1 00:00:00,017 --> 00:00:04,117 Hello everybody, I hope you can hear me and I'm very happy to be here. 2 00:00:04,237 --> 00:00:08,457 Thanks for showing up all of you and hello everybody out there on the internet 3 00:00:08,457 --> 00:00:10,117 who will watch this video later. 4 00:00:10,677 --> 00:00:15,437 The topic is open washing and the fact that you are here makes me assume that 5 00:00:15,437 --> 00:00:19,417 I don't need to explain what open washing is. Should I? 6 00:00:21,037 --> 00:00:26,477 Yes, please. Yes, please. Okay. Okay, so from two years ago, 7 00:00:26,557 --> 00:00:30,877 I noticed that there is a trend coming up in proprietary software vendors that 8 00:00:30,877 --> 00:00:38,017 wanted to use the term open source as a quality label for their software, 9 00:00:38,157 --> 00:00:43,857 despite the fact not publishing their software or not following the rules or 10 00:00:43,857 --> 00:00:46,877 the characteristics of open source that we all know. 11 00:00:47,797 --> 00:00:52,457 We know how to publish. We know the licenses and stuff. Just a short example, 12 00:00:52,797 --> 00:00:55,897 a company that was doing a digital twin asked me. 13 00:00:56,357 --> 00:01:00,517 So I'm Markus. I'm a senior journalist. I was a journalist at Linux Magazine 14 00:01:00,517 --> 00:01:02,917 and Heise here in Germany, tech magazines. 15 00:01:03,257 --> 00:01:06,517 I have a small company from Regensburg where I do open source consulting. 16 00:01:06,757 --> 00:01:10,897 And I'm working for Grumonio, which is doing an open source exchange replacement. 17 00:01:11,157 --> 00:01:12,597 A company startup from Jena. 18 00:01:12,957 --> 00:01:17,137 And with my own company in Regensburg, I was asked by a company who was doing 19 00:01:17,137 --> 00:01:24,377 digital twins whether I could help them to win open source tenders for the state, 20 00:01:25,077 --> 00:01:26,897 which have been very common. 21 00:01:26,997 --> 00:01:32,697 It's very common to request open source in tenders for public offers. 22 00:01:33,157 --> 00:01:37,457 And they said, we always lose the tenders because we're not open source. 23 00:01:37,657 --> 00:01:43,097 So we don't want to publish our IP. We will never publish our source code because it's so valuable. 24 00:01:43,457 --> 00:01:47,757 And so we will never publish. We will never share. but we need to label open 25 00:01:47,757 --> 00:01:49,517 source so that we win the tenders. 26 00:01:49,757 --> 00:01:52,917 Because our competitors who do open source, they win the tenders and we fail. 27 00:01:53,137 --> 00:01:58,177 And I suggested them to do a workshop of about three days on the basics of open 28 00:01:58,177 --> 00:02:00,797 source. And that was the end of the business relation. They never came back. 29 00:02:01,657 --> 00:02:05,437 And that was the first time I had this encounter with open washing. 30 00:02:05,657 --> 00:02:09,717 And since then, many things have happened that we will also talk about in the course of this panel. 31 00:02:10,877 --> 00:02:16,277 And did that explain? There's a lot of companies out there who also from the 32 00:02:16,277 --> 00:02:22,557 old and good old open source world who have retreated a little bit from the 33 00:02:22,557 --> 00:02:24,377 values that we actually share. 34 00:02:26,217 --> 00:02:31,557 And that leads me directly to my first guest, which is sitting right next to me. 35 00:02:31,577 --> 00:02:35,837 I just want to say shortly who is here. So this is Cornelius. You all know Cornelius. 36 00:02:36,297 --> 00:02:41,097 I think he's doing KDE longer than the KDE exists or something like that. 37 00:02:41,957 --> 00:02:46,117 I've known him for many decades. It's now, it's so, wow. 38 00:02:47,077 --> 00:02:50,297 Okay, that's also Cornelius from KDE. 39 00:02:50,597 --> 00:02:56,137 We have Holger Dürer from OwnCloud. I guess many of you also know him and you know own cloud, 40 00:02:57,017 --> 00:02:59,877 Yeah, we also know each other for decades and the same applies to 41 00:02:59,877 --> 00:03:04,457 Richard Heigl who is from Blue Spice media Wiki enterprise media wiki software 42 00:03:04,457 --> 00:03:08,517 from Regensburg like an Atlassian replacement in open source based on media 43 00:03:08,517 --> 00:03:13,517 wiki also completely open source And I'm very happy to have Leonard Kugler here 44 00:03:13,517 --> 00:03:17,397 from the senders the tender for digital sovereignty, 45 00:03:18,057 --> 00:03:21,897 Which is an institution of the how do you say in English the Secretary? 46 00:03:21,897 --> 00:03:26,857 State Secretary, the Ministry of Interior Affairs, the Bundesinnenministerium. 47 00:03:27,877 --> 00:03:33,077 Sorry. And the Center for Digital Sovereignty is running a few really, 48 00:03:33,117 --> 00:03:34,617 really cool open source things. 49 00:03:34,837 --> 00:03:38,997 And they are so cool that they've been invited to the United Nations only weeks 50 00:03:38,997 --> 00:03:43,577 ago, that they have invitations and guests from Japan, even from Taiwan and 51 00:03:43,577 --> 00:03:44,757 from other European countries. 52 00:03:44,937 --> 00:03:50,217 Collaboration with France is a big thing. And one of the things they do is open 53 00:03:50,217 --> 00:03:53,477 source desktop for the public administration called OpenDesk. 54 00:03:53,497 --> 00:03:54,977 You can read about it on Golem. 55 00:03:55,097 --> 00:03:57,417 Guess who wrote the article and did the testing? 56 00:03:57,657 --> 00:04:04,317 And what Mr. Kugler is for is standing for he's the director, 57 00:04:04,437 --> 00:04:08,457 I guess, for OpenCode, which is an open source repository for the public sector 58 00:04:08,457 --> 00:04:12,217 with some 6,000, more than 6,000 users, 59 00:04:12,437 --> 00:04:15,377 more than 5,000, more than 600 projects. 