Matrix.org
Tim: [00:00:00] I'm Tim Panton.
Vim: [00:00:03] And I'm Vimla Appadoo
Tim: [00:00:06] And this is the distributed future podcast. This is a podcast where we interview people who are doing interesting things and who may give us a hint into what the future might hold, based on what they're doing and how they're doing it. So this one is actually, more, almost more about kind of. Process than it is about technology. it's with, with Amandine , who's, the COO of matrix.org and a couple of associated companies, and it's about how they've structured that to try and protect their goals from the money. So there's sort of, it's deliberate structure of trying to separate where the venture capital investment has gone in from the sort of philosophical heart of the project. And it's interesting to see how they've constructed that and how early on that start thinking about that for it to work. So it's quite exciting, kind of, you know, although oddly process-driven conversation.
Vim: [00:01:10] Yeah, I can imagine. I think, as I started to explore more into different ways of bringing purpose and value into organizations, that recognition that it takes the whole system to understand it, which means processes, systems, services, as well as regular patients law with governance and everything in between.
Tim: [00:01:35] Right, exactly. And I think that that came over for me was, was how much they really had to, like put that in early and think it through and how much effort and time they, they had to put in in order to preserve the, the goals that they had from the start of the project. And I mean, it's kind of, they've had to move around in order to do it. But, and you know, adapt their strategy somewhat, but like, Really, interesting to see, see how that's, how that's worked out. And I think if you don't go in with the right goals and the right mindset, then you've got no chance in that. I think.
Vim: [00:02:15] Oh, definitely. And I think one of the biggest risks with, kind of parts of companies is the assumptions. I would always be good and it will just be based on the kind of, the foundations of that. Whereas if it's not written into everything, if it's not a kind of conscious consideration and how you operate and fundamentally run the organization, it will slip away. It will kind of, it will become secondary
Tim: [00:02:44] yeah. Yeah.
I mean, I had a little bit of an incident with that this week. The startup I've used for something who changed ownership and change policies and change character over the years, and now I'm less comfortable, significantly less comfortable with them than I was than I am three or four years ago. Because they've taken a bunch of venture capital in the, they've shifted, they've pivoted. To be fair, they pivoted away from what it was that I wanted them to do and what it is that they're now doing. Cause that's presumably where the money is. so it's kind of tricky to see how as a consumer, you can protect yourself around that
Vim: [00:03:25] What was kind of, the biggest learning from the conversation?
Tim: [00:03:29] Oh, I think for me was that, the legal structure actually turns out to be the thing that matters most, and that you can, you can have all the aspirations you like and all the people you like. But in the end, it comes down to. You know, what is the legal structure? Where does the money set sit? Who owns what? who has the right to push who around that, you know, that structure is the thing. I know obviously in the future that could change, but I guess then, then you get down to kind of notification and telling the customers what's happening, keeping them informed. well, what they've, I mean, what they've done is, is they're using the open source nature of it to mean that even if that happens, then people can just move off to another provider. and, and, The enabler there is the fact that they kept the protocol open so anyone can implement it. So in principle, like if you know all of the existing providers of that protocol suddenly decide to sell your data to China or whatever, then somebody else can spring up and build a replica based on the opensource. and that keeps everybody, at least to a larger extent, honest. The trouble then is how do you extract money from it? How do you get investment into it? so it's kind of two sided coin, but really, really exciting to kind of.
Vim: [00:04:58] Yeah, I'm going to save a cheeky pocket. But they, the paper that I was involved with, with a British Academy called the future of the corporation was really focused on that kind of like legality and governance structure around how you, how you embed it and how you . How do you make that clarity and transparency? But what I found really interesting was this was the guy through the discussions around the focus on, employee employees and like employee buy in and employee rights you right? And how you, if you're hiring someone in based on the premise of a purpose or, you know, whatever that structure is. How you then hold the business to account to any shifts or changes based on investment or stakeholders or you know, all of that kind of stuff. How you get that, that conversation going in that relationship built.
Tim: [00:05:54] Yeah. I
mean, I think, I think that's really important, but I think it's also important that you. You go wider than just the employees and you have to get the customers, you have to play fair with the customers as well, and changing the rules after they've been like customers for four years and you've got their data, is quite difficult. Like, you know, I mean, I understand the commercial realities of it, but I also understand that that doesn't make me happy as a customer. And I start looking for an alternative now. Yeah.
