WEBVTT 00:00.000 --> 00:12.000 So everybody, good morning, afternoon, whatever we are now, sorry for the late star. 00:12.000 --> 00:18.600 Turns out that this is an accurate representation of not just the road, some mainstream 00:18.600 --> 00:24.000 matrix, but also the combination of HDMI and USB-C devices before me. 00:24.000 --> 00:29.500 So I am now going to fit a probably 55 minute presentation into about 40 minutes, so buckle 00:29.500 --> 00:32.500 up because we are going for a ride. 00:32.500 --> 00:36.500 So first of all, hopefully everybody knows by now, matrix is an open network for 00:36.500 --> 00:38.500 security, centralised rule, time communication. 00:38.500 --> 00:43.500 The mission being to build the real-time communication network of the open web. 00:43.500 --> 00:47.500 The whole idea is that no single party into your conversations, they have replicated 00:47.500 --> 00:49.500 shared across all participating servers. 00:49.500 --> 00:54.500 So, different graph this time, looking at the 30 day retained users. 00:54.500 --> 00:58.500 So this is the actual secret metric that we have been staring at, it is literally on a 00:58.500 --> 01:02.500 different platform of dashboard in the will of the Make sure 01:02.500 --> 01:06.500 explanation offices in Chasic in London. 01:06.500 --> 01:11.500 Looking at how many people have actually been hanging around for at least 30 days on the platform 01:11.500 --> 01:12.500 at any given point. 01:12.500 --> 01:16.500 So if they stop using it for more than 30 days, it is clear if they are being added 01:16.500 --> 01:19.500 for more than 30 days, if they are being added from more than 30 days to the tele count. 01:19.500 --> 01:23.500 And this is based on the phone home stats that synapses across all reporting 01:23.500 --> 01:24.500 servers. 01:24.500 --> 01:28.180 You can see that, you know, many years ago now, six years ago, we were down at a couple 01:28.180 --> 01:29.180 of tens of thousands. 01:29.180 --> 01:34.220 It's been growing fairly linearly, we've had some kind of more exponential phases through 01:34.220 --> 01:39.380 to 2022, then a bit of a slow down in 2023, but then I'm happy to say in 2024 things are 01:39.380 --> 01:44.020 moving off, indeed, since the summer things are really heading on as slightly better track 01:44.020 --> 01:45.020 again. 01:45.020 --> 01:50.260 So, we are here today to talk about the road to mainstream matrix, and this is a road that 01:50.260 --> 01:52.820 has both a past and a future. 01:52.820 --> 01:57.820 So, I warn you, if you expect me to talk an awful lot of very geeky stuff about new 01:57.820 --> 02:00.020 MSCs, this is going to be a bit of a mix. 02:00.020 --> 02:05.140 We're going to look at things in the bigger picture for once, so I apologize if anybody 02:05.140 --> 02:09.420 flees in horror at this, but I promise we'll have some fun demos, which definitely 02:09.420 --> 02:12.620 won't work, if I'm doing this off-Ham and didn't slap top, but that's fine, don't 02:12.620 --> 02:13.620 worry about it. 02:13.620 --> 02:19.580 So, first of all, in the past, May 2014, the dream, we sat down and said, hey, why don't 02:19.580 --> 02:22.140 we try to build the most communication there with the web? 02:22.140 --> 02:25.580 I mean, I already had a team of about 10 folks who've been working together for about 10 years 02:25.580 --> 02:31.420 doing lots of sip, we wrote on Ron Vogue Stacks, and all that sort of thing, and our idea 02:31.420 --> 02:35.760 was that we were building implementation first, called Synapse, and in fact, originally 02:35.760 --> 02:40.460 matrix was called Project Synapse, that was the code name, and we would use this implementation 02:40.460 --> 02:45.860 to prove the protocol, work, and try to figure it out, and then have the implement of it 02:45.860 --> 02:49.580 when we spec it, updating the one-true spec. 02:49.580 --> 02:52.060 This Wikipedia is anchored in reality, hopefully. 02:52.060 --> 02:57.780 The whole idea had been that we would prioritize a killer app, a actual real-life production 02:57.780 --> 03:01.860 code, over a theoretical protocol. 03:01.860 --> 03:06.740 And this is, I apologize for the corporate PowerPoint, and this is the original Project Synapse 03:06.740 --> 03:12.220 timeline before it was even called matrix, and you can see, starting in May, we go and 03:12.220 --> 03:15.780 actually build the thing, we launch it at TechCrunch to stroke. 03:15.780 --> 03:21.300 We then go around the world, telling people how amazing it is, including Foz Dam in February 03:21.300 --> 03:27.580 2015, so this is our 10th year Foz Dam, and then we would try to get people to actually 03:27.580 --> 03:32.300 use it, and you can possibly read, it says, here, get the glossy app to differentiate, 03:32.300 --> 03:35.500 which is horrible business speed for saying, we want to have the killer app that people 03:35.500 --> 03:37.500 can actually use. 03:37.500 --> 03:43.220 So 2014 September, here we are at four or five or six o'clock in the morning of the TechCrunch 03:43.220 --> 03:47.180 to stroke hackathon, weirdly enough, nobody is hacking on matrix, because it is about 03:47.180 --> 03:51.300 six hours older at this point, and so we do so instead. 03:51.300 --> 03:56.580 We ship Synapse, we go and ship this in called CyWeb, which became renamed matrix console, 03:56.580 --> 03:58.700 and actually had an awful lot of stuff. 03:58.700 --> 04:04.620 You'd have chat, you'd have one to one vote, you'd get federation in there too, I think 04:04.620 --> 04:08.900 boy got added literally at the last hour, and again the whole idea is it would prove the 04:08.940 --> 04:11.860 implementations and them would spark it. 04:11.860 --> 04:16.940 And it looked like this, so this is a real screenshot from a few days before launch on 04:16.940 --> 04:23.540 September 3rd with a stupid inception joke and the XKCT joke, and you can see even at this 04:23.540 --> 04:28.300 point that people have got their real matrix IDs, you can see Eric on Eric J on JKI 04:28.300 --> 04:33.500 to RE, you can see me as Arathorn on Erisphere.net, and a whole bunch of other people, and 04:33.500 --> 04:37.940 there's ugliest sin, and if anybody can be bothered to click on that hyperlink, you would 04:37.980 --> 04:41.900 go there in your matrix client today, and see there is still there. 04:41.900 --> 04:46.500 The images won't work because this is pre-release, and so obviously we broke the API 04:46.500 --> 04:50.620 at the last minute before we actually released, but it's kind of fun that we can 04:50.620 --> 04:53.220 touch over all the way back there. 