WEBVTT 00:00.000 --> 00:08.000 So, let's give a round of applause to Yos. 00:08.000 --> 00:13.000 Thank you. 00:13.000 --> 00:17.000 So, we're going to misread the clock officially. 00:17.000 --> 00:19.000 And, of course, we're at Phos Dems. 00:19.000 --> 00:25.000 I'm going to obviously talk about something completely open source, with no shenanigans there. 00:25.000 --> 00:31.000 How many of you, like, know next lap quite well, and are here for an update, okay? 00:31.000 --> 00:33.000 That's pretty much everyone. 00:33.000 --> 00:36.000 And, how many really are like next lap next lap? 00:36.000 --> 00:38.000 I guess I'd like to know more, okay? 00:38.000 --> 00:40.000 That's really just a few. 00:40.000 --> 00:41.000 Oh, class, even, yeah. 00:41.000 --> 00:44.000 I'm happy to give you a one-on-one, even. 00:46.000 --> 00:50.000 Yeah, it's essentially at a placement for Microsoft 365. 00:50.000 --> 00:54.000 So, teams office, one drive, the whole shebang. 00:54.000 --> 00:58.000 But open source and running on your own server. 00:58.000 --> 01:00.000 That's a short of it. 01:00.000 --> 01:05.000 And you're going to see much of the features in the next, you know, 30-ish minutes. 01:05.000 --> 01:10.000 One thing I want to talk about very shortly is, so two features. 01:10.000 --> 01:14.000 I want to quickly talk about in general on this Federation. 01:14.000 --> 01:18.000 So, as next out is designed to Southhouse, that would mean you're all on your own. 01:18.000 --> 01:23.000 You're on little server with the 5, 10 or 5 million users, in some cases, 01:23.000 --> 01:25.000 that you have on the server. 01:25.000 --> 01:26.000 But thanks for Federation. 01:26.000 --> 01:29.000 Similar to what E-mail does, and also must've done these days, 01:29.000 --> 01:31.000 and a bunch of other tools. 01:31.000 --> 01:35.000 You can work between people from different next-out servers. 01:35.000 --> 01:37.000 So, you can share files to people on the other server. 01:37.000 --> 01:40.000 You can have a video call with people from another server, 01:40.000 --> 01:45.000 or a chat with people from another server, and all feels like a single installation. 01:45.000 --> 01:46.000 So, there's pretty neat. 01:47.000 --> 01:51.000 And the second thing, of course, security, which is always important 01:51.000 --> 01:53.000 with these kind of things. 01:53.000 --> 01:59.000 Next out has, well, the basic HCTPS, so when you go from your server to your client, 01:59.000 --> 02:02.000 I mean, you can say, well, it's a trade-off. 02:02.000 --> 02:06.000 There's still a performance impact, and sure there is, but I don't think anybody wants to 02:06.000 --> 02:09.000 center stuff without any protection anymore over the internet. 02:09.000 --> 02:13.000 So, find the amazing that that was still a thing 10 years ago. 02:13.000 --> 02:17.000 On top of that, you can encrypt your data on storage. 02:17.000 --> 02:21.000 In case you're, for example, afraid somebody comes in and steals your server at night, 02:21.000 --> 02:24.000 with Raspberry Pi that's relatively easy with a big server, 02:24.000 --> 02:25.000 a data center. 02:25.000 --> 02:28.000 It's a little bit less likely, but there's, of course, still a risk somebody breaks in. 02:28.000 --> 02:32.000 So, it's the data's encrypted at rest that works when they break in. 02:32.000 --> 02:36.000 But, of course, if they're in, you don't notice, 02:36.000 --> 02:39.000 and they can observe what's happening after a while, 02:39.000 --> 02:41.000 encryption at rest is not good enough. 02:41.000 --> 02:45.000 I mean, it's an encryption for that, which means encrypting on the client. 02:45.000 --> 02:46.000 Turn what off. 02:46.000 --> 02:49.000 There's still one, there's a case you can do that. 02:49.000 --> 02:53.000 Oh, sorry. Please do. 02:53.000 --> 03:01.000 So, if you're running an XOT server, and you have maybe a subset of data that you want to protect, 03:01.000 --> 03:05.000 even in case some of the breaks into the server, or maybe you don't trust the server admin, 03:05.000 --> 03:09.000 and actually you can run this at home, but you can also run this as a service provider, 03:09.000 --> 03:13.000 then you can use the antenna encryption, which basically encrypts the files on the client, 03:13.000 --> 03:19.000 mobile, or desktop, so that then on the server, they cannot be read at all. 03:19.000 --> 03:25.000 The keys are encrypted, also locally, and not, well, they're only encrypted sent to the server, 03:25.000 --> 03:29.000 and you have to exchange passwords manually between the clients, 03:29.000 --> 03:33.000 to be able to have everything only locally accessible. 