[00:00.000 --> 00:10.320] Everyone, we have Boris and Redo here, and they will be providing a talk on breaking [00:10.320 --> 00:16.280] away from Big Tech using open source infrastructure in a convenient way. [00:16.280 --> 00:17.280] Thank you. [00:17.280 --> 00:18.280] Thank you. [00:18.280 --> 00:23.280] So, as mentioned, today we'll be talking about how to break away from Big Tech, and we will [00:23.280 --> 00:29.800] be focusing mostly on small and medium teams, but what we're talking about applies more [00:29.800 --> 00:31.640] broadly as well. [00:31.640 --> 00:33.240] So a little bit about us. [00:33.240 --> 00:34.960] My name is Boris. [00:34.960 --> 00:35.960] And I'm Redon. [00:35.960 --> 00:36.960] Yep. [00:36.960 --> 00:42.240] We've been open source activists for a couple of years, involved in different projects. [00:42.240 --> 00:43.880] And yeah. [00:43.880 --> 00:45.040] Okay. [00:45.040 --> 00:49.520] So to start off, let's talk about Big Tech. [00:49.520 --> 00:52.240] So any fans of Big Tech here? [00:52.240 --> 00:53.240] Show of hands. [00:53.240 --> 00:54.240] Okay. [00:54.240 --> 00:55.240] No. [00:55.240 --> 00:57.400] This was not expected. [00:57.400 --> 00:59.960] So, okay, there are many issues with Big Tech. [00:59.960 --> 01:04.440] We're going to focus on some of them really fast, and we'll not say, probably, I know [01:04.440 --> 01:09.200] you with things that you already know it or hate. [01:09.200 --> 01:13.280] But one thing that sends out is that the amount of money these people have. [01:13.280 --> 01:15.440] It's like, in the beginning was millions. [01:15.440 --> 01:19.240] Some decades ago, now it's trillions. [01:19.240 --> 01:21.880] It's billions now, and millions, billions, and now it's trillions. [01:21.880 --> 01:27.840] So it's probably the market accumulation or wealth accumulation is more than the entire [01:27.840 --> 01:31.520] GDP of France, if you gather all around all this. [01:31.520 --> 01:34.080] And that's just in the first quarter of 2021? [01:34.080 --> 01:35.600] 2021, one year ago. [01:35.600 --> 01:42.200] So probably more money are printed, and they're not going anywhere close to here, right? [01:42.200 --> 01:47.200] Which is, and money is not the issue here, but money also brings power, which is a major [01:47.200 --> 01:48.320] problem. [01:48.320 --> 01:52.960] And in the beginning, in the 70s, as you know, the big oil companies were the one who had [01:52.960 --> 01:56.360] all this market capital and growth. [01:56.360 --> 02:03.880] Of course, Disney, with its nice copyright lobbying issues, and Warner Brothers, just [02:03.880 --> 02:08.240] like Disney, but less efficient in terms of copyright things, as you know. [02:08.240 --> 02:10.760] But again, we're talking about trillions. [02:10.760 --> 02:18.880] And one thing that concerned us a lot is that we've seen many of open source organizations [02:18.880 --> 02:26.880] or companies that are in open source that they do not use, they don't do the one thing [02:26.880 --> 02:28.680] called that we say it's dog-fooding. [02:28.680 --> 02:33.400] They don't use other open source tools for their own infrastructure. [02:33.400 --> 02:37.960] And they do this out of conveniences, you know, because when services like Gmail started, [02:37.960 --> 02:45.080] they were giving away this service for free, free as not as in freedom, which brought a [02:45.080 --> 02:51.080] lot of people into these platforms, Gmail is only one of them, right? [02:51.080 --> 02:55.800] And this was the main problem, because they could afford, because all of this market capital [02:55.800 --> 03:00.720] that they had, they could afford giving all these services for free. [03:00.720 --> 03:04.120] And of course, they also killed a lot of innovative stuff. [03:04.120 --> 03:09.680] I'm not saying that Instagram is innovative, but as you know, they bought Amazon as one [03:09.680 --> 03:11.080] of these good examples. [03:11.080 --> 03:17.800] They're buying everything that is around and which they consider a threat, which might [03:17.800 --> 03:21.560] be a very innovative product at some point. [03:21.560 --> 03:27.680] And one of the scholars, Tim Wu, mentioned that these kind of formula, but by purchasing [03:27.680 --> 03:34.000] other companies that can be considered competitive, this is anti-innovative, right? [03:34.000 --> 03:35.720] And create oligopoly kill zones. [03:35.720 --> 03:39.360] Okay, you know all of this. [03:39.360 --> 03:45.680] But the question is, why do a lot of these organizations, small teams or medium teams, [03:45.680 --> 03:49.760] continue to use this kind of big tech platforms? [03:49.760 --> 03:55.520] And even more specifically, how come that in the open source community, we are so reliant [03:55.520 --> 03:58.120] upon big tech infrastructure? [03:58.120 --> 04:00.400] And there's a couple of different reasons for that. [04:00.400 --> 04:05.000] For sure, one of them is the fact that they are free as in premium and not free as in [04:05.000 --> 04:06.200] freedom. [04:06.200 --> 04:09.800] And this makes it very convenient to sign up with them. [04:09.800 --> 04:16.280] Also the fact that they integrate their services with one another very tightly. [04:16.280 --> 04:20.880] At first, it makes us believe that it makes us more efficient in what we want to do. [04:20.880 --> 04:26.600] But in reality, this just leads to a vendor lock-in, where after a few years, it will [04:26.600 --> 04:29.920] be very hard for you to migrate to something else. [04:29.920 --> 04:37.360] So in our daily job, we work on providing digital infrastructure using open source tools. [04:37.360 --> 04:38.400] And we've seen this. [04:38.400 --> 04:44.320] These organizations that are, again, are into the open source, open science or open knowledge [04:44.320 --> 04:45.320] world. [04:45.320 --> 04:49.320] These organizations started by using the free-me model of these platforms. [04:49.320 --> 04:54.600] And because they couldn't move in easily, now they are locked, right? [04:54.600 --> 04:56.880] And that's why they keep using these platforms. [04:56.880 --> 04:57.880] Okay. [04:57.880 --> 05:04.760] Yeah, so the solution, we think, is we're going to talk about it in two parts. [05:04.760 --> 05:08.120] The first part is the ideological side. [05:08.120 --> 05:11.600] And the second part is the more hands-on and practical one. [05:11.600 --> 05:13.760] So there are proposals for this. [05:13.760 --> 05:17.120] One of those, we are just mentioning, because we don't have much time, we are just mentioning [05:17.120 --> 05:18.640] one of them. [05:18.640 --> 05:23.080] Usually we say big tech, but why should everybody go big, right? [05:23.080 --> 05:29.040] So we should use small tech, small companies, which use open source tools. [05:29.040 --> 05:36.880] There is a small tech foundation founded by, founded some years ago. [05:36.880 --> 05:39.720] And they say the small tech is should be easy to use, private by default, peer-to-peer, [05:39.720 --> 05:43.960] zero knowledge, non-colonial, personal, share-like, interoperable, non-commercial, inclusive, [05:43.960 --> 05:45.320] a lot of stuff. [05:45.320 --> 05:51.280] Again, I'm pretty sure there are other approaches, but this is only one of them that we want [05:51.280 --> 05:52.280] to mention. [05:52.280 --> 05:58.720] So the part two of the solution, all of the solutions, is hands-on, the hands-on approach. [05:58.720 --> 06:00.400] What, okay, this is good. [06:00.400 --> 06:04.960] We need to, we want to move away from big tech, but what to do, how to do it technically. [06:04.960 --> 06:05.960] Right. [06:05.960 --> 06:12.280] And the good, but also maybe a little bad thing is that you have a lot of choices as [06:12.280 --> 06:15.000] to how you do this migration. [06:15.000 --> 06:20.840] So we listed the five most important ones that we've seen. [06:20.840 --> 06:26.040] And we want to go over each and every one of them to talk a bit more about them. [06:26.040 --> 06:30.520] So the first option is what we believe will be an ideal world. [06:30.520 --> 06:35.480] And as a sysadmin, I