04:53.220 --> 05:00.700 Also, during the arc fun, we made animatrix, anybody know what animatrix is? 05:01.700 --> 05:09.140 Okay, about 10 people, so I'll show you about five seconds of a slightly cringy hack fun 05:09.140 --> 05:13.780 pitch showing this thing off, assuming that we've got an African activity. 05:13.780 --> 05:19.500 What we have here is a unity thing where you use in those kinematics to go and puppet 05:19.500 --> 05:24.700 this little character, hit send, and the world's first ever Android matrix client, and 05:24.700 --> 05:31.060 that was literally what became eventually Android, and nowadays, Alan and Android, was 05:31.060 --> 05:36.300 written in order to take this 3D telemetry to find your own animation, a district man going 05:36.300 --> 05:40.780 and wandering across the thing, hit send, send it to the other person, they re-render it. 05:40.780 --> 05:43.740 This is 4 years before the emoji, hit send to add. 05:43.740 --> 05:45.300 So what does this tell us? 05:45.300 --> 05:49.540 We really, really want matrix to be more than chat, we want it to be mainstream, not 05:49.540 --> 05:53.460 just for gigs, no offense, and governments, probably no offense. 05:53.460 --> 05:57.820 We had a lot more dev skills, frankly, than product skills, we had taken building glossy 05:57.820 --> 06:02.140 mainstream apps completely for granted, we'd done it before, but always with lots of support 06:02.140 --> 06:06.140 before we could do it by ourselves, and also worth noting that matrix really began as this 06:06.140 --> 06:10.540 existing, centralized team, unlike a typical force project. 06:10.540 --> 06:12.900 So 2015, how are we going to fund this? 06:12.900 --> 06:16.220 Still need a killer app, so let's go and build something on top of this. 06:16.220 --> 06:20.580 We figured out there are 75 different products you can build on matrix, and really irritating 06:20.580 --> 06:25.860 I can't find the notes from them, but I remember clearly, one of them was, hey, we should 06:25.860 --> 06:33.020 take this animatrix thing and create an increase of visual application on it, or we could 06:33.020 --> 06:39.540 do IoT over matrix, or we could do live video mixing for drones flying around sort of bigger 06:39.540 --> 06:43.940 reeners, and you mix it together on matrix, or we could build virtual worlds on matrix, 06:43.940 --> 06:48.780 we could do dating apps, or we could do gov tech comms, but seriously, it would be crazy 06:48.780 --> 06:53.260 if governments have a bit on matrix, and what we decided in the end was to converge on this 06:53.260 --> 06:57.700 thing that was originally codenamed Skype done right, and then subsequently got reenamed 06:57.700 --> 07:04.540 open hangouts, as the idea of basically professional collaboration out that was open source, 07:04.540 --> 07:06.180 how can it be? 07:06.180 --> 07:11.860 And we called it Vector, a Vector, to push matrix out into the world, and we launched it 07:11.940 --> 07:18.260 Oscar, which was an interesting experience, and that then became right, and became 07:18.260 --> 07:23.740 well-eminent, and had this ugly baby blue colour theme, but that is the same code base 07:23.740 --> 07:28.700 as element web on the day it was born, and to end encryption started at the work at the 07:28.700 --> 07:32.900 start at the same time, I think it's interesting to note, this could be the original 07:32.900 --> 07:38.780 sin of matrix, at least in terms of development philosophy, people who remember matrix 07:38.860 --> 07:43.860 before E2E, but then, I suspect it felt, oh my god, this thing is going incredibly first, 07:43.860 --> 07:47.820 I've got quite a lot of stuff, it's federation, it's amazing, state roads, my even works 07:47.820 --> 07:52.860 on day, and then we added an answer to encryption, and frankly, everything, ground to not 07:52.860 --> 07:56.940 a whole, but probably about 10x, so that we've been going before, because it turns out 07:56.940 --> 08:03.580 the decentralized encryption is rather hurt, then 2015 to 2016, I just note this is premature 08:03.620 --> 08:08.740 success, around this point, matrix is getting real attention, lots of people getting 08:08.740 --> 08:13.980 very excited about this new protocol that will liberate us from the silos, entirely 08:13.980 --> 08:19.540 mindful for hyping it to the heavens and back again, lots of pressure from the community 08:19.540 --> 08:26.420 to improve the spec, because it was very much following what we had been implementing, 08:26.420 --> 08:31.420 and the spec process wasn't there, the governance process wasn't there, it's interesting 08:31.500 --> 08:36.820 because we ended up investing a lot of time in trying to make the spec, something that 08:36.820 --> 08:43.500 everybody could build on, rather than actually investing in making it mainstream, and making 08:43.500 --> 08:47.900 sure that we had a killer app that people could use, and I think it's an interesting and 08:47.900 --> 08:52.540 very controversial viewpoint, and it's probably got people to throw rotten fruits at me, 08:52.540 --> 08:58.540 that perhaps in retrospect, we should have ruthlessly prioritised polishing the app or an app 08:58.620 --> 09:03.820 to drive mainstream adoption rather than perhaps prematurely focusing on the open ecosystem, 09:04.780 --> 09:11.100 so one thought experiment is that matrix itself perhaps should not have been primarily positioned 09:11.100 --> 09:17.020 as a protocol, matrix could have been the name of an app, which was a Trojan horse, 09:17.020 --> 09:21.340 unashamedly for a protocol, so rather than setting the expectations from the outset, 09:21.340 --> 09:25.020 as the missing communication layer of the web, we could have just shipped a really, really good 09:25.100 --> 09:31.340 thing like, say, Skype, which was originally decentralized, imagine if Skype was open source, 09:31.900 --> 09:37.020 and then you could have acted out a protocol from that, and it's interesting seeing what the 09:37.020 --> 09:42.860 blue sky folks are up to with an 80 pro, so, again pretty controversial, but frankly it seems to 09:42.860 --> 09:47.100 be working out for them, that they are effectively smuggling a genuinely decentralized protocol, 09:47.100 --> 09:53.020 and under the allspices of what looks like a centralized communication network. I personally know 09:53.100 --> 09:57.900 the guys and I believe that they are genuinely aiming to decentralise things, but they're basically 09:57.900 --> 10:02.940 following that track that we didn't follow, and as a result, they have a really mainstream 10:02.940 --> 10:09.580 successful decentralized social app, whereas we have been on this Escher-like infinite staircase. 