03:33.000 --> 03:37.000 Anyhow, what are we doing, all this effort for? 03:37.000 --> 03:41.000 Of course, because we don't want to live in a world where these companies, 03:41.000 --> 03:43.000 basically, control our entire digital lives. 03:43.000 --> 03:49.000 I heard yesterday, people saying that Europe is now a digital colony. 03:49.000 --> 03:53.000 It's not a really nice thing to say, but the Broly Garkey is definitely in charge, 03:53.000 --> 03:55.000 I would say, that's just a reality. 03:55.000 --> 04:01.000 If an American president would want to shut down the European economy, 04:01.000 --> 04:06.000 because Europe doesn't, I don't know, lower its tariffs, or stop selling cars, 04:06.000 --> 04:08.000 or whatever it is, they won't. 04:08.000 --> 04:10.000 They can actually do that. 04:10.000 --> 04:14.000 And they might not do it, but that will then be because Europe gave in, because of the threat. 04:14.000 --> 04:18.000 And that's, I think, not a healthy situation to be in as Europe. 04:18.000 --> 04:22.000 And obviously, on the personal level, I don't want these guys to have my data period, 04:22.000 --> 04:24.000 so that's that as well. 04:24.000 --> 04:28.000 So me, yours, was already mentioned, so I'm just going to move on here. 04:28.000 --> 04:32.000 And the thing is, I'm going to talk about the releases we did last year, 04:32.000 --> 04:34.000 and you know, we always do these release videos. 04:34.000 --> 04:36.000 They are one and a half, two hours long. 04:36.000 --> 04:38.000 We did about three releases. 04:38.000 --> 04:39.000 You can do the math. 04:39.000 --> 04:42.000 I have to do this in 30 minutes, so I'm going to go pretty quick. 04:42.000 --> 04:44.000 And I'm skipping a ton of stuff. 04:44.000 --> 04:48.000 So if you say at some point, like, hey, I'd like to learn more about XY or Z, 04:48.000 --> 04:52.000 you can also just wave at me, and if I'm paying attention, which isn't always a given, 04:52.000 --> 04:55.000 I will try and, you know, help you out there. 04:55.000 --> 04:56.000 All right. 04:56.000 --> 05:02.000 So I have tried to group it a little bit, so that everything is grouped by feature, 05:02.000 --> 05:04.000 not so much by release. 05:04.000 --> 05:07.000 But I'm still going to start with things from XY8, 05:07.000 --> 05:10.000 and then mix in XY9. 05:10.000 --> 05:13.000 And of course, there'll be also our special talk release we did. 05:13.000 --> 05:15.000 But let's start with XY8 files. 05:15.000 --> 05:17.000 So that's the file sink and share. 05:17.000 --> 05:19.000 With a desktop client, mobile clients, 05:19.000 --> 05:21.000 the document photos, videos live there. 05:21.000 --> 05:24.000 Yeah, easy to use. 05:24.000 --> 05:27.000 Of course, if you have, well, normal user these days, 05:27.000 --> 05:30.000 have thousands, tens of thousands, hundreds of thousands of files, 05:30.000 --> 05:32.000 that can get a little overwhelming. 05:32.000 --> 05:34.000 The next set has a ton of ways to help you. 05:34.000 --> 05:36.000 For example, we will recommend some files. 05:36.000 --> 05:40.000 For example, that we're recently edited or recently shared with you, 05:40.000 --> 05:43.000 just to help you find these files quicker. 05:43.000 --> 05:46.000 We keep track of relations between data. 05:46.000 --> 05:49.000 So if you have files that you're working on with a group of people, 05:49.000 --> 05:52.000 then when you click on one of these files, you will see the other files. 05:52.000 --> 05:57.000 But you will also see a chatroom, you will see maybe a task 05:57.000 --> 05:59.000 that is related with the files, et cetera. 05:59.000 --> 06:03.000 So the different content between the different next cloud components 06:03.000 --> 06:05.000 will all then be visible there. 06:05.000 --> 06:08.000 And you can say, oh, you know, we have a calendar together. 06:08.000 --> 06:10.000 We have a meeting room together. 06:10.000 --> 06:13.000 And that is all that related to these files. 