60 00:04:15,377 --> 00:04:24,337 And so we have the public sector, we have two leaders of small and medium enterprises, 61 00:04:24,597 --> 00:04:27,197 and we have the Wikipedia community, 62 00:04:27,617 --> 00:04:33,657 we have the cloud community, and I'm sure Holger also knows quite some stuff about AI. 63 00:04:33,657 --> 00:04:37,737 And we have Cornelius from KDE with whom I would like to start. 64 00:04:38,117 --> 00:04:42,417 And then I will be silent and just ask me questions, which you will appreciate. 65 00:04:42,717 --> 00:04:48,157 Because those are the guys that know shit. I'm only the one who asks and who 66 00:04:48,157 --> 00:04:50,717 brings the bad news or the messages. 67 00:04:51,537 --> 00:04:58,737 So when we were preparing for this panel, I had a video conference with Cornelius this week. 68 00:04:58,737 --> 00:05:02,577 And not to my great surprise, 69 00:05:02,757 --> 00:05:10,537 but it was striking that Cornelius said that open washing is the biggest threat 70 00:05:10,537 --> 00:05:15,457 currently to our open source values. Did I quote you right? 71 00:05:16,137 --> 00:05:18,497 And that you're pretty scared of what's going on. 72 00:05:19,943 --> 00:05:26,663 Can you elaborate why there is such a big fear in you? Sure. 73 00:05:28,283 --> 00:05:32,243 One thing you said, I'm in KDE for decades. 74 00:05:32,643 --> 00:05:36,903 It's pretty much exactly 25 years now since I'm doing KDE. 75 00:05:37,423 --> 00:05:42,403 I'm not the first or oldest contributor. There are people who are around for even longer time. 76 00:05:42,803 --> 00:05:48,143 And the reason why I'm here and still here is that we are doing free software. 77 00:05:48,143 --> 00:05:51,083 I wouldn't be here if this would be just a job. 78 00:05:51,823 --> 00:05:54,763 And this doing free software, I think that's a call. I mean, 79 00:05:54,803 --> 00:05:56,823 that's why all of you are here as well. 80 00:05:57,043 --> 00:06:00,983 And free software for me has these two parts. The software, that's technology, 81 00:06:01,483 --> 00:06:05,003 writing great applications and so on. We want to use that. We all want to have 82 00:06:05,003 --> 00:06:06,163 fancy stuff on our computers. 83 00:06:06,323 --> 00:06:08,683 That's all fine and nice and it's enjoyable. 84 00:06:09,003 --> 00:06:13,663 But we are also here because it's free. There are values behind that. 85 00:06:14,463 --> 00:06:19,123 Sometimes you call that open source for pragmatic reasons. It's probably more 86 00:06:19,123 --> 00:06:22,943 or less equivalent, but they're really saying the freedoms behind that, 87 00:06:23,103 --> 00:06:27,003 which are formulated in the four freedoms on the open source definition, 88 00:06:27,363 --> 00:06:32,923 which is also reflected in KDE's vision, this giving people control about their digital life. 89 00:06:33,183 --> 00:06:35,203 So they're really strong values behind that. 90 00:06:35,763 --> 00:06:40,463 And the software part, I think, is something, I mean, I won't deny that other 91 00:06:40,463 --> 00:06:44,403 The organizations, also companies which have no interest in freedom, 92 00:06:44,563 --> 00:06:50,083 are able to create usable and useful software, which people love to use. 93 00:06:50,783 --> 00:06:56,143 But what they don't do is this freedom part, which enables us to work together 94 00:06:56,143 --> 00:06:59,703 as a community, but also which gives users these freedoms. 95 00:06:59,863 --> 00:07:04,163 And that's the very important part, that users get the freedoms as well, 96 00:07:04,203 --> 00:07:05,743 not only the developing people. 97 00:07:06,899 --> 00:07:14,739 And with open washing, what we see is that many companies see that there is 98 00:07:14,739 --> 00:07:17,299 advantages in using the term. 99 00:07:17,359 --> 00:07:22,079 You can use it for marketing and they try to take this term for their own purposes 100 00:07:22,079 --> 00:07:23,579 without actually being open. 101 00:07:23,779 --> 00:07:26,039 And in the end, what is happening? 102 00:07:26,279 --> 00:07:32,979 I mean, they are taking away our language. So we can't clearly communicate anymore 103 00:07:32,979 --> 00:07:36,019 about our values because other people do it in a different way, 104 00:07:36,039 --> 00:07:40,419 mean different things, and people get used to Open as something different. 105 00:07:40,679 --> 00:07:45,319 And I think that's really a threat, a threat to us because we are not able to 106 00:07:45,319 --> 00:07:48,379 communicate our values clearly anymore. 107 00:07:48,539 --> 00:07:52,759 People don't understand that anymore. more, users don't understand anymore what 108 00:07:52,759 --> 00:07:56,319 open actually means, that this open thing they get from a proprietary company 109 00:07:56,319 --> 00:08:02,039 which takes away all their freedoms and concerns their privacy and all that, 110 00:08:02,099 --> 00:08:03,299 that this is not really open. 111 00:08:03,539 --> 00:08:08,119 So I think for us it's very important that we have clarity about the terms so 112 00:08:08,119 --> 00:08:10,759 we can communicate and we should fight for this. 113 00:08:11,019 --> 00:08:15,639 And when I say we, I mean us as KDE, as a free software community. 114 00:08:15,959 --> 00:08:20,159 All other free software I think they also need to do that and also the companies 115 00:08:20,159 --> 00:08:24,439 who are generally doing free software are generally doing open source, 116 00:08:24,639 --> 00:08:27,939 I think they are sitting in the same boat and also all the companies who are 117 00:08:27,939 --> 00:08:32,879 using open source and want to have clarity about how this works and what they get. 118 00:08:33,159 --> 00:08:36,439 So that's why I think it's really a threat we have to fight. 