Vim: [00:06:30] I think that's the rare case though. I think the majority of people will be unhappy, but continue to use it because that isn't that instinct to find new alternatives or you know, those that, that old saying that you're more likely to buy a new house than to change your bank. You just, you just get used to the same, it's more hassle to try and shift it rather than be completely happy with the service.
Tim: [00:06:57] Right. And the thing with the bank though, is that they traditionally less so now, but they traditionally made it fantastically difficult to move. and you know, everyone still has the perception that moving bank banks is, is difficult and painful, and therefore, like something you'd avoid if you possibly could.
Vim: [00:07:18] Yeah, definitely.
Tim: [00:07:20] And, and, and so there's a. In the case of the banks, there's this whole issue about regulation, making it easier and more reliable. I mean like on a regulatory basis to move banks in order to create competition where historically there hasn't really been any.
Vim: [00:07:38] Yeah. But then it comes down to that, can shame the mindset as well. Even if we did regulate it to make it as easy as possible. How long would it take to shift us as consumers because of. The baggage we see alongside that. And you know, I think you see the same in government services as well, where the, it doesn't matter how easy or simple it is to do things online. I still though, well, we're going to have to print off a form somewhere or it's not going to work anyway, so might as well go and do it in person
Amandine: [00:08:10] yeah.
Tim: [00:08:11] Yeah. I mean the go and do it in part. I had one of those this morning. I was sorely tempted. I in the end did, I didn't, but I was solely attempted just to rock up at somebody's office this morning cause I felt fairly sure that me having me sitting grumpily in reception, would sort it out more effectively than making a phone call. And I'm going to read it. Yeah. I'm going to reserve that for like later in the week. when they, when they, right. When they haven't rung me back, you know, kind of rock up looking, looking like, Like, I'm messing up their office, so untidilly sit in reception until it solves itself. But I mean, hopefully that doesn't turn out to be necessary, but
Vim: [00:08:55] That's the thing though like it's not out of the kind of remit of reality that that is the case. That is what is going to happen.
Tim: [00:09:07] yeah. Yeah. I mean, although the nice thing about that is that like. I know I'm doing when I do that, it's pushing their costs up that what that does is it gives them a disincentive for handling me badly on the phone or because, cause it costs them, like I saw some figures from the banks that it cost them or hundred times more to process a transaction in branch than it does online.
So every, every failed transaction online costs them a hundred times more. If, if the person rocks up in a branch.
Vim: [00:09:46] Yeah.
Tim: [00:09:47] So like there's a big incentive for getting on online actually to do what it is that, you know, I as a consumer want it to do. I'm not sure that's totally got into people's heads yet that, that massive drive
Vim: [00:10:01] no it hasn't, especially when you see a job risk as well. So my job in the branch is going to be made redundant by a really good online experience.
Tim: [00:10:13] Yeah, I mean, that is the other side of it, of course, that like, you know, the reason it costs a hundred times more to have people in, in branch is because it's people. rather than, than automation. I must tell you a funny story about automation, which is like unrelated to this, topic of this podcast, but it happened to me this week and it's sort of relevant to the other podcasts topics we've done about, about bias. so I was doing a, I did a, a robo car race. So the challenge is you build a little robot and you train it to drive it around this, track the tracks about, I don't know, three meters by five or something. And basically you're your training a vision processing. AI to follow the dotted line down the middle of the road and stay between the two solid lines. Like, you know, that's the sort of the whole game is to do this. And, so I did this and the thing kind of mostly worked and then it went completely berserk at one point. And I had no idea, sort of started lurching towards me and I'm like, I couldn't work out what was happening. So I ended up capturing the video, and that's one of the things, just like, it's not obvious what it's reasoning is about it suddenly doing something random. So I captured the video and I realized it took me a little while looking at the video, but w what is it done? Is it seen the white lines around the edge of my converse and it was heading for my shoes.
Vim: [00:11:37] So, right, right.
Tim: [00:11:39] So the, the, the meta lesson there is not to rock up, not to rock up to a robo cars competition wearing converse cause they look like the middle of the road, but like, but, but that's also like. You know, there's a, there's a, there might be a lesson there about like not wearing converse in public. If self-driving driving cars takeover, it's like we might end up being a target, but it was interesting and kind of revelation for me actually how hard it was to work out. Why, why it was doing what it was doing.