06:13.000 --> 06:16.000 In the left side bar, you have your favorites. 06:16.000 --> 06:19.000 You recently changed files, another quick filters. 06:19.000 --> 06:23.000 And new and up eight was the personal file field that just sees the files 06:23.000 --> 06:26.000 that are completely yours and not shared by others. 06:26.000 --> 06:28.000 Because, you know, especially if you live in the company, 06:28.000 --> 06:31.000 you have tons of data that is actually shared by colleagues 06:31.000 --> 06:33.000 and not really yours only. 06:33.000 --> 06:36.000 It can sometimes be helpful to quickly see that stuff. 06:36.000 --> 06:38.000 Now, there's, of course, another way to find files. 06:38.000 --> 06:39.000 That's a search. 06:39.000 --> 06:44.000 And you can there find, again, content from all over next cloud 06:44.000 --> 06:47.000 through emails, chat messages, well, and your files. 06:47.000 --> 06:50.000 And of course, those files might even be shared in the chat room. 06:50.000 --> 06:53.000 And next, let's have a video chat application, right? 06:53.000 --> 06:54.000 Next, I'll talk. 06:54.000 --> 06:57.000 And that's deeply integrated with the rest of next cloud. 06:57.000 --> 07:01.000 So if you share a document, you can just edit the document and view it right there 07:01.000 --> 07:02.000 in the chat. 07:02.000 --> 07:06.000 There's no need to navigate away open a tab or download it and edit it on your desktop. 07:06.000 --> 07:08.000 You can just go and work on your files. 07:08.000 --> 07:10.000 And this also works during a video call. 07:10.000 --> 07:13.000 So it can be, well, a quick way to edit something. 07:13.000 --> 07:19.000 Obviously, this is not where you write your presentations or big documents. 07:19.000 --> 07:23.000 But for that, you can open it and separate tab as well. 07:23.000 --> 07:25.000 You can also insert files in deck. 07:25.000 --> 07:26.000 Is it playing? 07:26.000 --> 07:27.000 No, it is not. 07:27.000 --> 07:29.000 So you can insert all kinds of content. 07:29.000 --> 07:31.000 It is our lightweight project management app. 07:31.000 --> 07:34.000 And again, you can just insert the document there. 07:34.000 --> 07:36.000 And you can edit it right in there. 07:36.000 --> 07:39.000 You can choose to view it as a link or you want to see the full file. 07:39.000 --> 07:44.000 You can scroll and see the button and then go and edit away. 07:44.000 --> 07:49.000 So this tool that you just saw the slash and then it gives you a bunch of choices. 07:49.000 --> 07:51.000 This lets you also do other stuff. 07:51.000 --> 07:56.000 Besides inserting files, you can insert the link to a location or a street map. 07:56.000 --> 07:57.000 You can insert an image. 07:57.000 --> 07:58.000 You can insert a video. 07:58.000 --> 07:59.000 You can insert tasks. 07:59.000 --> 08:00.000 You can insert calendars. 08:00.000 --> 08:02.000 You can insert chat rooms. 08:02.000 --> 08:05.000 Basically all kinds of content from all over next cloud. 08:05.000 --> 08:07.000 It can be inserted in any of the places. 08:07.000 --> 08:09.000 So this works in chat. 08:09.000 --> 08:10.000 This works in the task. 08:10.000 --> 08:11.000 This works. 08:11.000 --> 08:12.000 Basically anywhere. 08:12.000 --> 08:13.000 And that's all. 08:13.000 --> 08:15.000 This is our notes app. 08:15.000 --> 08:19.000 But again, when you have text input fields, you can put all kinds of content there. 08:19.000 --> 08:21.000 That works all over next cloud. 08:21.000 --> 08:25.000 Now sharing, of course, a pretty core part of next cloud files. 08:25.000 --> 08:28.000 When you share, you have a bunch of options. 08:28.000 --> 08:29.000 You can add a note. 08:29.000 --> 08:31.000 You can hide the download option. 08:31.000 --> 08:32.000 You can limit downloads. 08:32.000 --> 08:34.000 It was new in Hub8. 08:34.000 --> 08:37.000 You can also show a QR code. 08:37.000 --> 08:41.000 I never really understand why you want to have it on your screen. 08:41.000 --> 08:42.000 If you want to share it. 08:42.000 --> 08:45.000 But I'm guessing you can send the QR code to people. 08:45.000 --> 08:46.000 Or print it. 08:46.000 --> 08:50.000 Now, these were all things that changed in Hub8. 08:50.000 --> 08:52.000 And I promised that I would mix in Hub9. 08:52.000 --> 08:55.000 But the thing is Hub9 made design chains. 