119 00:08:37,372 --> 00:08:42,612 I think it's very common also in the cloud, in the world of clouds, isn't it, Holger? 120 00:08:43,932 --> 00:08:47,632 I've heard the term open source as a service. 121 00:08:47,732 --> 00:08:54,332 That is, for example, companies that use open source, produce open source code, 122 00:08:54,392 --> 00:08:58,452 license it with open source code, but they don't hand over the source code to their customers. 123 00:08:58,492 --> 00:09:01,952 They're just developing stuff in order to run it on their cloud, 124 00:09:02,032 --> 00:09:05,932 on their SaaS system, and the customers will never get it. 125 00:09:06,812 --> 00:09:12,192 And since you are from the cloud, well, maybe you also can refer to the world 126 00:09:12,192 --> 00:09:14,652 of AI, because I know you're also deeply involved in that. 127 00:09:14,752 --> 00:09:19,372 And that's the same thing. There is a lot of products and solutions out there 128 00:09:19,372 --> 00:09:21,992 that claim to be open source, but are they? 129 00:09:22,672 --> 00:09:28,052 Yeah, exactly. We do see this a lot in the cloud with offerings. 130 00:09:28,052 --> 00:09:33,332 We also see changes to other licenses from open source licenses, 131 00:09:33,672 --> 00:09:39,972 which continue to be named open, which is a challenge because, in fact, they aren't. 132 00:09:40,092 --> 00:09:45,932 And if we look at the world of AI, everybody thinks that meta and the models 133 00:09:45,932 --> 00:09:50,272 they do with Lama, Lama Sri, etc., while they're great models, 134 00:09:50,572 --> 00:09:53,572 they are often termed as open. 135 00:09:53,772 --> 00:09:57,312 They are termed as open in all the press releases from them, 136 00:09:57,332 --> 00:10:00,612 etc. But they have a very important restriction. 137 00:10:01,012 --> 00:10:07,192 You can't do more than 700 million users with them, which in the world of open 138 00:10:07,192 --> 00:10:10,612 source communities is very difficult because how do you measure? 139 00:10:10,752 --> 00:10:13,552 How do you count how many people you have there? 140 00:10:13,632 --> 00:10:18,852 And it's just not open source. So, there is an ongoing trend, 141 00:10:18,992 --> 00:10:23,732 you mentioned, Cornelius, the open source definitions, so in the world of AI, 142 00:10:23,992 --> 00:10:28,192 there is currently a definition ongoing from OSI. 143 00:10:29,981 --> 00:10:33,621 Having strict rules, what is open source and what isn't. 144 00:10:33,701 --> 00:10:39,261 And some of my hope is when that is finalized, that we get way more clarity 145 00:10:39,261 --> 00:10:44,061 and a back push to the open washing attempts we currently see. 146 00:10:44,381 --> 00:10:49,221 And it was mentioned in the talk before here, from an open source perspective, 147 00:10:49,421 --> 00:10:54,281 from a true open source AI system perspective, which involves all the weights 148 00:10:54,281 --> 00:10:56,381 and the training data and how you 149 00:10:56,421 --> 00:10:59,701 actually get to the model so we can understand 150 00:10:59,701 --> 00:11:03,041 how they got there and if absolutely 151 00:11:03,041 --> 00:11:06,201 needed of course with some money for training we could 152 00:11:06,201 --> 00:11:08,941 redo the same work but at least we 153 00:11:08,941 --> 00:11:12,741 can understand and have the necessary transparency so 154 00:11:12,741 --> 00:11:19,401 right now only the olmo model really delivers those type of criterias the challenge 155 00:11:19,401 --> 00:11:25,801 with that is it's not as good as the other models are at this time so So more 156 00:11:25,801 --> 00:11:32,721 effort needs to go into making more true open source models and making them better. 157 00:11:32,881 --> 00:11:35,481 And in any case, we should be honest. 158 00:11:35,681 --> 00:11:42,921 The companies should be honest. And we shouldn't be tired of pointing out that 159 00:11:42,921 --> 00:11:44,881 those things are not open if they aren't. 160 00:11:45,281 --> 00:11:48,361 So one more question, Holger, before you hand over. 161 00:11:49,521 --> 00:11:55,781 OnCloud has done its newest product, Infinite Scale, scale at a completely open 162 00:11:55,781 --> 00:11:58,421 source and very radical open source approach. 163 00:11:58,661 --> 00:12:02,141 There is nothing closed source, nothing limited. 164 00:12:02,281 --> 00:12:08,001 There's just an EULA that limits hosters or something like, but the whole software 165 00:12:08,001 --> 00:12:10,721 is freely available to an extent completely. 166 00:12:11,341 --> 00:12:16,481 What was the decision to do that? Why did you trust in this model? 167 00:12:16,721 --> 00:12:20,681 And the second thing is what can other companies, and And that's also a question 168 00:12:20,681 --> 00:12:25,881 that I will ask Richard then, because you are the representatives of open source 169 00:12:25,881 --> 00:12:27,961 companies, of small and medium-sized companies. 170 00:12:29,561 --> 00:12:35,701 And what can others learn from that? And what is the model for success with that? 171 00:12:35,781 --> 00:12:39,621 Why should, I mean, open washing is an evasion strategy. 172 00:12:40,261 --> 00:12:46,581 And we have to show and tell a different model to compensate for that. 173 00:12:46,581 --> 00:12:54,341 So what was your reasons for going that much open source with your newest product, for example? 