10:10.620 --> 10:18.540 2017, a new vector, if you get the gag, teams that are relevant to the work that have been happening 10:18.620 --> 10:23.100 for matrix. At the same time, we started setting out the matrix foundation as a nonprofit 10:23.100 --> 10:28.300 guardian to make sure that matrix would be independent from element and other commercial matrix 10:28.300 --> 10:34.060 vendors. Community stepped up with donations to the foundation in 2017. Thank you so much, 10:34.060 --> 10:39.420 we would never have been able to raise money element by saying, hey, look, people believe in matrix, 10:39.420 --> 10:45.420 because they donate and fund it, and subsequently element managed to raise cash from status, 10:45.420 --> 10:51.100 and then later from financial vcs and other investors. Something that people don't always seem 10:51.100 --> 10:57.580 to remember is the element has spent 70-80% of its funding on basically building out matrix for 10:57.580 --> 11:02.700 everybody, for many, many, many, many years transferring the IP through to the matrix foundation. 11:03.740 --> 11:09.420 So then, happy days, 2018 to 2021, how's the end days? The matrix grew, matches, we finally 11:09.420 --> 11:15.020 set one-point-owned June 2020, lots of projects and companies and organizations grew on the protocol. 11:15.020 --> 11:19.820 Open source communities, like Foztern, Zeller, KDE, Nov, and so many others go matrix, 11:19.820 --> 11:25.260 and government start rolling it out to provide digital softnessy. The killer apps do start to, 11:25.260 --> 11:29.260 no, you see some hints of killer apps forming, element actually has a design and a product 11:29.260 --> 11:33.980 team and starts fixing the baby blue color scheme we saw earlier. Beeper forks, and then 11:33.980 --> 11:38.300 pre-exaliment, and shows that you can build a really nice killer app on top of matrix and do 11:39.260 --> 11:43.740 lots of amazing experimental projects at the same time, this is all mindful going around saying 11:43.740 --> 11:47.740 right, we're building the missing communication layer of the web, we need to grab as much 11:47.740 --> 11:52.300 attention and funding as possible and show the potential and inspire people to build on it, 11:52.300 --> 11:56.860 when people build the metaverse, they should build it on matrix, if they want to, 11:56.860 --> 12:00.940 they should have local first matrix, we'll build pit appear, we'll do low bandwidth, we'll put 12:00.940 --> 12:06.540 over HF radios, not anyone's in there, we've done the hydrogen for lightweight matrix, 12:06.540 --> 12:10.940 we've got dendrite for next generation, we went and merged guest room to matrix, 12:10.940 --> 12:16.220 3D cooling, VR demos, CRDTs, element drive, applications we on chat, and all of this inspirational 12:16.220 --> 12:22.540 work, I think made for some fun demos at Fozden, I think at least I had fun, but it did still 12:22.540 --> 12:30.140 a lot of energy from actually writing that killer mainstream app. So 2022, the wheels come off, 12:30.140 --> 12:36.380 a series of unfortunate events, markets crash, post-COVID, no more investment really, 12:36.380 --> 12:40.460 element isn't profitable, yeah, and so not on a position to raise much more investment, 12:40.460 --> 12:44.940 having been chucking all of the dollars into matrix, now if we're playing in the ecosystem, 12:44.940 --> 12:49.180 it's in contributing dollars at the same level, but hey, obviously it's going to be fine, 12:49.180 --> 12:54.780 matrix is doing great, huge, strong-up to public sector, governments realise that the idea 12:54.780 --> 13:00.140 that your country is operationally dependent on a private US tech company like Microsoft 13:00.220 --> 13:04.380 is bananas, it makes no sense if you're like Sweden or Somedy. 13:09.660 --> 13:13.580 And honestly, the evidence is obviously mounted since then when you see the terrifying 13:13.580 --> 13:17.740 and easier signing key is going out the door and all of the teams comprises and 13:17.740 --> 13:22.460 exchange compromises there have been basically people who wait up that they need to run their 13:22.460 --> 13:27.580 own comms, which is great. And so I'm and so as hey, in order to keep funding matrix, we'll go and 13:27.580 --> 13:31.580 focus on government implementations in a bad bid to get profitable and sustainable. 13:32.380 --> 13:38.300 So this turns out to have two big failure modes. The government said, we want matrix, so here's 13:38.300 --> 13:44.860 a public tender to the world, please provide us with matrix deployment. Then big system integrators 13:44.860 --> 13:49.740 who spend their lives going and selling big IT projects to governments pitch for the tender. 13:50.300 --> 13:53.660 They're well positioned to win because they do this all the time, they've got local staff, 13:53.660 --> 13:57.660 they've got the billion dollars of revenue they need to qualify, they've got existing contracts, 13:57.660 --> 14:02.620 they know the right people in the ministry and all that sort of thing. And then the tender typically 14:02.620 --> 14:07.740 is awarded to the cheapest proposal. So in order to win the tender and keep the price term, 14:07.740 --> 14:13.740 the SI just follows the law of capitalism and the law of the market and says, right, we will pick 14:13.740 --> 14:19.420 up open source synapses on GitHub and we will deploy it and we will support ourselves and we will take 14:19.420 --> 14:23.420 all of the money from the government for this tender and we will not root any of it to the upstream 14:23.420 --> 14:28.780 project. And you think, hey, okay, that's not ideal. And it happened a few times, 14:29.660 --> 14:38.860 exotin. And it was, and it was no bit cringy, but there we go. It turned out not to be a one-off, 14:39.100 --> 14:45.260 particularly in 2022, basically almost every deal, I think, that we hand on the element side 14:45.260 --> 14:50.220 for providing matrix to governments ended up disappearing to somebody who basically could win it 14:50.220 --> 14:58.060 because they didn't have the cost of developing matrix. So this is three emojis of skull barred. 14:59.340 --> 15:03.980 Failure mode two is a bit different. The government just says, hey, the FSFee have told me 15:03.980 --> 15:08.540 that public money for public codes, this is a great idea. We want to make sure that all of our 15:08.540 --> 15:13.660 development will be open source, but we don't have budget for maintenance or support fees. So what 15:13.660 --> 15:18.220 would do is that we'll run this ourselves, otherwise, you know, we can't spend money on sort of 15:18.220 --> 15:22.620 a recurring thing to pay for this. Well, we'll go buy some features, but sorry, we don't have 15:22.620 --> 15:27.180 budget. Otherwise, my mind is, we'll be buying proprietary software for how expensive it could be. 15:27.180 --> 15:32.940 So the upstream gets paid to do features, which is good, you know, all of element acts for instance 15:32.940 --> 15:39.180 as pretty much funded by doing development like that. But then again, the features you do, 15:39.180 --> 15:43.260 might not actually create a mainstream product. They might not be particularly coherent. 