and I'm trying to think who we did. We've talked to a couple of people back in the early episodes around around bias and, and you know, misconfigured AI. So, so there's a little little shout back to, our past and no doubt we'll come back to it as well cause I think it's, it's going to remain a hot topic. One of the, one of the thing I wanted to say was like an UN kind of to the audience is, if you're enjoying what you're hearing. Please like, like it or put it on Twitter or kind of spread the word, just word of mouth, your friends, whatever. Because the more people we get listening, the happier everyone is, hopefully. And obviously, you know, send us, send us commentary by, by Twitter or, or. You know, any other means you find to get hold of us. There are links on the distributed future website that, that will help you find us. If you, if you've got comments cause we, we would appreciate kind of feedback about what we're doing, right, what we're doing wrong, what you like, what you don't like. we'd, we'd obviously also do that in the, in your podcast app cause that's also really informative. So anyway, I'll, I'll introduce let Amandine introduce herself then I think .
Amandine: [00:13:29] I am Amandine Le Pape and I am the cofounder of the matrix.org projects, which is an open standard for secure and decentralized communications. So the idea is to provide a layer on top of the web where everyone can have. Real time communication and where we can actually break the silos of the big communication apps that everyone used. Today we're trying to democratize, instant messaging and voice over IP and any type of real time communication. I'm, so I'm one of the director of the matrix.org foundation. There are five of them of, of us. And I'm also the co founder and COO of new vector, who is a for profit company, building apps and services on top of the matrix of a matrix itself.
Tim: [00:14:23] Okay. Cool. So, so I guess maybe track a little bit back and try and work out why this matters. So you're, I mean, this is a communication system for businesses or for. People, individuals.
It's for anyone really. today. we can. so basically we've, we've built an app on top of matrix with defined the protocol and the technology itself. And we're, in order for it to grow, you need users. If you want to grow a network, you need users. So we've also built a flagship application on top, which is called riot. And, today it looks very similar to Slack. is, used mostly for collaboration purposes, in enterprises and, between communities like open source communities and these kind of things. But the, in the end, the goal is to make it available to everyone, to everyone. And we also want to address the more family, friends, and family use case for individuals. as well as, as well. Then as well as the collaboration angle.
So kind of differentiator from, I mean, I suppose it sits somewhere between kind of the three things. It's like. If you're aiming for the ubiquity of SMS, you want the kind of behavior or functionality to some extent of of Slack, but you're aiming also for the privacy of something like Signal. So there's a sort of, you're trying to do all three things at once. Is that fair?
Amandine: [00:15:55] Yes. I think that's quite a good, a, a good, explanation. So definitely the ubiquity of SMS and even more email, the security of, signal and WhatsApp, any end to an encrypted communication. And in terms of feature, it can be Slack, right? It can be WhatsApp as well. So just be able to exchange messaging, file a voice calling. So the different, I think the, the missing bit is to highlight how close it is to email in the, in the fact that. Everyone can choose who is hosting the data. So the same way you create your email address on a service you like or you trust, it can be email, it can be your company's email, it can be Hotmail, whatever. We want to be able to do the same for instant messaging. And if you want to, if you don't trust a service, you don't register an account there, but you can still communicate with someone using this service.
Tim: [00:16:51] Okay. So those, so that's a true federated protocol, which at least in principle, anyone who's got kind of the software can join.
Amandine: [00:17:02] Yes, like email, if you have an email address, you can send an email to anyone and the idea is to do the same for instant messaging
Tim: [00:17:11] in practice though, email now running your own email server is really hard. I mean, I used to and now I don't because it's just, it's got too difficult about five or six years ago. It just got too hard to do as a, as a kind of part time task. what's your aspiration there? Like. Do you think that lots of people will be running their own matrix services, or do you think that it's going to be, you know, there are seven or eight around the world.
Amandine: [00:17:40] it should be more than seven or eight. We're not expecting anyone to be able to run their own matrix server unless we set up. we've seen people like a freedom box in, in France. It's an opensource project allowing you to have your small box and a in your house where you can have your email, your matrix server your file, storage, everything open source. If, if it's, if I managed to find a way to package it like this or partner with someone packaging it like this so that anyone can just turn on a box at home and get their matrix server, that'd be great. Otherwise, the idea is more like, like yourself, like you were running your email server. people who are a bit technical, should be able to do it, but we don't expect anyone to do it. But we also don't expect to have only five matrix servers in the world. More, like probably. Hundreds of thousands or a few millions, would, would make sense like companies or institutions and, these kind of things.
Tim: [00:18:40] Okay. So, so then you really do hit up against the, the governance issue of like, you know, how does this get managed? How did those things cooperate? How do they all know they're speaking the same protocol? Like, you know, how do you know that one email address or one. Matrix address isn't being used by somebody else. Like how do you start to enforce those things in, in a highly federated environment?