174 00:12:55,683 --> 00:13:01,703 When we redid OnCloud in a cloud-native way, etc., we were thinking, 175 00:13:01,823 --> 00:13:05,003 what do we do in regards of openness? 176 00:13:05,303 --> 00:13:11,763 And we decided to go into a mono-repository on GitHub. 177 00:13:12,423 --> 00:13:19,263 We decided to even do our roadmap in the open and have as many discussions as 178 00:13:19,263 --> 00:13:24,743 we can have together with the community. I guess part of it because it got started 179 00:13:24,743 --> 00:13:26,743 by our customer community. 180 00:13:27,103 --> 00:13:31,723 CERN was one of the major original contributors to own cloud infinite scale. 181 00:13:31,983 --> 00:13:38,383 And we decided to not leave that path and clearly stay in the open there, 182 00:13:38,503 --> 00:13:41,683 continue to evolve it in that way. 183 00:13:41,683 --> 00:13:48,723 As I said, including a roadmap, which is often a little tricky because you also 184 00:13:48,723 --> 00:13:54,203 see bugs and discussions and those type of things in the issues which for a 185 00:13:54,203 --> 00:13:57,443 commercial company can sometimes be tricky. 186 00:13:58,323 --> 00:14:04,923 However, till now we continue down that path and it serves us well. 187 00:14:05,023 --> 00:14:09,883 However, the system still needs to develop. And we've chosen an Apache license 188 00:14:09,883 --> 00:14:17,503 for it so we could get rid of the community contributor agreement for it, 189 00:14:17,643 --> 00:14:20,623 which also was something people wanted. 190 00:14:20,803 --> 00:14:26,743 And that way, that's the path we're on right now. And if it's successful, we have yet to see. 191 00:14:26,823 --> 00:14:30,343 So everybody is, of course, welcome to look at that. 192 00:14:32,803 --> 00:14:37,303 Richard, your product is also completely open source. 193 00:14:37,663 --> 00:14:42,763 You're in a market where there is one kind of monopolist who is closed source. 194 00:14:43,063 --> 00:14:48,943 And there is also, yeah, there's a variety of products. Not everybody is open source. 195 00:14:49,163 --> 00:14:55,883 But how can you justify being completely open and giving away most? 196 00:14:56,043 --> 00:14:59,483 Well, you have some add-ons that are not. Do you have some add-ons that It would 197 00:14:59,483 --> 00:15:05,723 be considered open core, but how can you justify for your product that you're 198 00:15:05,723 --> 00:15:06,723 giving away the source code? 199 00:15:09,394 --> 00:15:17,094 Yeah. So look, every software has its history, right? 200 00:15:17,334 --> 00:15:22,534 So we come from MediaWiki. MediaWiki is the software of Wikipedia. 201 00:15:22,974 --> 00:15:30,694 So what we do is we take already a free software and reuse it and we upstream again. 202 00:15:30,834 --> 00:15:37,814 And so we are very well connected to the community. And from this, 203 00:15:37,974 --> 00:15:44,474 it's only, of course, we need to be open source. There is no alternative to that. 204 00:15:45,974 --> 00:15:50,034 There is another thing which is, in my opinion, central. 205 00:15:50,154 --> 00:15:58,394 So today you cannot produce any software without using a piece of software which 206 00:15:58,394 --> 00:16:01,514 you haven't produced or is not maintained. 207 00:16:01,934 --> 00:16:05,934 So that is what some open source companies don't get today. 208 00:16:06,314 --> 00:16:12,034 Is that, okay, yes, they want to be in the market, stay in the market and want 209 00:16:12,034 --> 00:16:14,434 to grow. Of course, we'll talk about this a bit later. 210 00:16:14,914 --> 00:16:21,134 But it's also true that they are relying on software they don't provide, right? 211 00:16:21,774 --> 00:16:23,834 Even the big companies don't do this. 212 00:16:24,534 --> 00:16:31,114 So from this perspective, this is an approach that could be or should be, 213 00:16:31,174 --> 00:16:37,654 yeah, Yeah, forced by the community to make this clear in a public view. 214 00:16:38,054 --> 00:16:43,714 But what I would say is, and now I'm coming back to Cornelius saying this is a big threat. 215 00:16:44,694 --> 00:16:50,114 Yes, but we're not only living in a world of ideas, we're living in a world of capitalism. 216 00:16:50,374 --> 00:16:55,874 And if you ask what is interesting for a company, then, of course, 217 00:16:55,974 --> 00:16:59,374 first of all, it's always profit, yes. 218 00:16:59,374 --> 00:17:10,414 Yes, and together with that, you need to stay in the market and you want to reduce costs. 219 00:17:10,694 --> 00:17:14,254 Just take this. I don't justify it. That's how it is. 220 00:17:16,114 --> 00:17:24,934 And what makes any sort of free software and open source successful in our world 221 00:17:24,934 --> 00:17:28,014 is just, okay, you can reduce costs. 222 00:17:28,014 --> 00:17:32,874 You have any sort of benefit out of it and many companies don't see the benefit 223 00:17:32,874 --> 00:17:38,574 and for me the benefit is is clear i don't see there is any alternative and 224 00:17:38,574 --> 00:17:45,194 the notion that you need to have a product and you need to have the property 225 00:17:45,194 --> 00:17:47,834 is so deeply rooted in this capitalist system, 226 00:17:49,319 --> 00:17:52,239 um this is really hard against to fight 227 00:17:52,239 --> 00:17:54,959 against but uh if you if you 228 00:17:54,959 --> 00:18:01,739 look closer than say we have no alternative in in going in an open source in 229 00:18:01,739 --> 00:18:07,499 a collaborative way of producing services and goods that's where we have to 230 00:18:07,499 --> 00:18:13,739 go and we can we need to be realistic in that but i'm not sure if, 231 00:18:14,339 --> 00:18:18,559 other companies just look at their own costs. 