15:43.340 --> 15:48.860 The funding is typically one-off. Say, hey, here's 500k to go and implement threads in whatever, 15:49.900 --> 15:56.060 however, then you've got threads, which are a maintenance nightmare, and no one providing cash 15:56.060 --> 15:59.660 to basically keep folks paid at least as their day drop to maintain it. 16:00.940 --> 16:08.860 Boo. So survival time, this is probably end of 2023 beginning in 2024. Matrix has taken 16:08.940 --> 16:14.380 this grunted. As a commons, hooray, we succeeded everybody. Open source creators across the 16:14.380 --> 16:21.420 matrix ecosystem, not just element, but basically most of the matrix vendors are suffering the 16:21.420 --> 16:25.820 same problems, because they release this open source, and there is no direct incentive for 16:25.820 --> 16:31.500 anybody to actually pay open source project, anything. So, started to get fairly drastic at the 16:31.500 --> 16:37.100 end of 2023. So trying to get elements available so that it can keep donating to matrix, 16:37.820 --> 16:41.820 ironically, elements stop donating most of the work to the foundation switch from a patchy to 16:41.820 --> 16:46.540 a GPL in order to sell a GPL exceptions to the people like the system integrators who want to 16:46.540 --> 16:52.380 customise it. R&D projects get shelved in favor of focusing on stability and quality. 16:53.020 --> 16:57.820 I'll give you probably a good thing for the ecosystem overall, but even then, three writing is 16:57.820 --> 17:02.860 still posing an existential threat to the whole ecosystem. Three writing is the technical 17:02.940 --> 17:08.140 economic term for this failure mode when people take the free thing and milk it for all its 17:08.140 --> 17:14.940 worth and don't maintain it. Interestingly, the press is starting to notice this. There's an article 17:14.940 --> 17:20.620 yesterday in the press about the German healthcare deployments of matrix for health insurance 17:20.620 --> 17:27.580 companies, where there's a recurring theme that folks are picking up open source implementations, 17:27.580 --> 17:34.940 whether it's servers from element or clients from other matrix vendors and going and using the 17:35.500 --> 17:41.340 including as a GPL without reaching any cashback. Now, these are large amounts of money to go and 17:41.340 --> 17:47.100 provide a service for tens of millions of citizens and users and so we're basically starting to 17:47.100 --> 17:53.340 talk into the press and saying like this is a bit of a scandal really because it's not sustainable 17:53.420 --> 17:57.740 if they're going to do these tenders and it then falls apart because there's nobody to write 17:57.740 --> 18:03.420 the software, it's not going to look very good. So the press is starting to pick it up, but there is hope 18:04.220 --> 18:09.500 there is a really simple answer. Basically, if you are buying a commercial matrix deployment 18:09.500 --> 18:15.580 you should mandate that the upstream project is funded by buying that upstream products if they have 18:15.580 --> 18:20.460 one. Now sometimes it's just going to be a little open source project, but these things are 18:20.460 --> 18:24.780 complicated or when they look as if they're successful even the one man bands go and say right 18:24.780 --> 18:30.700 I'm going to try to pay my rent using this and set up a small outfit in order to support the 18:30.700 --> 18:36.780 development costs and the role of firm the idea is that you as somebody wanting to run a commercial 18:36.780 --> 18:42.300 matrix deployment should just pay for what you use in order to keep the project free, free as in 18:42.300 --> 18:48.540 labor. So only buy from people who are actually going to fund upstream maintenance and that can 18:48.540 --> 18:53.260 be a system integrator. It can be anybody as long as you actually have a paper trail showing, 18:53.260 --> 18:57.740 but some of that cash might go back to the paying the rent of the people who are going working on it. 18:58.540 --> 19:03.340 So the idea is that you would normalize paying for open source products and that's as folks pay 19:03.340 --> 19:09.740 for baratry ones. I mean we were looking at the costs, I think one of the German ministries 19:09.820 --> 19:16.940 of how much they pay Adobe for like acrobat licenses every year or was it Microsoft for licenses 19:16.940 --> 19:22.460 too and we're literally talking towards the billions of euros a year and then they say oh yeah 19:22.460 --> 19:27.900 but we can't possibly we don't have the two million a year that would be needed to go and 19:27.900 --> 19:34.060 provide some support for the matrix stuff. It's really weird thing and we just want to normalize it 19:34.940 --> 19:40.300 so that it's not embarrassing to actually pay the people who write open source software and people 19:40.300 --> 19:46.540 might say oh but the upstream is trying to vendor lock you and it's evil element tool whoever 19:46.540 --> 19:51.980 going and trying to force people to buy their thing and you know there are other implementations out 19:51.980 --> 19:56.700 there if the upstream product isn't suitable then you switch and this would then be tree 19:56.700 --> 20:01.180 public money for public code and some people get it really right. I think the flagship here is 20:01.340 --> 20:06.700 the center of digital sovereignty in Germany where they're literally done this. I think they work 20:06.700 --> 20:16.380 with eight different vendors yeah brand of applause for vendors. Seriously thank you because it was 20:16.380 --> 20:21.020 a very very helpful for the vendors to say well we were with eight different soon folks to 20:21.020 --> 20:26.220 incorporate them into open desk we got element we got next cloud we've got open project we've got x 20:26.220 --> 20:31.660 wiki we've got open exchange but all of these and our grand rules are that we will actually just go 20:31.660 --> 20:38.780 and pay for their profession you supported product for stop and it was like oh my god I wish that 20:38.780 --> 20:45.100 everybody did that and Sweden again doing the right thing. US NATO doing the right thing UK kind of trying 20:45.100 --> 20:53.820 to do the right thing so some people definitely get it anyway where are we now? It's 2025 by the way 20:53.900 --> 20:58.860 I would say that matrix is closer to mainstream adoption than ever. We're obviously not there yeah 20:58.860 --> 21:05.100 no it's well 400,000 or 30 years as was the graph on the first page. The path has been