Amandine: [00:19:08] Well, everything is baked into the protocol itself. so that you get, yes, the unique, unique identifiers to be able to connect, to one another. Basically.
Tim: [00:19:20] So this is a trust, the math situation. You're not, let, you've got a kind of cryptographic, structure, which enforces some of the rules, at least.
Amandine: [00:19:31] in terms of unicity of the, of the identifiers, yes. It's, it has to be each on the server. you can not create multiple, similar address and your address is linked to your server.
Tim: [00:19:44] Okay. No, I'm just thinking about how, you know, you spam emails that claim to be from people who they aren't from. So that's that classic Federation problem. trusting the, the origin of things is not that easy, but so, so in terms of governance, how you, you mentioned you are on the board of a number of different structures. Can you kind of talk through how those fit together and what influence those have on the project?
Amandine: [00:20:12] Sure. So the protocol itself is managed by the nonprofits matrix or foundation. the idea is that we're building an ecosystem here. We want, as many people as possible building on top of matrix using matrix and developing matrix apps and services and making money from matrix. So the goal of the foundation is really to grow this ecosystem. So it needs to be a completely independent entity. And it's holds, the, the, the standard itself. So th the specification documents of the standard, which are open of course, and, on open as well, some of the open source implementations of the servers or services on these kinds of things. So there are five people on the board. the two cofounders, Matthew Hodgson and myself. And we have three other independent people, who are, Jon Crowcroft who is a professor of communication at Cambridge. Ross Schulman who is, tech lawyer in the U S and, Dr. Jutta Steiner, who is the CEO of, a company called Parity who's in using, Matrix and a riot for their communication since almost the start. so it's meant to be an independent entity, who is managing the open standard itself. And then on the side, there is a for profit company called NewVector, who is a startup of 30 people. it hires most of the matrix core teams. So the team who created. The protocol who is, like 90% of the work new vector does, is actually contributing to the matrix open, open standard and open source implementations. And, but it's very much, so it's, we've, we've raised, we've raised some funds from VCs, a month ago, two months ago now, and we actually have customers and so professional services. And, and hosting services as well. So the idea is on one side, we build, paid services and applications to actually grow the ecosystem. And on the other side, we have this nonprofit entity who is meant to be independent from NewVector and hence the five directors, on it. And, who is taking care of the ecosystem itself.
Tim: [00:22:31] Right. So then how do you, how do you fund those, the, the, the foundation, how does that, the foundation, and then I guess it probably doesn't have huge costs, but must have some costs. How's that funded?
Amandine: [00:22:46] So the foundation, it doesn't have many courses, as you say, because the only things would be, the, so there is some marketing things when you want to attend conferences and these kind of things. otherwise. It's, it doesn't have the only income. It has are donations from people who wants to support the project. And then, it actually, so we have a service agreement between the foundation and new vector, other, other companies, I think, so that's new vector can contribute to the foundation. So. There are not, there isn't much money going through the foundation. It's contracting new vector to do some developments and new vector provide developments for the foundation for free.
Tim: [00:23:29] Right, right. And how do you, I mean, I guess the, the other kind of side of this is where is the intellectual property? I'm sitting who owns it and who, who has to defend it? I mean, if you get a Peyton. case crop up or something like that. Who's, who's caught does that ball land in ?.
Amandine: [00:23:50] So the, the entailed to an intellectual property of the standard itself. the open source implementations, stand in. Well, basically the foundation is, an IP, holding company, and that's, it has the copyrights on the standard and on the open source implementations. up the servers. So it will be responsible for that and all the work that new vector does goes into the foundation itself.
Tim: [00:24:17] Okay. So that's quite a nice structure. I mean, I've seen, I've seen this done, you know, I've seen a bunch of different ways that open source projects are, are funded and done. And quite often it comes down to them being funded either by. Hardware sales, which are kind of pushed forward by the existence of the open source or a or made possible by the open source. Or the other option is that they're funded by. By services quite often with like people, you know, if you think of somebody like red hat or whatever, that they are essentially a service business based around, an open source tree. So this is a slightly different model. I guess it's kind of more red Hatish than it is. I was thinking of asterisk in terms of the hardware, version so that there's no, there's no hardware version at the moment or is that something that might happen.