08:55.000 --> 08:57.000 So it would have been quite visible when it was Hub9. 08:57.000 --> 08:59.000 So I thought that first show that design. 08:59.000 --> 09:02.000 So with Hub9, we had a new background. 09:03.000 --> 09:06.000 And we changed button and border radius, basically, 09:06.000 --> 09:09.000 to take a little bit less space, make the interface more efficient. 09:09.000 --> 09:12.000 And that denser interface is very visible, especially in talk. 09:12.000 --> 09:18.000 Because you can noticeably add more chats in the room and have so more messages visible. 09:18.000 --> 09:20.000 Yeah. 09:20.000 --> 09:23.000 This is an exit files where we also updated a lot of icons. 09:23.000 --> 09:26.000 I made a sidebar a lot more compact. 09:26.000 --> 09:29.000 And the breadcrumbs is more compact, etc. 09:29.000 --> 09:34.000 So I think this is nice for people who like to have a little bit of a denser interface. 09:34.000 --> 09:36.000 Oh, that was an exit mail. 09:36.000 --> 09:38.000 And I'm going a little quick. 09:38.000 --> 09:42.000 So there everything closer together in collectives and knowledge management app. 09:42.000 --> 09:45.000 Also improve the layout of the pages a bit. 09:45.000 --> 09:50.000 And of course, for those with a dark soul, like me, a preferred dark mode. 09:50.000 --> 09:54.000 Then, well, this is of course also available in the next slide. 09:54.000 --> 09:55.000 All right. 09:55.000 --> 09:59.000 One of the new things that we did are filters. 09:59.000 --> 10:02.000 I, yes, exactly, it does play. 10:02.000 --> 10:04.000 So you can quickly filter for specific file types. 10:04.000 --> 10:06.000 There's also on the top left I search. 10:06.000 --> 10:08.000 So you can also search, of course, by the words. 10:08.000 --> 10:11.000 And find a file that you're looking for specifically. 10:11.000 --> 10:15.000 A lot of people have also been asking about a folder preview on the sidebar, 10:15.000 --> 10:17.000 which we introduce now. 10:17.000 --> 10:20.000 So you can navigate your files a folder faster. 10:21.000 --> 10:22.000 Another cool thing. 10:22.000 --> 10:25.000 You can also upload folders, not just files. 10:25.000 --> 10:29.000 And we added a folder quest feature. 10:29.000 --> 10:34.000 You could, of course, always send a folder to people where they can upload files to. 10:34.000 --> 10:36.000 But with a further quest. 10:36.000 --> 10:37.000 Come on. 10:37.000 --> 10:38.000 Go play. 10:38.000 --> 10:43.000 We created a nice dialogue where, well, a non technical person who find this whole, like, 10:43.000 --> 10:47.000 create a public link and then make it possible to upload files, etc. 10:47.000 --> 10:48.000 Who finds it complicated. 10:48.000 --> 10:51.000 You can just answer a couple of simple questions. 10:51.000 --> 10:53.000 Who do you want to send it to given email address. 10:53.000 --> 10:58.000 And then they will get an email with a nice request that says, please upload the files here. 10:58.000 --> 11:03.000 And then later, of course, you will get notified when the files are uploaded, etc. 11:03.000 --> 11:06.000 It just makes the same thing a little easier to use. 11:06.000 --> 11:08.000 Tons of other improvements. 11:08.000 --> 11:10.000 But I don't have time for that. 11:10.000 --> 11:12.000 I need to get a group error. 11:12.000 --> 11:16.000 So mail app is, of course, designed to make your life easier. 11:16.000 --> 11:17.000 So you can do things, etc. 11:17.000 --> 11:18.000 A minor for a mail. 11:18.000 --> 11:21.000 We have a AI inbox. 11:21.000 --> 11:25.000 We actually had that already four years ago or something. 11:25.000 --> 11:31.000 Before the LLM-Crisse, this is just a locally running neural network that gets trained on 11:31.000 --> 11:38.000 your emails and what you click and what subjects you like or people you mail with. 11:38.000 --> 11:41.000 But we do use the LLM-Crisse as well. 11:41.000 --> 11:42.000 So you can summarize emails. 11:42.000 --> 11:44.000 It will put little summaries in there. 11:44.000 --> 11:47.000 If you have a long email thread, it will summarize it. 11:47.000 --> 11:51.000 It can suggest replies, you know, and we have a pretty cool feature. 11:51.000 --> 11:53.000 And I think that's in the next slide. 