winding 21:05.100 --> 21:09.740 like a lot of matrix vendors including element but others as well have ended up focusing on 21:09.740 --> 21:15.980 gov tech and health care to survive and as a result you end up with enterprise features getting prioritised 21:15.980 --> 21:21.660 over the mainstream consumer and community centric features. That said what's good for gov tech 21:21.740 --> 21:26.780 is often good for mainstream uptake to the competition here is basically what's Apple, 21:26.780 --> 21:33.740 Signal, Snack or Teams. Now on the matrix side the strategy has been the most important stuff now 21:33.740 --> 21:42.860 is to actually get everything fast beautiful or able to support beautiful apps and basically excellent 21:42.860 --> 21:49.180 not necessarily chase shiny mainstream features but make sure that the underpinnings are rock solid 21:49.180 --> 21:55.740 so the other folks whether that's people or elements or fluffy chairs or anybody have the best 21:55.740 --> 22:01.500 possible infrastructure and APIs to build whatever killer app that happens to be and so matrix 22:01.500 --> 22:06.700 coupon O is basically that long running project to try to ensure the matrix can support mainstream 22:06.700 --> 22:14.140 uptake so matrix coupon O first announced to two years ago to the day in foster them on the stage 22:15.100 --> 22:19.500 whole idea is that you make it work make it work right now and coupon O you actually make it not 22:19.500 --> 22:24.940 so you make it fast and useable you can outcompete the mainstream apps implementations to prove 22:24.940 --> 22:32.620 that the MSC is for matrix coupon O can work landed in September 2024 and they are next 22:32.620 --> 22:38.300 show north which gives you a log in and two-factor off and multi-factor off native IDC and also 22:38.300 --> 22:43.420 sort of amazing good stuff implemented all over the shop now there's a dendrite PR very kindly 22:43.420 --> 22:50.380 contributed by someone called midnight and LNX fractal I think landed it last Friday as well 22:50.380 --> 22:57.100 and the idea on this is that it will get turned on on matrix dot org basically as soon as the 22:57.100 --> 23:04.540 MSC passes FCP so that we drag everybody on matrix dot org instance kicking and screaming into 23:04.540 --> 23:10.940 the brave new IDC world feature and I will show you how cool and why that is a good thing in a few 23:11.900 --> 23:16.540 then we've got simplified sliding sink this is instant logging instant launch instant sting 23:17.420 --> 23:21.580 natively implemented and snaps nice will the con do we wait has landed it to 23:22.380 --> 23:29.020 or he says PR for it this is very close to being FCP on the MSC it has one edge case I believe 23:29.020 --> 23:36.620 remaining about handling deleted state and then matrix coupon O matrix RTC is a native 23:36.620 --> 23:40.860 encrypt to group by calling implemented an element cool web and element acts I believe that the 23:40.860 --> 23:46.060 famously stack also has matrix RTC support in this one we asked an incorporating feedback from the 23:46.060 --> 23:51.900 MSCs from the September launch and then finally invisible crypto some people don't understand 23:51.900 --> 23:56.220 invisible crypto at all and say oh this is crap that removing all of the warnings it will not be 23:56.220 --> 24:01.980 safe anymore this is the opposite of what indisible encryption is the whole idea is to make it more safe 24:01.980 --> 24:07.980 by removing the unactionable cryptic if you see what I did that warnings about message authenticity 24:07.980 --> 24:13.980 or verify devices and also add identity pending to this is a big change it will require users 24:13.980 --> 24:19.580 to verify their devices at log in but then we can just simplify the experience massively don't have 24:19.580 --> 24:24.940 this legacy concept of where devices hanging around which are probably an attacker because they 24:24.940 --> 24:30.940 didn't prove that they were owned by their actual owner so when we launch this at the matrix 24:30.940 --> 24:36.140 conference in September showed off a bunch of demos and everybody said well that's nice how do I 24:36.140 --> 24:42.060 use it and we said oh crap we kind of forgot to actually ship this in a community friendly 24:42.060 --> 24:49.100 distribution sorry about that good news though is that on the element side there will be open 24:49.100 --> 24:55.740 so I'll say GPL helmetshots and coming out soon and coming months in the meantime though I have a 24:55.740 --> 25:01.820 hacky work around which is going to be very interesting to demo without my lamp so how are we 25:01.820 --> 25:06.140 going to do this so if I can break in time it is not up and actually shell into my dangerous post 25:06.140 --> 25:11.340 it might interest anything can you hold mic for me because we don't do break in mic anymore 25:11.340 --> 25:18.780 burnt thank you right let's see if I can just shell in now it was literally what I was hoping 25:18.860 --> 25:23.340 would have been fixed let me see if I can switch back to my lamp's up 25:32.700 --> 25:36.700 I need to hurry up with this coming up 25:41.500 --> 25:45.820 I'm going to have fun here so the problem here is I need to figure out how to shell into my 25:45.820 --> 25:51.660 myself obviously so to come in so for a second I'm hoping that somebody might have done this but 25:51.660 --> 25:53.660 so I'm just doing a talk 26:08.860 --> 26:10.860 right almost there 26:21.660 --> 26:31.100 okay so the fact that I'm using the interest demos here is going to be from the chances of 26:31.100 --> 26:37.100 the screenshot are working and presumably now K-R-M-R we pray there we go right 26:39.500 --> 26:42.460 that is a single most stressful SSH restart I've ever done in my life 26:42.460 --> 26:55.500 come on please let me in no I should have messed around with SSH keys instead 27:01.020 --> 27:04.300 and that's of course that's just a wrong password I'm probably not having a password on it 27:05.020 --> 27:10.700 okay that's because I reset my password for me thanks guys 27:15.100 --> 27:21.500 well at least they tested it right okay so the interest demos is this location BPS which is completely 27:22.540 --> 27:29.580 naked and new I do have a test checkout here called element docket demo which I am going to 27:29.980 --> 27:46.060 go yesterday and mine post 27:46.060 --> 27:52.300 i've gone unchanged my password for me right and I am going to cut this 27:52.300 --> 28:02.300 something out of fresh. So this is a really basic Docker compose set up that I did just 28:02.300 --> 28:08.300 to go and be able to show off Matrix 2.0 stack for things like this. Do not use it in production. 28:08.300 --> 28:17.300 It is. God, but not using subversion anymore, are we? Come on, focus Matthew. 