Amandine: [00:25:08] no, there is the name hardware, version so far. And, but yes, in terms of, so we've, so we've set up a very complicated governance process, but not complicated, more comprehensive governance process to actually define the role of the foundation and I will share the links, where, which explains how the standard is being developed. We have a spec core team. Who is, going through it. Everything is done in the open and anyone can contribute. So we're being very, very transparent on how things work. And then the fact that, the new vector team works for the foundation is both, there's definitely some big discussions in terms of what needs to be done and how to make sure that. we keep working for the foundation and making sure for the ecosystem rather than prioritizing new vectors, interest. But that's where the independent directors on the foundation come in and ensure that they're basically keeping us honest and making sure that new vector is not going rogue. Basically.
Tim: [00:26:15] And, and would you, I mean, I suppose we should maybe track back and talk about why there's this all matters. so what we're looking at here is, is a messaging system that companies and governments I think, are going to use as a way of like basically supporting their day to day businesses and, and in future individuals we use to communicate with their loved ones. This is a, like, it's the kind of. Almost like the modern telephone system, a critical piece of infrastructure. So it sort of does matter how it's done and, and I guess to work with, again, we're drawing a contrast with sort of the other three I mentioned with the SMS and the and the Signal type things, and also Slack where. they have certainly Slack and signal have much higher degrees of centralization and single points of control on them. whereas your, your aiming for a real decentralized solution, I guess, is that, I mean, if you. Thought philosophically about kind of all that, and if so, can you talk about it?
Amandine: [00:27:21] Yes. It's, it's really bringing back the cultural of their communication to the users and indeed, Slack and signal very centralized. You have to trust the person, the people who run them. And some might say that one single server you trust. Is good. You still have to trust the server on who knows if it can be trusted forever. the SMS side of things, it's been defined by big corporations and it's still, even if it's done in the open, it's a controlled by, by the GSMA itself. Basically. And it's, it's a very, it's not like anyone, well, I guess anyone can run their, their SMS gateway somewhere, on the have it, but it's quite well regulated. so we asked, what we're trying to do is very much like the web, give everyone the opportunity to be here and on their communication wise, participating in the wider network and basically free. Everyone's communication from, from the control of the corporation.
Tim: [00:28:28] But, but you haven't chosen to do this through existing standards bodies. I mean, the web has got a couple of standards bodies, which mostly agree. And the Internet's got a couple of standards bodies. Have you, have you kind of gone down that path or did it seem too early or what.
Amandine: [00:28:46] Yeah, it seemed a bit too early. And also we, the team who've, who built matrix have been working together for 15 years on instant messaging. I've been playing with XMPP, have been playing with SAPE, have been building their own proprietary stuff, and basically we . Relied on our experience and the idea of, making sure rather than specifying something and maybe it works. And then it's a slow process and it's designed by committee. We wanted to check whether the idea of big, moving fast and implementing things, the why we were defining them. would be working, because then it allows you to be much more flexible in the definition. And it's. It's been the de facto standard that we've written, basically relied on our experience in the domain to put something together in the open that anyone can use without the heavy lifting of having to deal with a big standards body. However, the idea is very much that at some point we, it would be ideal to actually. Start talking to W3C or IETF and see how to embed a matrix in. I'm in one of these bodies, but we reached 1.0 only in June, and until we had a stable definition of the standard, it was much more easy to actually. iterate quickly, because otherwise we would never be there. It was, it already took us like five years to get to the point where we decided the standard was stable enough and could be upset out of beta. So if you, if we had been dealing with, standards, building an addition to that, it would have been much slower.
Tim: [00:30:35] Yeah. I mean, you're, so you're emphasizing the kind of running code aspect rather than, you know, the, the theoretical kind of standardization first and, and, yeah. Okay. That, yeah. No, that's, that's a, that's a working model. So when you say you've got to one.zero do you have any sense of how many users you have.
Amandine: [00:30:59] So today the matrix network is 12 million users, and we estimate about 40,000 deployments. I say we estimate because, people don't have to actually participate in the wider network, and we don't necessarily know about every single server run out there. There are people are not forced to report their a servers from home. So it's, it's, yeah. It's an estimation. Sometime we find out that there is one server running with a few hundred people on it and then it disappears again because they had enabled federation briefly by mistake or something, which is quite interesting. So, so yeah, that's roughly the size we see today.
Tim: [00:31:38] That's quite an achievement for five year project. Congratulations. You must be pleased.
Amandine: [00:31:44] Thank you. Let's switch. Can be even better. Wholeness. Nestle
Tim: [00:31:51] and, and, and I, we mentioned earlier, if I understand it correctly, you are actually getting interest from, from governments as well as from individuals in open source projects is a pretty wide range of user base.