11:53.000 --> 11:55.000 I'll wait with that. 11:55.000 --> 11:58.000 We have, as my signatures and encryption, all this stuff. 11:58.000 --> 12:00.000 And shared mailboxes as well. 12:00.000 --> 12:02.000 Now, yeah. 12:02.000 --> 12:08.000 So one cool thing we did for hub nine is that we created an urgent thing on top, which 12:08.000 --> 12:13.000 the LLM actually reads the email and decides if it might be urgent or not. 12:13.000 --> 12:19.000 Again, this stuff has a perfect, as lots of online videos about all kinds of stuff show. 12:19.000 --> 12:22.000 But it can help you, you know, make sure. 12:22.000 --> 12:24.000 And this also helps you with reminders. 12:24.000 --> 12:27.000 So it basically, when it sees like, oh, get back to me next week, 12:27.000 --> 12:30.000 then the week after this email has put on top of your inbox. 12:30.000 --> 12:36.000 So it'll automatically decide when emails need to need to follow up. 12:36.000 --> 12:39.000 While, of course, if the email is just like, okay, thank you. 12:39.000 --> 12:41.000 Then you don't need to follow up. 12:41.000 --> 12:44.000 So it basically filters out the emails, what you need to do something with. 12:44.000 --> 12:50.000 Rather than emails that are maybe, yeah, just for your information or newsletter, etc. 12:50.000 --> 12:53.000 All right. 12:53.000 --> 12:55.000 Well, we're at the calendar app. 12:55.000 --> 12:59.000 And this is connected with NexatTalk. 12:59.000 --> 13:04.000 So you can, for example, insert calendar invitations and also see them in talk. 13:04.000 --> 13:10.000 Another cool thing is that if you create a calendar invite in the calendar app, 13:10.000 --> 13:15.000 you can add a talk room and then in the talk room, you can see when the upcoming call is, for example. 13:15.000 --> 13:17.000 So this is also nicely integrated. 13:17.000 --> 13:21.000 But this is where you can share, you know, schedules and stuff. 13:21.000 --> 13:25.000 We improved the filtering and search email a lot. 13:25.000 --> 13:28.000 It now supports C filters. 13:28.000 --> 13:31.000 So that means the filters actually run on the server instead of on the client, 13:31.000 --> 13:36.000 which means if you use multiple email clients, including the Nexat webmail client, 13:36.000 --> 13:41.000 you will have the same filters everywhere because they're processed on the server automatically. 13:41.000 --> 13:44.000 The mail server itself. 13:44.000 --> 13:47.000 Well, back to the calendar again. 13:47.000 --> 13:51.000 We worked a lot on the availability few resource booking that went a little too quick. 13:51.000 --> 13:56.000 Room search and so you can put in resources and then book a room or a book, you know, 13:56.000 --> 14:00.000 a projector and these kind of things quite nice in the company. 14:00.000 --> 14:04.000 And I already mentioned NexatTalk. You can see upcoming meetings. 14:05.000 --> 14:08.000 With Hub8, we included a global out of office. 14:08.000 --> 14:10.000 So there's a fact, of course, your email. 14:10.000 --> 14:15.000 But people who try to chat with you also will see a message that you're out of office until a certain date. 14:15.000 --> 14:20.000 And if you try to book a calendar meeting with you, then the calendar will also say you're out of office. 14:20.000 --> 14:23.000 So you don't need to put a blocker in, do your calendar manually. 14:23.000 --> 14:26.000 And your status also shows that you're out of office. 14:26.000 --> 14:29.000 So you only need to set this once. 14:29.000 --> 14:33.000 There's a ton more that happened, and I really couldn't make it very readable. 14:33.000 --> 14:37.000 But there a lot of work that was done last year in group air. 14:37.000 --> 14:40.000 Office, I'm going to keep even shorter than the other stuff. 14:40.000 --> 14:42.000 We developed this together with Klebera. 14:42.000 --> 14:46.000 Michael Meeks is not in the room, but at some point I'm sure you can follow the talker. 14:46.000 --> 14:49.000 He talks about a lot of the improvements they did. 14:49.000 --> 14:53.000 Of course, a number of things is done specifically for NexCloud. 14:53.000 --> 14:56.000 The general performance and security work, etc. 14:56.000 --> 14:58.000 You can then get for Michael Meeks. 14:58.000 --> 15:01.000 One cool thing that we did in NexCloud itself is support for filling forms. 