28:17.300 --> 28:25.300 The connectivity is giving me about two seconds of latency on typing here. Right. So this thing is a really, really, really simple repository. 28:25.300 --> 28:31.300 If I go and look at it, you can see here. You've got some scripts. You've got a template for some configuration. 28:31.300 --> 28:39.300 And if we look, what we have here, there's a set up script. So I'm going to run a set up script. 28:39.300 --> 28:47.300 And I'm going to say, I'm going to create a new server called foster.dentress.net. 28:47.300 --> 28:53.300 I'm going to not use MKSR. I'm going to use Let's Encrypts. I'm going to check it over to Matthew and Matrix.org. 28:53.300 --> 28:59.300 And we're going to set up Let's Encrypt quickly. And yes, it will be nice and that the FF spam me. 28:59.300 --> 29:05.300 And I'm pray that Let's Encrypt is not going to write limit my IP or something because that would be bad. 29:05.300 --> 29:15.300 Right. And it's gone and set up. That's some trips to get for me. Then I do Docker compose up and we're in. 29:15.300 --> 29:25.300 That's it. 29:25.300 --> 29:33.300 Let's see if it even works. Oh, yeah. Come on. So again, look in the background. There's synapse. 29:33.300 --> 29:37.300 Desperately flying around the place. And here's an element in. 29:37.300 --> 29:43.300 Now, wait a little enough out of the box. It doesn't have. 29:43.300 --> 29:51.300 I can't registration turned on. So I guess I should have set this running in the screen session, shouldn't I? 29:51.300 --> 29:57.300 Let me control see on that. And perhaps set that running in a screen session instead. 29:57.300 --> 30:03.300 Docker compose up. Right. So in the background. And I go and look at the read me and it gives you a little. 30:03.300 --> 30:11.300 Come on, blind to run, which is to register a user. So this is using mass. 30:11.300 --> 30:19.300 This is the Matrix 2.0 next run all stuff. And if I run that, it will say, hello. Do you want to go and create a new user? 30:19.300 --> 30:25.300 I'll create one called Matthew. And I will give it a email address. 30:25.300 --> 30:33.300 And I will give it a password. 30:33.300 --> 30:41.300 Right. And that's it. Create the user. Confirm? Yes. Brilliant. So I should now be able to go over to 30:41.300 --> 30:51.300 and sign in. Hit continue on this. And it takes me to mass. So this is now the authentication system provided not by the client by the server. 30:51.300 --> 30:59.300 So I can help them to the username here. And trying to remember the password I just said. 30:59.300 --> 31:05.300 And now this is quite a different setup. In the, we're giving this over. 31:05.300 --> 31:13.300 Well, this client access to my account. Hit continue on that. Make sure that you trust it. 31:13.300 --> 31:17.300 And hopefully, let's go back and see what's happening in my screen session. 31:17.300 --> 31:27.300 Lots of turning around. Come on. There we go. Woo. So now enter this brand new server. 31:27.300 --> 31:35.300 So I will disable notifications. And I will go, and then publish, I'll DM you. 31:35.300 --> 31:41.300 Okay, right, fine. So I'll go and put an md on matrix. All here and see if Federation is going to work. 31:41.300 --> 31:47.300 Oh, there she is. Go. Hello. 31:47.300 --> 31:55.300 Goes in creates room. Set up recovery. Probably good idea for us to do recovery stuff. 31:55.300 --> 32:07.300 Yeah. That was, yeah. So what happens if you type on someone else's laptop. 32:07.300 --> 32:15.300 All right, let's get back up successful, brilliant. So do you read me? 32:15.300 --> 32:21.300 Well, whilst I'm indeed logs in, thank God for that. 32:21.300 --> 32:23.300 Hey, Federation. 32:23.300 --> 32:31.300 So let's go look at the next thing, which is the matrix RTC, the native calling. 32:31.300 --> 32:37.300 So we do still have old one to one calling, but in practice everything should be using element call these days. 32:37.300 --> 32:41.300 We need to migrate from legacy to element call and the matrix RTC stuff. 32:41.300 --> 32:49.300 And allow this time, and I'm going to hide my thing there. Then what this is done is to basically use a 32:49.300 --> 32:55.300 Metroscopy widget to embed the element call web app into here and start a video call. 32:55.300 --> 33:01.300 And Emily in there is just logging in on her phone there and hopefully going to turn up. 33:01.300 --> 33:07.300 And even though she's on a matrix.org server, we have Federation stuff happening there. 33:07.300 --> 33:15.300 Now you don't have Vicerin stores. I'm not going to be up to screen show my phone nicely. 33:15.300 --> 33:21.300 So I have to use the analog gap for this bit, which is a shame because it's in some ways the best bit, but let's use. 33:21.300 --> 33:28.300 So now let's see one of the benefits of matrix authentication service, which is I can just go here to link new device. 33:28.300 --> 33:32.300 Give me a QR code scan, and I'm going to take. 33:32.300 --> 33:40.300 So can you point this or, oh, you know, this is really not going to be able to show me. 33:40.300 --> 33:46.300 I'm going to just go and scan that. 33:46.300 --> 33:54.300 So obviously it's like, oh, so my problem that my phone doesn't have internet connectivity. 33:54.300 --> 33:58.300 It tends that you need internet connectivity for this to work. 33:58.300 --> 34:04.300 Right. So now it's actually done it. It gives you a check code to make sure that it was me who scanned that. 34:04.300 --> 34:12.300 I've then gave the access to my account here and can you point this. 34:12.300 --> 34:13.300 Yeah. 34:13.300 --> 34:14.300 Yeah. 34:14.300 --> 34:15.300 Okay. Oh, sure. Yeah. So you are. 34:15.300 --> 34:17.300 Right. So let's get rid of this. 34:17.300 --> 34:20.300 So you can now see there is a new login on the IP address there. 34:20.300 --> 34:24.300 And all analog gaps sharing. This is what has happened on the iOS. 34:24.300 --> 34:28.300 If I hit OK, then go into here to the chat with Amazon's in. 34:28.300 --> 34:36.300 You can literally see that conversation with its end-to-end encrypted history has been synced in like that onto iOS. 34:36.300 --> 34:41.300 So it's certainly didn't say it happened in real time, but it happened basically instantly once I gave it connectivity. 34:41.300 --> 34:46.300 That I've now logged in on my phone on this brand new server on this brand new account using a QR code. 34:46.300 --> 34:53.300 And if I open up the chat view here, you can see the, oh, in fact, really obviously I could do a bit to just join the call. 34:53.300 --> 34:56.300 I'm just joined the call on this new login here. 34:56.300 --> 35:01.300 So there I am going to unmute the video on that too. 35:01.300 --> 35:04.300 And it's actually working. 