Amandine: [00:32:03] Yes, indeed. we had the, one of the first users of matrix, not the first, but it's since two years. we've been working with the French government who's been deploying matrix as their internal communication mechanism, basically replacing telegram and WhatsApp. So they've deployed it across the entirety of the governments across all 16 ministries, and they have actually 60 deployments out there. With, I think it's almost a hundred thousand users now, so it's, it's growing. And, they have forked a riot who is, which is opensource to build their own app, which is called Tchap, which is open source as well. And, and yeah, it's basically every ministry keeps control, complete control over their server. They can apply the level of security they require because of course, different subculture don't have the same antiviruses for example. And yet they can still talk to one another and, install the app on there. There were no phones if they don't have a professional ones. And, and they, they organize the G7 earlier this year on the, all the security was coordinated across Tchap, which is quite interesting.
Tim: [00:33:14] That's cool. That's very cool. So, so to what extent you feel this is a European project or shaped by kind of European values?
Amandine: [00:33:24] we're. I think the fact that the French government is using it is giving a great example to others and we've had interest from other governments, looking at, at what they've been doing. It's a good case study to go around and there is definitely interest for, from governments to actually use it to communicate from one between them. so there is on one side, we need a secure messenger for our governments. On the other side, there is. Would be great to have a secure messenger to actually talk to other governments. So, that's something which is really interesting.
Tim: [00:33:58] Yeah, no, that's, that's quite exciting actually. and, and do you find, I mean, are you finding that it's taking root, you mentioned America, but are you finding that it's taking root in, say, Asia anywhere else, or is it fairly European based still?
Amandine: [00:34:16] there is, there is some level of usage in, in Asia, although they're pretty independent and we don't, well, we have a bit of visibility, but not much. But yeah, basically, so far we've been working a lot with Europe and, North America, so Canada and the U S governments, all the, emergency management. Department of Texas is actually coordinating as well on, on matrix. And I'm trying to, to push this towards the other States. So that's the kind of things we've been saying.
Tim: [00:34:49] So you said something quite, which I want to go back to. You said, and I'm trying to remember the exact words, but it was something to the effect of. having to trust the organization that was providing the software and the service. So, you know, example might be signal on Slack now and into the future. And, and there's a hint there about kind of the funding models. I mean, we've seen a couple of examples of this where the funding model has forced the company to do something which. It's initial users might not have expected, did that some, is that something that you kind of worried about throughout the process
Amandine: [00:35:27] while raising money from, VCs? this year was definitely, A bit scary, I would say because yeah, we needed to find people who actually believed in the big plan. We're trying, and the big thing we're trying to build here and making sure they understood. We're building an ecosystem and we were, we're working on something which has 10 year return rather than two. So it's, we needed to find people who we, who we could trust and actually buy into the vision. And I think we did, we had great, blog posts from all of the three VCs who, who funded new vector this year about the dream of Matrix, and, and how they're, they're actually looking forward for the, to them. It's time now to actually succeed. So, so yeah, it's been an interesting exercise. but I think we found good partners who are actually understand the plan and understand that the bigger ecosystem you build, the more opportunities to make money on top of it. And it's much better than build a closed silo who may spiral up to the top in a few years, but won't last.
Tim: [00:36:38] Yeah. I mean, I th I think that's, that's always a challenge of like, you know, how do you. How do you monetize it in the future? How do you get your 10 X return or whatever it is that they're looking for without, you know, doing something too, too radical against privacy or, or overcharging or something like that. There's always like a bit of tension at some point in there in the process, but if I understand it correctly, you built the governance structure in order to protect the protocol. Is that kind of fair.
Amandine: [00:37:12] Yeah, yes. Because the, we've spent a lot of time and a lot of money putting this together, and, we believe that we have a structure, which is a transparent and fair, which is protecting the standard itself. by having. The founders as a minority on the board of director of the foundation by having a spec core team, which has, is pretty transparent on everything it does. And we have a big community and a big vocal community who definitely be keeping an eye on what we're doing and keeping us honest as well on this side of things. And if we do something wrong, we'll be. told so, if, if it looks like new vector is not going in the right direction, then the three directors, three other directors of the foundation can definitely, keep, take back control, and make sure that the protocol itself is safe and working for them, for the, for the good of the ecosystem.
