at all [23:32.480 --> 23:37.600] okay someone mentioned ilia before didn't they so ilia was like one of the german cloud providers [23:37.600 --> 23:43.200] i was sorry not cloud provider um energy provider um what we need to do is we need to get carbon [23:43.200 --> 23:47.840] emission figures from the energy providers so like they can tell you how dirty or clean the energy [23:47.840 --> 23:52.800] is at any moment in time so you need that basically scafandra is very very very good because it gives [23:52.800 --> 23:58.320] you very very good numbers but you need hardware support to actually share the information about [23:58.320 --> 24:03.760] this so intel they have something called wrapper which is running average power limit which gives you [24:03.760 --> 24:09.040] idea of the energy used by the cpu but not necessarily the gpu or the disk so you need all [24:09.040 --> 24:15.200] this stuff kind of instrumented for you to get this um kepler uh because you're going into the [24:15.200 --> 24:19.920] kernel you can get really really really detailed information but i'm kind of out of my lay at [24:19.920 --> 24:24.880] my league here because i can write docs but i can't write rust which is why i contribute to docs to [24:24.880 --> 24:31.440] scafandra instead of rust to to just scafandra um greed coding metrics there's also a talk online [24:31.440 --> 24:36.160] where the guy taught where arna turrara he talks about this in really really good detail and he's [24:36.160 --> 24:40.320] actually very very knowledgeable and cloud carbon footprint everything is open source and it's all [24:40.320 --> 24:44.960] in type script where they are very explicit about the assumptions they make all along and they're [24:44.960 --> 24:50.640] also open to pull requests i've committed to code to that one and yeah i haven't committed code to [24:50.640 --> 24:54.880] the other two but um they're cool people and i i i really support the projects i hope that helps [24:54.880 --> 25:10.800] and i'm again happy to chat in more detail thank you cool folks thank you