definitely want this to be real, will be where everybody is able [06:35.480 --> 06:39.360] to host their own stuff on their computers. [06:39.360 --> 06:44.000] And this means that everybody does their own deployments, their own maintenance, their [06:44.000 --> 06:47.680] own security optimizations, and so on. [06:47.680 --> 06:52.040] This can work quite good, especially for personal projects. [06:52.040 --> 06:56.800] But when we're speaking about infrastructure for organizations, it can be a bit harder [06:56.800 --> 07:02.200] to maintain as compared to your own personal data, case in point, an instance shall not [07:02.200 --> 07:05.720] be down for an organization in the same way it can be down for you. [07:05.720 --> 07:10.040] So with more users, there's more complexity as well. [07:10.040 --> 07:14.360] And also it depends on who is using it, like do it yourself. [07:14.360 --> 07:19.160] For example, I don't have the time to do it for my parents, for example, right? [07:19.160 --> 07:21.200] But I can do it for myself. [07:21.200 --> 07:26.080] And the idea is sometimes it's a very good scenario, sometimes it doesn't work. [07:26.080 --> 07:30.640] For the punk movement in the late 70s, 80s, it worked doing a DIY. [07:30.640 --> 07:33.400] We need to see if it also going to work for us as well. [07:33.400 --> 07:38.160] But when we talk about DIY, you need to be a bit careful about what we call the Dropbox [07:38.160 --> 07:39.160] problem. [07:39.160 --> 07:44.120] And what the Dropbox problem is, is that when Dropbox was initially launched, one of the [07:44.120 --> 07:46.280] first comments on Hacker News was... [07:46.280 --> 07:48.400] This is quite famous, by the way. [07:48.400 --> 07:50.480] Or for some people. [07:50.480 --> 07:54.720] One of the first comments was saying, I don't understand why Dropbox needs to exist, because [07:54.720 --> 07:59.200] you can host your own FTP server and use this and that library to do it. [07:59.200 --> 08:01.720] And yeah, from a technical perspective, you can. [08:01.720 --> 08:06.920] But I think time has shown that most users will not want to host their own FTP server. [08:06.920 --> 08:10.480] Is the person who commented this in here? [08:10.480 --> 08:16.960] So if not, please don't have this approach, don't have this approach. [08:16.960 --> 08:21.560] Because this is one of the reasons we have so much big tech right now. [08:21.560 --> 08:24.200] Like we say, yes, we can do it. [08:24.200 --> 08:28.400] And if you do all this and you have your own server at your own place and you do this and [08:28.400 --> 08:34.520] you update and you have SSL certificates and FTP and all that stuff, that works. [08:34.520 --> 08:39.200] But for the wider audience, that's a major issue. [08:39.200 --> 08:42.440] So we need to have a different approach when we propose this. [08:42.440 --> 08:47.800] So the second approach that one can take are what we call no-code platforms. [08:47.800 --> 08:52.120] So these are platforms like joingardens.com or Unihose. [08:52.120 --> 08:57.680] And essentially what they do is that they lower the barrier needed to start self-housing, [08:57.680 --> 09:00.320] because they automate a lot of the processes. [09:00.320 --> 09:05.720] And this is a really great way to not only set up, but also maintain your infrastructure. [09:05.720 --> 09:12.840] However, depending on your specific needs, if you want to do some custom features, it [09:12.840 --> 09:17.440] might be a bit trickier to get them to work exactly the way you want. [09:17.440 --> 09:21.880] But as long as you don't want something very custom, they're a very great way to get started. [09:21.880 --> 09:29.720] And it makes you win a lot of time from the first solution, which is self-hosting. [09:29.720 --> 09:34.240] Because you have a lot of the tools that you need to do in order to automate stuff, right? [09:34.240 --> 09:38.600] And third option is having an internal team. [09:38.600 --> 09:46.520] So again, if you are an organization promoting open knowledge, you might