Tim: [00:38:12] And I mean, you've already mentioned that the, Implementations, both of the client and I think also of the server are opensource and that they've already been forked because people had differing different requirements, which is great. I mean, that kind of shows you, it started, it started as open. It's always been open source. It's never been closed. Is that fair?
Amandine: [00:38:34] Yeah. Yeah, yeah. Yes. We, the, when we sat down to start working on matrix, there were a whole bunch of requirements for it. And open was like the top one, a open, decentralized, secure bake end to end encryption in, from in the get go, in it. And, yes. Transparency is hugely important. Otherwise people cannot trust it. And that's also why the governments like it in terms of being able to do a full audit of the end to end encryption, making sure they're not locked into one vendor and can switch to anyone. because it's open source, it's basically what attracts them as well.
Tim: [00:39:15] And has, have any of them published their security reviews or have you had a security review done in yourselves.
So we've made a no data of our end to end encryption, in 2016 and we have a public report of it. we are finalizing the last bits of the user experience, for, to make them to an encryption much easier to use. And once this is done, we will do another audit. It's off. The whole stack. Last time was just, the protocol itself and this time is going to be like from the top, like from the user experience down to the protocol because most of the problems usually come from how you actually use the end to end encryption and how you present it to the user rather than. For the, the cryptography itself. So that's going to happen. The governments have been doing some level of auditing and some other are on the in progress, but, yeah, not, not everything is, is published. I think.
So one of the other things I noticed was we have your 12 million users, a substantial fraction are using your own infrastructure. that's matrix.org, I guess. how much does that. Represent a success. And how much does that represent a challenge to the kind of decentralized nature of the project?
Amandine: [00:40:41] Honestly it is more on the challenge side of things. On one side we need, it's because it's, it's a good way for people to try it out so that just go register an account on these big public free server, but it's breaks the idea that. We want people to be able to choose who they trust, choose, run their own server or get it from someone they trust or, it's kind of things. it's also quite a pain to run because it's really big, I think with 5 million users on it. So it's, we are looking forward to the day where we can turn it off and actually have people spread across multiple servers.
Tim: [00:41:19] You know, that'll never happen.
Amandine: [00:41:22] I'm not sure. You know, Matthew is pretty pushing for it. We are still trying to move towards a more peer to peer structure, so we don't believe that severs are going to go anywhere because it's still useful to have an archive of your discussions. And if you're switching between devices, have a way to catch up on this kind of things. If one of your devices have, and relying only on your phone . The not being stolen or not falling down the sea or whatever, is not always ideal, but. that should help, in terms of, some people actually not wanting, not caring about history or archiving and not choosing servers anymore, but yeah. So, yeah, it's, so
Tim: [00:42:11] yeah, you accidentally become an infrastructure company, which isn't, which I guess wasn't on your list of priorities.
Amandine: [00:42:18] Yeah. But it's part of them that comes with the game.
Tim: [00:42:23] Yeah. Yeah. I mean, I guess you have to prove the point, but it's, I mean, it's nice that it's already under half the users. That's a kind of, that's actually, I didn't know what the numbers were, but it's, it's nice that it's, it's not, not more than half the users. That's kind of comforting in a way. yeah. So just talking about what you were saying about. Using a server, the necessity of servers. So what you're saying is that if I, if I lose my phone or I drop it in the bath or whatever, and, and I need to restore, I can restore from my server or can I restore from my other devices? I mean, we'd kind of digging into technical practicalities here, but kind of interesting a differentiation point.
Amandine: [00:43:06] so that's all part of the peer to peer matrix work, which is barely started. so honestly right now, I'm not sure how things are going to happen. I suspect that will be a way of being able to restore from devices. but then it probably means that your other device needs to be online and turned on so that you can retrieve history from it.
Tim: [00:43:29] Right.
Amandine: [00:43:29] My best guess with the level of technicality, Matthew, would probably have a three hour six variations can give you.
Tim: [00:43:38] Yeah. I know. I, I've, I kind of, I'm genuinely much more interested in the, in what you're doing and the foundation side, because I think the, it's kind of. Well documented and well-described and there've been a ton of technical talks about like how matrix works and that sort of thing. But I think it's really, for me, it's really interesting to see how what you've done is effectively, if somewhat of a new model of how to do a mass distributed open source project. I mean, if you look at some of the, you know, some of the kind of, Crypto coin people, they trying to do something similar, but then when you look at how they're financed, they're really not taking that, that decision fully on board. And, and you know, once you start being funded by your own tokens, then, you're not really independent of anything. Yeah.