232 00:18:18,679 --> 00:18:22,819 So they see just cost, okay, if I maintain a community, if I do upstream, 233 00:18:22,939 --> 00:18:26,039 I have costs. But you get a lot of those things back. 234 00:18:28,359 --> 00:18:34,799 And it's a way to be in a market and stay in a market. And I think that's the point where we are now. 235 00:18:35,519 --> 00:18:41,979 Because the public sector has changed now and say, If you want to sell us something, 236 00:18:42,339 --> 00:18:44,399 you need to be open source. 237 00:18:44,479 --> 00:18:53,799 So in Switzerland, there is a law that they now do open source first. 238 00:18:53,879 --> 00:18:59,419 And in Schleswig-Holstein again and on a federal government level, we see the same. 239 00:18:59,659 --> 00:19:06,419 And that makes many companies now, yeah, they are worried because they don't have open source. 240 00:19:06,419 --> 00:19:11,479 They don't know how this works and how they can change it or should they and 241 00:19:11,479 --> 00:19:13,279 now they try to avoid everything, 242 00:19:14,759 --> 00:19:19,479 and I can understand this but the question is how can we, 243 00:19:20,879 --> 00:19:27,859 help them to convince them this is a good way and not only for single companies 244 00:19:27,859 --> 00:19:32,779 but for the whole economies and the society, 245 00:19:34,259 --> 00:19:35,119 in some. 246 00:19:37,339 --> 00:19:42,979 Leonhard, how can you and the senders help those companies like that? 247 00:19:43,099 --> 00:19:47,859 You're providing an open code repository for the public sector where they could, 248 00:19:48,019 --> 00:19:49,819 for example, upload their source code. 249 00:19:50,179 --> 00:19:54,539 I guess if it matches some conditions. Yeah. 250 00:19:54,659 --> 00:20:01,039 Can you are you like a gatekeeper for making sure that only true real open source 251 00:20:01,039 --> 00:20:05,979 according to our values are in there or how can I imagine that? 252 00:20:07,432 --> 00:20:11,552 I think I would start with the question of the why. Why are we doing open source? 253 00:20:11,792 --> 00:20:15,072 Because I think to have a common understanding of why we are doing it, 254 00:20:15,112 --> 00:20:21,532 what's the value for us, for society, is one of the key factors we need to agree on in a way. 255 00:20:21,732 --> 00:20:29,412 So with Zendesk, we have this mission to work on digital sovereignty. 256 00:20:29,812 --> 00:20:35,692 And the government decided to found Zendesk because there is a realization, 257 00:20:35,992 --> 00:20:41,892 especially in the digital parts of society and how our government is running, 258 00:20:42,192 --> 00:20:48,372 we have really a strong addiction to and clustering on hyperscaler and specific vendors, 259 00:20:48,752 --> 00:20:54,732 which makes us un-sullerant in a way. So we are not able to choose. 260 00:20:54,992 --> 00:20:58,812 So choosing between options is an indicator for sovereignty. 261 00:20:59,212 --> 00:21:04,392 So that's why we say, OK, we need to have the possibility to switch. That's the first one. 262 00:21:04,672 --> 00:21:09,872 The second one is we want to influence those technologies. So if I have requirements 263 00:21:09,872 --> 00:21:13,952 from the public sector perspective, I want to have them in the software. 264 00:21:14,112 --> 00:21:16,892 I want to be able to address those within the process. 265 00:21:18,152 --> 00:21:26,332 So on the requirements side, we need to be able to work with the software and 266 00:21:26,332 --> 00:21:28,092 bring in our own features. 267 00:21:28,832 --> 00:21:32,592 And the third one is that we need to have the power to do so. 268 00:21:32,592 --> 00:21:37,052 So that means like there's no licensing or whatever market saying, oh, you can't do that. 269 00:21:37,732 --> 00:21:43,232 And with public sector, if I take those three very central options, 270 00:21:43,332 --> 00:21:44,772 I would say, yeah, it's open source. 271 00:21:45,852 --> 00:21:50,172 That's the reason why we came to open source with Ascendance. 272 00:21:50,232 --> 00:21:51,692 It wasn't open source in the first place. 273 00:21:51,772 --> 00:21:57,772 If you really think about digital sovereignty, open source is the solution for. 274 00:21:58,961 --> 00:22:02,661 So I think that's the main value if you want to have a democratic and sovereign 275 00:22:02,661 --> 00:22:04,461 society and societal infrastructure. 276 00:22:05,081 --> 00:22:10,721 That's the entry door. And the second important thing is what is open source? 277 00:22:10,941 --> 00:22:15,401 What does it mean? So how can we figure out? And that's the second thing we 278 00:22:15,401 --> 00:22:18,881 did with the platform OpenCode for the public sector. 279 00:22:19,081 --> 00:22:21,221 We said, OK, what do we need to be clear about? 280 00:22:21,521 --> 00:22:26,781 So what are the definitions and rules of open source? With OpenCode, 281 00:22:26,901 --> 00:22:29,021 we said, okay, we don't invent new rules. 282 00:22:29,221 --> 00:22:34,741 There are rules out there. Let's have a look and see how suitable they are for us. 283 00:22:35,101 --> 00:22:39,101 And we started and said, okay, we take the OZ layer and the OZ definition. 284 00:22:39,281 --> 00:22:40,481 That's our starting point. 