either, again, do [09:46.520 --> 09:48.680] it yourself and get all the know-how. [09:48.680 --> 09:53.000] But you might also have one system administrator or a team inside. [09:53.000 --> 09:59.120] This is good, because you can deploy things as you want and customize them. [09:59.120 --> 10:05.640] Usually there are costs for the hardware or if you go to the cloud, also for the team [10:05.640 --> 10:06.840] that you're going to go. [10:06.840 --> 10:11.040] If you have the budget that's good, if you don't, that's a tricky one. [10:11.040 --> 10:15.720] You should either go to option one or two or to the other two ones that are like option [10:15.720 --> 10:18.360] four or five. [10:18.360 --> 10:26.280] Hosting collectives are, if you know Chateau, lots of small hosting collectives from France. [10:26.280 --> 10:30.080] They are mainly in France. [10:30.080 --> 10:36.760] And they are mainly focused on the collective side of it, which is great, because it's also [10:36.760 --> 10:39.960] a very good approach of providing solidarity. [10:39.960 --> 10:45.600] Some of them are not-for-profit, some of them are collective, some of them are small companies. [10:45.600 --> 10:52.680] But the idea is to provide a good step, a very easy step for other collectives or other [10:52.680 --> 10:58.520] small companies to have open source infrastructure on their own. [10:58.520 --> 11:04.000] And some of them are very-that we know, they also provide some training so that it's easier [11:04.000 --> 11:09.240] than migration from all these evil platforms to the new platforms that they're using. [11:09.240 --> 11:18.920] We usually tend to-lots of people who are technically know how we tend to underestimate [11:18.920 --> 11:23.520] how hard it is to change the routine from one platform to the other. [11:23.520 --> 11:29.040] And these setups, these collectives are great in doing that. [11:29.040 --> 11:34.320] And the solution number four and five are kind of very close together. [11:34.320 --> 11:40.520] But the solution number five is that recently there has been an increasing number of providers [11:40.520 --> 11:43.520] that focus on open source infrastructure. [11:43.520 --> 11:48.760] For example, there's a GitHub repository on NextCloud called Providers. [11:48.760 --> 11:54.560] And these providers are not officially endorsed by NextCloud, but they take care of setting [11:54.560 --> 11:57.880] up NextCloud and maintaining it for you. [11:57.880 --> 12:05.920] And this is great because, for example, the NextCloud ecosystem has also official partners [12:05.920 --> 12:06.920] of NextCloud. [12:06.920 --> 12:09.640] So if you are a big organization, you just go there. [12:09.640 --> 12:13.160] If you are a small one, you just go to the list and research and do your own. [12:13.160 --> 12:16.000] Or you can do it yourself, as we mentioned before. [12:16.000 --> 12:21.960] Having all these options in a clear way makes it easier for people to just migrate for, [12:21.960 --> 12:27.520] I don't know, for my parents to start using NextCloud for their photos on their Android [12:27.520 --> 12:33.880] phone, which, by the way, I should update at some point because they asked them to-for [12:33.880 --> 12:34.880] me to help them. [12:34.880 --> 12:40.120] But it's important to keep in mind that it's not necessarily a one-size-fits-all. [12:40.120 --> 12:45.000] And even if you find something that fits your needs, your needs might change in the future. [12:45.000 --> 12:50.040] So it's important to think of these solutions as different alternatives and different steps [12:50.040 --> 12:54.960] that you can take on the journey to have open source infrastructure. [12:54.960 --> 13:01.080] So yeah, about managed service providers, as we mentioned, we think it should be very [13:01.080 --> 13:07.480] important to focus on platforms that are open source, so therefore open platforms. [13:07.480 --> 13:13.280] And with such service providers, usually you don't need to have technical knowledge to [13:13.280 --> 13:14.600] get things up and running. [13:14.600 --> 13:20.720] You need to have, though, some knowledge, which is legal, but also from the provider [13:20.720 --> 13:26.520] that you are choosing, of where your data are, if it's compliant, and all these things. [13:26.520 --> 13:31.600] So you need some basic knowledge to understand what the other side is doing with all your [13:31.600 --> 13:34.080] infrastructure that you are managing, right? [13:34.080 --> 13:39.160] And these providers should also offer you with not just technical support when something [13:39.160 --> 13:42.520] is down, but we believe also with user support. [13:42.520 --> 13:48.320] Because if you have an infrastructure, but nobody uses it, what's the point? [13:48.320 --> 13:52.760] And because managed service providers should not be the end goal, the end goal is to everybody [13:52.760 --> 13:54.800] to self-host, right? [13:54.800 --> 14:00.160] But until then, we should be able to understand the terms of service to read them because [14:00.160 --> 14:03.800] somebody else is doing the maintenance, and they can do something. [14:03.800 --> 14:08.560] They might have higher expectations, but there are also mistakes that people do usually in [14:08.560 --> 14:10.160] these cases, right? [14:10.160 --> 14:14.520] And so in this case, terms of service are very important, not in this case always. [14:14.520 --> 14:18.680] And also to understand the legal coverage, as I mentioned before, where the service are, [14:18.680 --> 14:20.560] where they deployed, et cetera, et cetera. [14:20.560 --> 14:24.960] And of course, very important, service continuity. [14:24.960 --> 14:31.840] There are many such service providers that provide open source platforms as managed service [14:31.840 --> 14:36.040] that, you know, they popped up, especially some years ago, pop up, they are there, and [14:36.040 --> 14:40.560] after like two years, they say, oh, didn't work out for us, the pricing that we calculated [14:40.560 --> 14:42.520] was bad, and they shut down. [14:42.520 --> 14:50.120] So you also need to review your service provider if they seem to have a business continuity [14:50.120 --> 14:53.280] and sustainability plan. [14:53.280 --> 14:54.280] So yeah. [14:54.280 --> 14:58.000] So one example is, for example, Mastodon, right? [14:58.000 --> 15:00.960] There's a lot of talk about hosting Mastodon these days, et cetera. [15:00.960 --> 15:05.320] So you need to know where to deploy it, and what you can do it again. [15:05.320 --> 15:07.720] You can do it yourself, but you can show someone else to do it for you. [15:07.720 --> 15:11.800] You need to have technical know-how with all the platforms. [15:11.800 --> 15:13.040] Platform-specific know-how. [15:13.040 --> 15:14.840] What does federation mean, for example? [15:14.840 --> 15:16.320] Or what's the toot? [15:16.320 --> 15:19.560] Legal implications and who does the moderation. [15:19.560 --> 15:25.280] All these things are very, very, very important for you to know. [15:25.280 --> 15:27.800] And five more seconds. [15:27.800 --> 15:34.120] So that's why, for example, for us, it took us months to understand and read a lot of [15:34.120 --> 15:42.640] legal paperwork and also research the platform before deciding to offer it to other people, [15:42.640 --> 15:45.200] which we are planning to do this week. [15:45.200 --> 15:49.160] And this is something that we are announcing just here today. [15:49.160 --> 15:51.080] And that's, yeah. [15:51.080 --> 15:52.080] Yeah. [15:52.080 --> 15:56.040] Something to keep in mind is that regardless of the option that you choose. [15:56.040 --> 15:57.640] You can go very quick. [15:57.640 --> 15:58.640] Yeah. [15:58.640 --> 15:59.640] Yeah. [15:59.640 --> 16:00.640] Very quickly. [16:00.640 --> 16:04.280] Be careful to not be vendor locked in, because that's a very important aspect. [16:04.280 --> 16:09.760] And if you want the sticker, we have it around with us so you can get it later. [16:09.760 --> 16:10.760] So thank you. [16:10.760 --> 16:11.760] Thank you so much. [16:11.760 --> 16:28.800] Thank you very much.