35:04.300 --> 35:13.300 So, sorry. Thank you. 35:13.300 --> 35:19.300 And what this is showing here is also, we've got multiple device support now in matrix RTC. 35:19.300 --> 35:24.300 This started off as a one-to-one call, but then we just casually added a new device. 35:24.300 --> 35:29.300 Literally a new device in the middle of the call and just go and put a spate on it too. 35:29.300 --> 35:32.300 If we had more time, I'd now do a screen share and take you through all the things. 35:32.300 --> 35:35.300 But you get the idea that it's rather cool. 35:35.300 --> 35:38.300 This, I would say, is starting to fill mainstream. 35:38.300 --> 35:43.300 We've got discord style QR login that just works with all of the encryption state. 35:43.300 --> 35:45.300 I didn't need to specify my home server. 35:45.300 --> 35:47.300 I didn't need to specify my recovery key. 35:47.300 --> 35:49.300 I didn't need to even remember my username or my password. 35:49.300 --> 35:53.300 I just split the link and it took me in there with the same identity. 35:53.300 --> 35:55.300 Verified out of the box. 35:55.300 --> 35:57.300 Woo! 36:03.300 --> 36:07.300 I'm so running out of time, but let's keep going for it. 36:07.300 --> 36:09.300 Then other things in 2025. 36:09.300 --> 36:13.300 No, September 24 we did the first ever global matrix conference. 36:13.300 --> 36:15.300 It was really cool. 36:15.300 --> 36:17.300 I highly recommend going to the next one. 36:17.300 --> 36:21.300 Hopefully in front next year of this year in September. 36:21.300 --> 36:25.300 But one of the things that truly struck me is that this is not. 36:25.300 --> 36:27.300 This is a properly heterogeneous ecosystem. 36:27.300 --> 36:29.300 You've got, obviously, the native rust stack. 36:29.300 --> 36:31.300 Matrix rust SDK that is powering the element stack. 36:31.300 --> 36:36.300 But now you've got a totally separate stack without any lines of code overlap at all. 36:36.300 --> 36:39.300 Over then I guess the crypto and the form of flutter. 36:39.300 --> 36:41.300 It requires pot and multi-platform of tricks. 36:41.300 --> 36:43.300 Like wise go, which I think. 36:43.300 --> 36:45.300 I'll be using go, oh yeah. 36:45.300 --> 36:47.300 Yeah, too. 36:47.300 --> 36:49.300 So there's no code overlap at all. 36:49.300 --> 36:51.300 I see Brad say maybe. 36:51.300 --> 36:53.300 Yeah. 36:53.300 --> 36:55.300 And then react native. 36:55.300 --> 36:57.300 So this is a new one coming from, you know, 36:57.300 --> 36:59.300 made an fulfillment. 36:59.300 --> 37:01.300 You've taken rust SDK and put a really good react native binding over the top. 37:01.300 --> 37:03.300 I was playing it with it the other day. 37:03.300 --> 37:07.300 It's literally about 10 lines of react native to get your matrix stuff up and running. 37:07.300 --> 37:11.300 There's plus times like matrix client from echo and many, many more. 37:11.300 --> 37:15.300 So when people say matrix is controlled by element and it is not a tree ecosystem, 37:15.300 --> 37:16.300 I say bullshit. 37:16.300 --> 37:19.300 Look at this list of folks who have implemented 37:19.300 --> 37:21.300 entire separate stacks independently. 37:21.300 --> 37:23.300 And the talks were amazing. 37:23.300 --> 37:25.300 Not quite as good as fostering, but still amazing. 37:25.300 --> 37:29.300 Around about 50 years and you can see all the thumbnails there. 37:29.300 --> 37:30.300 I'm relatively small. 37:30.300 --> 37:32.300 How many people were there by the way? 37:32.300 --> 37:33.300 Okay. 37:33.300 --> 37:35.300 So you should all come next to you. 37:35.300 --> 37:36.300 We'll be great. 37:36.300 --> 37:40.300 And some of the highlights we heard from TI Messenger in Germany, 37:40.300 --> 37:43.300 which is running our matrix for all of healthcare. 37:43.300 --> 37:46.300 We heard from Open Desk and the Zendes folks I mentioned earlier, 37:46.300 --> 37:52.300 who are using us as the sort of chat sub system for this team's killer initiative. 37:52.300 --> 37:54.300 You've got E post in Switzerland. 37:54.300 --> 38:01.300 No, this is really exciting where they are pivoting the Swiss postal service from snail mail to Matrix. 38:01.300 --> 38:02.300 Obviously. 38:02.300 --> 38:03.300 Why wouldn't you? 38:04.300 --> 38:07.300 And that may sound crazy, but it's literally what they're up to. 38:07.300 --> 38:12.300 Going and deploying matrix and servers on premise for big corporate customers, 38:12.300 --> 38:14.300 providing a SaaS hosting to themselves, 38:14.300 --> 38:17.300 and then building a great big super app on top, 38:17.300 --> 38:20.300 where all of the businesses have their own identity is a bit like, 38:20.300 --> 38:22.300 we chat or something like that, 38:22.300 --> 38:27.300 but also communicating out to the wider network. 38:27.300 --> 38:29.300 You've got chat in France. 38:29.300 --> 38:31.300 Did you lead to the great talk? 38:32.300 --> 38:37.300 Yeah, it must be the biggest public sector deployment with about 300,000 monthly active users, 38:37.300 --> 38:40.300 millions of messages sent to the moment for each ministry, 38:40.300 --> 38:41.300 getting its own home server. 38:41.300 --> 38:45.300 The stance for this, by the way, a public starts to incubate, 38:45.300 --> 38:47.300 to beat it up to FRL or something. 38:47.300 --> 38:49.300 If you want to go and look at their database, 38:49.300 --> 38:51.300 dashboard and see how many people are logged in right now. 38:51.300 --> 38:54.300 No, for really interesting one was NATO, 38:54.300 --> 38:57.300 using a fork of element called nice for low-side communication 38:58.300 --> 39:00.300 to replace what's happening signal. 39:00.300 --> 39:04.300 So, that's where things are today, what comes next. 39:04.300 --> 39:07.300 Well, one of the big initiatives, which I don't have time to talk about, 39:07.300 --> 39:09.300 is state res V3. 39:09.300 --> 39:15.300 This is an overhaul of state resolution that we're in relatively late stages doing. 39:15.300 --> 39:20.300 This involves TARDIS, which stands for time-agnostic room-dagon inspection service, 39:20.300 --> 39:24.300 which lets you visualize precisely, I'm glad somebody likes my joke. 39:24.300 --> 39:27.300 This is one of my projects with civilisation, 39:27.300 --> 39:31.300 so you can see what's happening in your rooms and debug them. 39:31.300 --> 39:34.300 And then an ounce today, in about half an hour, 39:34.300 --> 39:36.300 will be chaos, actually one hour. 39:36.300 --> 39:39.300 Chaos is a new tool, which is basically a chaos monkey, 39:39.300 --> 39:41.300 petition testing tool for home servers. 39:41.300 --> 39:43.300 It causes networks, breaks network, 39:43.300 --> 39:45.300 in a deterministic fashion, and then checks, 39:45.300 --> 39:46.300 where things converge or not. 39:46.300 --> 39:49.300 You can see why this might be useful when debugging 39:49.300 --> 39:50.300 state res issues. 