285 00:22:40,721 --> 00:22:46,341 And we can build up our own open source definition on these OZ definitions and 286 00:22:46,341 --> 00:22:48,101 say, okay, this is the international rule. 287 00:22:48,241 --> 00:22:52,521 We do the same for the public sector. We describe it on the platform, opencode.de. 288 00:22:52,621 --> 00:22:54,121 It's publicly visible. You can 289 00:22:54,121 --> 00:22:58,001 read it in English and in German and we describe it. What does it mean? 290 00:22:58,881 --> 00:23:02,081 Translating in German is also really important because if you want to address 291 00:23:02,081 --> 00:23:05,641 public sector in Germany, you need to make these efforts to understand, 292 00:23:06,281 --> 00:23:08,261 doing a common understanding of that. 293 00:23:09,601 --> 00:23:15,021 And then in the next step, it's about the licensing. Okay, we have a wide range 294 00:23:15,021 --> 00:23:16,881 of open source licensing out there. 295 00:23:16,881 --> 00:23:24,221 And if we have 11,000 public sector communes in Germany, and they all need to 296 00:23:24,221 --> 00:23:28,501 be experts in open source licensing, it will bring us nowhere. 297 00:23:28,761 --> 00:23:33,261 And we don't solve open source and digitalization anywhere else. 298 00:23:33,301 --> 00:23:37,941 What we did was a platform saying, okay, let's do that for the whole public sector. 299 00:23:38,281 --> 00:23:45,201 We already checked around 500 open source licenses and put them on an allow 300 00:23:45,201 --> 00:23:50,401 list and on a D9 list also publicly available on the platform and say, okay, 301 00:23:50,541 --> 00:23:54,001 these are allowed and they stick to our rule of open source. 302 00:23:54,221 --> 00:23:58,841 So that's all the entry doors where we said, okay, these are our policies, 303 00:23:59,001 --> 00:24:03,481 these are our rules, that's how we define it. We do it all public, it's open source. 304 00:24:03,761 --> 00:24:09,741 And now let's get a catalog and bring together all those projects fulfilling those rules. 305 00:24:09,741 --> 00:24:18,901 And now you said it, one thing is having research on public sector and there 306 00:24:18,901 --> 00:24:20,921 is so much open source software in the public sector. 307 00:24:21,161 --> 00:24:23,881 In the last decades it got developed a lot. 308 00:24:24,041 --> 00:24:28,541 So we just started collecting it on the platform and we have now a Git repository 309 00:24:28,541 --> 00:24:31,261 where we get more and more of these repositories quickly. 310 00:24:32,602 --> 00:24:37,842 But the second thing is like the end-to-end process, maybe it's a lot important 311 00:24:37,842 --> 00:24:41,042 point, but it comes to procurement and what are our rules, 312 00:24:41,222 --> 00:24:48,742 ensuring that we bring those software into platform and also into the market. 313 00:24:49,102 --> 00:24:52,322 At this point, I want to encourage everybody of you, ask questions, 314 00:24:52,402 --> 00:24:56,902 interrupt us, raise your hand and ask. And how do we do it with the microphone? 315 00:24:57,442 --> 00:24:59,442 Is there somebody with a microphone or should I? 316 00:25:00,602 --> 00:25:03,422 Yeah. So we already have a question here. 317 00:25:03,502 --> 00:25:11,462 Just before, so the reason that you need a login for OpenCode is also because 318 00:25:11,462 --> 00:25:19,582 the senders or you who run OpenCode are meant to be like a gatekeeper for this. 319 00:25:19,722 --> 00:25:23,922 Yes. With the quality assurance of that there is, for example, 320 00:25:23,942 --> 00:25:29,382 not some marketing buzz project that is just pure open washing in there. 321 00:25:29,442 --> 00:25:30,642 Yeah, that's right. Cool. 322 00:25:31,082 --> 00:25:34,762 The public sector needs someone who they can trust in, and they need to be from 323 00:25:34,762 --> 00:25:35,862 the sector. That's we are. 324 00:25:36,042 --> 00:25:39,642 And we say this platform is from the public. We are working for the public sector, 325 00:25:40,822 --> 00:25:42,142 and we're doing the gatekeeping. 326 00:25:44,341 --> 00:25:50,781 So first of all, I think thank you for externalizing the frustration that I 327 00:25:50,781 --> 00:25:52,401 think many in the KDE community feel. 328 00:25:52,581 --> 00:25:56,161 Because if you look at the demographics of the KDE community, 329 00:25:56,361 --> 00:25:58,701 we are mainly comprised of developers. 330 00:25:59,001 --> 00:26:02,701 We don't have all that many people doing outreach or communication. 331 00:26:03,061 --> 00:26:07,961 So increasingly, when we attend an open source event and we introduce what we 332 00:26:07,961 --> 00:26:12,321 do, you run into people who are like, oh, you're writing code. What a curiosity. 333 00:26:13,981 --> 00:26:16,701 That's also open source and for them it's something 334 00:26:16,701 --> 00:26:19,701 very different my question is 335 00:26:19,701 --> 00:26:23,941 the following I did not 336 00:26:23,941 --> 00:26:28,361 know the term open washing before and of course it immediately made me think 337 00:26:28,361 --> 00:26:33,781 of the similar term greenwashing that is maybe inspired by it that's where it 338 00:26:33,781 --> 00:26:39,341 comes from and I live in Berlin which is full of companies that offer SAS software 339 00:26:39,341 --> 00:26:42,141 for carbon footprint monitoring And of course, 340 00:26:42,141 --> 00:26:46,501 they also sell you an eco-label where you're green just for doing footprint monitoring. 341 00:26:47,241 --> 00:26:52,841 And of course, you have carbon offsets that don't always promise what they actually do. 342 00:26:53,001 --> 00:26:59,101 And I think if I look at the arc of the public reaction to these things, 343 00:26:59,301 --> 00:27:02,721 then at this point, the term greenwashing is well established. 344 00:27:02,821 --> 00:27:06,101 And I think particularly consumers have turned to cynicism. 345 00:27:06,161 --> 00:27:10,161 They hear something is green or they have the checkbox offset your flight. 346 00:27:10,161 --> 00:27:14,381 And at this point, everybody knows it's not holding its promises. 