Amandine: [00:44:29] Yeah. It's, probably much harder team to manage.
Tim: [00:44:34] Have you found that. Being a decentralized cryptographic system means that people immediately lump you in with blockchain.
Amandine: [00:44:47] when blockchain with huge in 2017, that was, it was often the, confusion. It also helps sometimes to explain because matrix can be similar in the fact that all the, history of messages are cryptographically dependents of the others. But, but we're managing to. Yeah, it's when we explain how it works, it's quite obvious that we're, we're different. And the fact that we use, yeah. Basically all the client server, thing, also helps. So we're, yeah. We're managing to differentiate.
Tim: [00:45:28] Okay. No, that's, that's a, that's been an interesting challenge for us as well. So I'm, I'm kind of fascinated to see how you've got on with that. Cool. So I think if, if you, unless you've got kind of something else that you think I've missed out on, I'm just gonna I've got, after that, I've got one last question. So is there anything else you think we haven't really covered.
Amandine: [00:45:50] I don't think so. I think we've gone through mostly everything, but yes, in terms of business models to fund the open source projects, on one side, the startup is VC funded, but then yes, professional services to help people deploy the, deploy and run the open source software. And of course, hosted services where people can just go pick up their own setup, their own matrix server with, with, their DNS is pointing at it, they control it. So that's, that's what we do.
Tim: [00:46:22] So, given that we're kind of, the podcast is distributed future, I'm very interested in finding out where you see. This technology and this business structure being in five years time. I mean, you know, I know it's a guess, but where do you think you'll be or where do you hope to be in five years time?
Amandine: [00:46:43] the, I would go even further than five years. That may be even the 10 year plan that theidea in the end is really, that matrix becomes a layer of communication and everyone has an app. Built from matrix on their phone and their, these apps may be built by a new vector or may not. And we hope to see many companies actually selling hosted services and professional services, on top of matrix, like new vector do. So we're expecting the model, the new vector model to duplicate and replicate. so that's. User actually have a choice of who they want to go with and who they trust. And I'm definitely expecting the foundation to stay where it is and roughly the same structure being the independent custodian of the standard. And while other people are just making the most of it and providing the best services. And then you'll get into healthy competition where it's the best service, the best app, which actually wins rather than the one which has the biggest.
Tim: [00:47:50] Okay. That's interesting. Cause then you know, cause the big selling point for all of the communications app apps at the moment is, Oh well, you know, there are more people on it, more of my friends are on whatever, so I have to be there. And so, and your decoupling. Yeah. Interesting. You're decoupling the app from the protocol. It's the same way that SMS did, I suppose.
Amandine: [00:48:12] Yeah. And then it's the best service. So if you, it's the best service, but also it allows people to build very niche applications. And ideally I'm quite a power user of magi changing apps at this point. I think I can say so what I use is probably not going to be the same that my grandmother is using on her. Smartphone. Then she wants something simple and easily accessible, but she still wants to talk to me. And that's basically where we, I mean, for.
Tim: [00:48:41] And that I guess, I mean this is somewhat out of the, left field to some extent, but I guess that, you know, in principle you could put matrix on one of these smart devices so you could have a thing, I mean, not Alexa because that's very centralized or so, not Google home, cause that's very centralized. But one of these voice assistants could actually be. a matrix endpoints. So your grandmother or whoever could actually literally talk to the device and then it would communicate with you over whatever channel over the matrix. Cloud do call it the cloud
Amandine: [00:49:18] network
Tim: [00:49:18] network and a, and then that would end up on your smart phone, which the only thing in common there is the protocol rather than, Any ownership or any branding or anything else. Oh, actually, that's something I should have asked you. What's the situation with the brand? Do you like who owns the brand? The name matrix, and who enforces it? Or did you answer that and I forgotten.
Amandine: [00:49:42] it's so it's owned by the foundation. I think with the matrix logo, I think it's trademarked, but, otherwise the name is a common name, so we cannot trademark it, but it's basically the foundation who owns it and manages it.
Tim: [00:49:56] Okay. Cool. Listen, thank you so much for that. I really do appreciate it. It's always kind of good to dig into this stuff and, and, and find out more. So on which note, if you have got some links that you think are, would be interesting to people, that would be great. If you could send them through. I'll put them on the, on the, what we call show notes
Amandine: [00:50:15] great. Thank you for having me. And, and yeah, giving me the opportunity to talk about matrix.
Tim: [00:50:20] No, it's great. Lovely. Thank you very much.
Okay. Thank you.
Bye.