39:50.300 --> 39:54.300 And you can see this, and you can see us talk at two in the devrim. 39:54.300 --> 39:59.300 MLS, if I had another three hours, I would talk about MLS and great detail. 39:59.300 --> 40:04.300 Lots of work on the ITF and to end encrypted and standard over the years. 40:04.300 --> 40:06.300 The main one on our side, and the call team, 40:06.300 --> 40:09.300 has been decentralized MLS, but it's a bit exotic, 40:09.300 --> 40:13.300 it deviates from how RFC MLS is meant to work. 40:13.300 --> 40:18.300 We now have two MSEs out there, one from Travis and myself, 40:18.300 --> 40:21.300 one from BWY, proposing slightly different ways, 40:21.300 --> 40:23.300 which you could add MLS to matrix. 40:23.300 --> 40:25.300 We also have things like linearized matrix, 40:25.300 --> 40:27.300 and Mimi, which is pretty similar. 40:27.300 --> 40:31.300 Progress on MLS is frankly blocked on funding right now. 40:31.300 --> 40:34.300 It is a very active topic of research. 40:34.300 --> 40:36.300 Also, post-content, last fall student, 40:36.300 --> 40:39.300 he teased and pulled out some PQXDH implementation 40:39.300 --> 40:40.300 for for those maps. 40:40.300 --> 40:42.300 And we assume that somebody would leap out to the audience 40:42.300 --> 40:45.300 and say, oh my god, post-content matrix, I'll fund that. 40:45.300 --> 40:46.300 And they didn't. 40:46.300 --> 40:47.300 So here we are. 40:47.300 --> 40:50.300 Yeah, later, and absolutely nothing has happened on it. 40:50.300 --> 40:52.300 So if anybody wants post-content matrix, 40:52.300 --> 40:54.300 please can you put some money behind the bar. 40:54.300 --> 40:57.300 Likewise, MLS is experimenting with PQS, 40:57.300 --> 41:00.300 as well, so we might get PQ3 with MLS, 41:00.300 --> 41:03.300 maybe, but again, for progress, blocks on funding. 41:03.300 --> 41:06.300 Then, Rust SDK has been evolving first. 41:06.300 --> 41:08.300 Loads of fun stuff. 41:08.300 --> 41:12.300 I'll talk about probably in the states of the union in five minutes. 41:12.300 --> 41:14.300 But big stuff is offline support, 41:14.300 --> 41:17.300 persistent event cashes, threads coming pretty soon. 41:17.300 --> 41:19.300 Space is coming a bit later, 41:19.300 --> 41:22.300 also we're seeing these additional bindings I mentioned. 41:22.300 --> 41:24.300 We're considering there may be a world 41:24.300 --> 41:26.300 without matrix JSSDK, 41:26.300 --> 41:28.300 with Wazon bindings on the horizon, 41:28.300 --> 41:29.300 thanks to Mozilla, 41:29.300 --> 41:30.300 and filament, 41:30.300 --> 41:32.300 who are implementing Uniafify, 41:32.300 --> 41:34.300 or Unify support for Wazon, 41:34.300 --> 41:36.300 so that we can take the existing Rust SDK bindings, 41:36.300 --> 41:38.300 the support iOS and Android, 41:38.300 --> 41:40.300 and just target them for the web. 41:40.300 --> 41:42.300 Very quick, final demo. 41:42.300 --> 41:45.300 So I'm not going to be able to show you on MLS laptop, 41:45.300 --> 41:46.300 which sucks, 41:46.300 --> 41:49.300 but come to the matrix every human CT mode, 41:49.300 --> 41:51.300 because there's actually his project, 41:51.300 --> 41:53.300 where his taken dioxys, 41:53.300 --> 41:54.300 as a quick hack, 41:54.300 --> 41:56.300 to take Rust SDK and put it in a browser, 41:56.300 --> 41:58.300 and I got on the skin that looked like element, 41:58.300 --> 42:00.300 and it would have been a really cool demo 42:00.300 --> 42:02.300 to see element X web, 42:02.300 --> 42:06.300 as a possible approach using dioxys. 42:06.300 --> 42:08.300 Lots of mainstream features, 42:08.300 --> 42:09.300 messing, 42:10.300 --> 42:11.300 both on the client, 42:11.300 --> 42:12.300 and the server, 42:12.300 --> 42:14.300 lots of trust and safety work to be done. 42:14.300 --> 42:16.300 We are expanding the team there, 42:16.300 --> 42:18.300 and making some progress. 42:18.300 --> 42:19.300 And then finally, 42:19.300 --> 42:21.300 perhaps matrix 3.0 in future. 42:21.300 --> 42:22.300 We don't know what this will be. 42:22.300 --> 42:23.300 It could be all about CNS. 42:23.300 --> 42:24.300 It could be MLS. 42:24.300 --> 42:25.300 It could be converging with Mimi. 42:25.300 --> 42:27.300 It could be peer matrix. 42:27.300 --> 42:28.300 Place your books. 42:28.300 --> 42:30.300 But the foundation is key to this. 42:30.300 --> 42:33.300 The foundation is the future of matrix, 42:33.300 --> 42:35.300 in terms of protecting it from the commercial 42:35.300 --> 42:38.300 of the institutions of vendors like element, 42:38.300 --> 42:40.300 or anyone else. 42:40.300 --> 42:42.300 And the next 10 years of matrix 42:42.300 --> 42:44.300 will be nurtured by the foundation, 42:44.300 --> 42:45.300 if it has funding. 42:45.300 --> 42:46.300 Right now, 42:46.300 --> 42:48.300 we have a bunch of funding members. 42:48.300 --> 42:50.300 We'd very much appreciate them. 42:50.300 --> 42:52.300 Thank you for putting money behind the bar 42:52.300 --> 42:53.300 to keep the foundation alive. 42:53.300 --> 42:55.300 We need more. 42:55.300 --> 42:58.300 We elected the governing board in May. 42:58.300 --> 43:00.300 There are now 24 people, 43:00.300 --> 43:01.300 I think, 43:01.300 --> 43:04.300 looking after the overall direction of the foundation. 43:04.300 --> 43:06.300 If you're not a member already, 43:06.300 --> 43:08.300 or even better bully your employer to join, 43:08.300 --> 43:10.300 if they like matrix and use matrix. 43:10.300 --> 43:12.300 I'm looking at you, 43:12.300 --> 43:13.300 big tech companies, 43:13.300 --> 43:14.300 use matrix, 43:14.300 --> 43:15.300 perhaps a bank communication, 43:15.300 --> 43:17.300 but I'm members of the foundation. 43:17.300 --> 43:18.300 Please. 43:18.300 --> 43:19.300 Then governing board, 43:19.300 --> 43:20.300 24 people, 43:20.300 --> 43:21.300 11 orgs, 43:21.300 --> 43:22.300 come, 43:22.300 --> 43:24.300 listen to Greg in the dev room, 43:24.300 --> 43:25.300 say what's going on there. 43:25.300 --> 43:26.300 And we need help. 43:26.300 --> 43:27.300 Please, 43:27.300 --> 43:28.300 support matrix, 43:28.300 --> 43:29.300 commercially, 43:29.300 --> 43:30.300 probably from upstream vendor. 43:30.300 --> 43:31.300 Doesn't have to be element. 43:31.300 --> 43:33.300 Can be any upstream vendor, 43:33.300 --> 43:34.300 build on matrix, 43:34.300 --> 43:35.300 follow us on, 43:35.300 --> 43:36.300 ask it on, 43:36.300 --> 43:37.300 please go and spread the word. 43:37.300 --> 43:38.300 Thank you very much. 43:38.300 --> 43:39.300 Thank you very much. 43:52.300 --> 43:53.300 Probably don't have time for questions. 43:53.300 --> 43:55.300 Come to the dev room if you want to ask questions.