347 00:27:14,601 --> 00:27:19,701 And so people react with cynicism to the green language, let's say. 348 00:27:20,421 --> 00:27:24,181 Can we still avoid this for the term open source and how? 349 00:27:24,921 --> 00:27:30,201 That people will just make that leap to, well, it's open source. 350 00:27:30,301 --> 00:27:34,881 I don't trust that. It's probably trying to sell me onto something that isn't quite true. 351 00:27:35,541 --> 00:27:40,721 Are we there yet? How close are we to that cliff? and can we still turn away from it? 352 00:27:42,568 --> 00:27:45,408 Cornelius you you you moved so i guess 353 00:27:45,408 --> 00:27:48,628 you would like to answer yeah i i think an opinion 354 00:27:48,628 --> 00:27:51,968 about that um i think if we are talking about software the 355 00:27:51,968 --> 00:27:55,388 term open source is well defined we have the osi definition we 356 00:27:55,388 --> 00:27:58,168 have a list of licenses and i think we have more 357 00:27:58,168 --> 00:28:01,008 or less clarity and if a company tries to sell and 358 00:28:01,008 --> 00:28:03,928 license as open source license which is actually not open source 359 00:28:03,928 --> 00:28:07,208 they they get heavy pushback and so far 360 00:28:07,208 --> 00:28:10,508 companies have not been successful with 361 00:28:10,508 --> 00:28:13,428 at least not on on the on the level 362 00:28:13,428 --> 00:28:16,328 that really licenses would publicly get accepted i 363 00:28:16,328 --> 00:28:19,468 mean maybe in terms of marketing terms and having this 364 00:28:19,468 --> 00:28:22,268 open impression but but that's still i think we still 365 00:28:22,268 --> 00:28:26,388 have a chance there i would say we have to be very careful there we have to 366 00:28:26,388 --> 00:28:32,468 fight that um attempts to to open wash the the term we have to support the osi 367 00:28:32,468 --> 00:28:37,408 in keeping up the the standard definition i think that's very important what 368 00:28:37,408 --> 00:28:41,368 actually concerns me is that i think we were talking about AI, 369 00:28:41,548 --> 00:28:44,608 that in this area, I think the ship has sailed. 370 00:28:45,208 --> 00:28:53,328 On the LAMA webpage, it says, this is the open source AI model from and whatever, whatever. 371 00:28:53,628 --> 00:28:57,208 And then you read the license and you see, okay, there's all these things, 372 00:28:57,268 --> 00:29:00,688 there are these additional restrictions, there's this, and the term open source 373 00:29:00,688 --> 00:29:02,908 AI, everybody talks about that. 374 00:29:03,168 --> 00:29:07,768 There's a lot of companies which are publishing open source AI and if you read 375 00:29:07,768 --> 00:29:13,408 the licenses in many many cases they have these restrictions where it's not 376 00:29:13,408 --> 00:29:19,628 open anymore in the sense of what you would apply for open source software. 377 00:29:19,828 --> 00:29:24,548 Of course the situation is more complicated because there are models, 378 00:29:24,668 --> 00:29:25,668 there are straining data, 379 00:29:25,848 --> 00:29:29,408 there is software around that and that's why I also appreciate the efforts of 380 00:29:29,408 --> 00:29:35,028 the OSI or the Linux Foundation and all that to find a better definition there. 381 00:29:35,868 --> 00:29:40,708 But I wouldn't be too confident with this massive money which is flowing into 382 00:29:40,708 --> 00:29:44,948 that and all these commercial activities and a lot of really high stakes which 383 00:29:44,948 --> 00:29:48,788 are there that we as open source software community, 384 00:29:48,968 --> 00:29:53,568 which are not even very much close to this AI community because they are coming 385 00:29:53,568 --> 00:29:59,608 from a different angle and that there it's really hard to save this. 386 00:29:59,808 --> 00:30:03,948 So I think we have to be careful there. We have to make sure that it works for open source software. 387 00:30:04,108 --> 00:30:08,608 And I think we have to find a way how we have this nice positive vision on the 388 00:30:08,608 --> 00:30:12,228 slides before, in the keynote before, how this nice positive vision, 389 00:30:12,348 --> 00:30:14,088 what we can do in terms of activism, 390 00:30:14,328 --> 00:30:20,148 in terms of communication to preserve the spirit of open source there. 391 00:30:20,148 --> 00:30:25,308 Maybe with other words, but I don't think open source AI will ever work for 392 00:30:25,308 --> 00:30:27,808 us in a favorable way as a term. 393 00:30:28,777 --> 00:30:34,537 I have to disagree in a positive way, because I think it's essential that for 394 00:30:34,537 --> 00:30:38,117 those terms, we also make sure that in the world of AI, 395 00:30:38,377 --> 00:30:44,637 it follows our understanding and we have to have a positive push on it. 396 00:30:44,717 --> 00:30:46,837 I don't think the world is over there. 397 00:30:46,997 --> 00:30:50,077 And the rule or the solution, I think, is competition. 398 00:30:50,497 --> 00:30:55,617 If you look into newer models, like from Mistral together with NVIDIA, 399 00:30:55,777 --> 00:30:58,417 they are publishing it now under an Apache license. 400 00:30:58,777 --> 00:31:06,157 And what we're missing now is the data and with the training data for full transparency, 401 00:31:06,417 --> 00:31:08,597 etc. There are lots of copyright issues. 402 00:31:08,957 --> 00:31:12,197 I think that's a matter of time that we find rules for that. 403 00:31:12,317 --> 00:31:16,617 There were great ideas in the talk before that about how that can accomplish. 404 00:31:16,857 --> 00:31:24,317 But I think we have to be positive. if we don't allow people to steal the term 405 00:31:24,317 --> 00:31:29,777 open source for their BSL or SSPL or whatever licenses, 406 00:31:30,097 --> 00:31:35,657 we have as a community to make sure it doesn't happen in AI either, 407 00:31:35,777 --> 00:31:42,257 which more and more will be part of every software solution out there. 408 00:31:42,937 --> 00:31:47,097 Hand over the microphone. I have a question to Richard and then Leon. 409 00:31:47,097 --> 00:31:53,417 I think that we have to separate between we have only some few minutes left 410 00:31:53,417 --> 00:31:58,617 by the way we have to separate between the general AI the gen AI topic and the 411 00:31:58,617 --> 00:32:00,217 small AI topic which is like the. 412 00:32:02,077 --> 00:32:03,817 Autocorrect on steroids I like to, 413 00:32:04,750 --> 00:32:09,390 That's like tools that, for example, that's why I want Richard to say something about that. 414 00:32:09,510 --> 00:32:13,850 That's what I know that also cities like the city of Treusling, 415 00:32:13,870 --> 00:32:20,770 they're using tools like that to do summaries of meetings of the board of the 416 00:32:20,770 --> 00:32:23,650 city, the mayor's board, whatever. 417 00:32:23,890 --> 00:32:29,930 So there is what I call small AI tools and they are present everywhere. 418 00:32:30,010 --> 00:32:34,170 They're also in your tools and they are open source as far as I understand. 419 00:32:34,170 --> 00:32:36,870 And with these tools, they can be completely open source. 420 00:32:37,090 --> 00:32:39,550 Am I right? Yeah. 421 00:32:41,950 --> 00:32:49,590 It's always the same story. So we are not providing LLMs because that's too much for us. 422 00:32:49,790 --> 00:32:55,530 So you can connect to different LLMs, but it's sometimes a bit ridiculous if 423 00:32:55,530 --> 00:33:00,690 you have a sponsor in the public sector which is saying, hey, 424 00:33:00,790 --> 00:33:03,610 we want to use open source. 425 00:33:03,610 --> 00:33:06,770 We want to try it out and then the the 426 00:33:06,770 --> 00:33:09,710 provider on the other hand says oh but chat gpt 427 00:33:09,710 --> 00:33:13,110 is far ahead as you see you see uh that's 428 00:33:13,110 --> 00:33:18,430 like like google analytics and matomo if you don't force the and and and use 429 00:33:18,430 --> 00:33:24,610 matomo then matomo will take much longer to keep up and that is what i see here 430 00:33:24,610 --> 00:33:32,890 coming again because it's easier to use chat gpt and an open eye and stuff like that which is 431 00:33:33,170 --> 00:33:41,970 heavily heavily financed with venture capital and we will see what will be the 432 00:33:41,970 --> 00:33:44,210 result but i'm a bit pessimistic in that way. 433 00:33:45,909 --> 00:33:53,449 So thank you uh yeah maybe maybe i close again with a more optimistic perspective on that, 434 00:33:54,749 --> 00:34:00,069 because i would say if if our our mission was to bring open source and and right 435 00:34:00,069 --> 00:34:05,129 in the heart of how we define society we are we are winning that's what i would 436 00:34:05,129 --> 00:34:09,909 say yes and if if we were then as we had the opportunity you said it in the 437 00:34:09,909 --> 00:34:12,809 beginning to to be in new york at the osmos for Good Conference, 438 00:34:13,249 --> 00:34:15,189 which is the second one. 439 00:34:19,049 --> 00:34:23,189 And 700 people from all over the world came there and discussing those things. 440 00:34:23,329 --> 00:34:26,529 And what we're doing right now with policies and what Switzerland is doing and 441 00:34:26,529 --> 00:34:30,469 France is doing and Germany is doing with policies, but especially not only 442 00:34:30,469 --> 00:34:35,249 from the government side, really like creating real code. 443 00:34:35,429 --> 00:34:39,229 What we're doing right now from government perspective, we're creating code. 444 00:34:39,229 --> 00:34:43,069 We will launch OpenDesk in October for public sector in Germany, 445 00:34:43,189 --> 00:34:47,229 which means we really are on a point where I would say we're winning with open source. 446 00:34:47,669 --> 00:34:53,449 And this recognition brings also we need to discuss again and again what are 447 00:34:53,449 --> 00:34:55,269 the rules for open source, why are we doing it. 448 00:34:55,449 --> 00:35:00,989 And so that's my more optimistic point on that because that's a really good why we're doing it. 449 00:35:01,589 --> 00:35:06,949 That's a wonderful closing statement. I wish we had two or three more hours to discuss this. 450 00:35:08,629 --> 00:35:12,929 I just, I can, if you want to know more about open code and about the Tzendis 451 00:35:12,929 --> 00:35:15,589 and about the United Nations story, go to Gulen.de. 452 00:35:15,689 --> 00:35:17,649 That's where the article is that I wrote. 453 00:35:17,789 --> 00:35:19,729 It's got some hundred thousand readers. 454 00:35:19,949 --> 00:35:23,409 We were totally surprised, but it's really something where, believe it or not, 455 00:35:23,469 --> 00:35:30,069 Germany is leading in the open source world in digitalization of the public sector. 456 00:35:30,109 --> 00:35:32,769 No, you don't believe that. That can't be. No, no, no. 457 00:35:32,889 --> 00:35:36,049 But thanks a lot. And I'm sorry we didn't have more time. 458 00:35:36,469 --> 00:35:39,589 Thanks a lot to all four of us. Thanks, thanks, thanks. 459 00:35:39,789 --> 00:35:45,849 And we try to make a panel or a talk or a whole room at FOSDEM next year about 460 00:35:45,849 --> 00:35:46,989 the topic of open washing. 461 00:35:47,349 --> 00:35:53,509 So feel free to contact me, us, whoever, if you want to be in there. And thanks. Thank you.