15:01.000 --> 15:06.000 So if you have a document with forms, you can use the NexCloud interface to fill in the form. 15:06.000 --> 15:09.000 And then you get the document, you can then sell mic changes. 15:09.000 --> 15:14.000 Or you just got a PDF out of it that you can immediately send out to other people. 15:14.000 --> 15:17.000 So that kind of integration is quite nice. 15:17.000 --> 15:19.000 Now, I already talked a little bit about talk. 15:19.000 --> 15:22.000 I'm going to dive into it a little deeper now. 15:22.000 --> 15:25.000 So this is for online meetings, video conferencing and chat. 15:26.000 --> 15:29.000 It's effectively Zoom and Slack into one. 15:29.000 --> 15:30.000 So think of mics of teams. 15:30.000 --> 15:33.000 That's effectively, of course, what it aims to replace. 15:33.000 --> 15:37.000 Scales to hundreds of participants, a call with 500 people. 15:37.000 --> 15:39.000 That's a reasonable amount of problem. 15:39.000 --> 15:43.000 You have mobile interfaces, Android, iOS, desktop interface. 15:43.000 --> 15:45.000 Feet is like breakout rooms. 15:45.000 --> 15:48.000 Federation that I mentioned earlier is available for talk as well. 15:48.000 --> 15:52.000 Yeah, essentially complete mics of team replacement. 15:52.000 --> 15:55.000 We've had nine interviews tonight's whiteboard. 15:55.000 --> 15:58.000 We in December are desktop clients became 1.0. 15:58.000 --> 16:04.000 So that is now officially shouldn't just be playing officially stable. 16:04.000 --> 16:06.000 A ton of other new features. 16:06.000 --> 16:09.000 For example, you can now save a poll as a draft. 16:09.000 --> 16:11.000 You can even export it as a file. 16:11.000 --> 16:15.000 So if you do webinars a lot, my team does a lot of webinars. 16:15.000 --> 16:18.000 And then you can create polls, save them into a file. 16:18.000 --> 16:21.000 And then when you create a new webinar, you just upload the file. 16:21.000 --> 16:25.000 With the polls and then you can, you know, you have them all the drafts. 16:25.000 --> 16:27.000 And you can just one by one just click. 16:27.000 --> 16:28.000 Publish publish publish. 16:28.000 --> 16:30.000 You don't need to manually work on that. 16:30.000 --> 16:33.000 So that kind of stuff is quite nice, I think. 16:33.000 --> 16:37.000 You can also invite people by mail or by public links. 16:37.000 --> 16:40.000 You can send people a public link and they can share it with others. 16:40.000 --> 16:46.000 But if you invite them by mail, you can see that that person with that specific email address actually joined. 16:46.000 --> 16:50.000 So again, for webinars, you want to track which of the people that got invited. 16:50.000 --> 16:51.000 Join. 16:51.000 --> 16:53.000 So you can upload the spreadsheet with email addresses. 16:53.000 --> 16:55.000 And then all these people will get a unique link. 16:55.000 --> 17:01.000 And then when they're joined, you can see you can also in advance already make some of the moderator for example. 17:01.000 --> 17:04.000 I have their speaker, if they're a participant. 17:04.000 --> 17:07.000 You can then set the permissions for that already. 17:07.000 --> 17:11.000 Yeah, and then later you can export the people who joined. 17:11.000 --> 17:17.000 So that you can see from the people who you invited, which ones showed up. 17:17.000 --> 17:22.000 Now, thinking about it, I could have put this in the section completely at the beginning. 17:22.000 --> 17:24.000 I already mentioned a bunch of LLM things. 17:24.000 --> 17:29.000 There's quite a bit of AI features in the Naround next slide that we've been implementing. 17:29.000 --> 17:35.000 But of course, you know, we're making an on-prem open source self hosted privacy solution. 17:35.000 --> 17:41.000 And so integrating said GPT was an exactly like top on our agenda as you can imagine. 17:41.000 --> 17:48.000 So instead, we talked actually at the first time a couple of years ago about AI and where we want to go. 17:48.000 --> 17:52.000 Because on one hand, it can be helpful for some things, especially summaries. 17:52.000 --> 17:55.000 It doesn't hallucinate too much in summaries. 17:55.000 --> 17:58.000 And we thought, well, we want to bring those benefits to users. 17:58.000 --> 18:01.000 But again, we don't want to give up on their privacy. 18:01.000 --> 18:04.000 But the thing is, of course, you don't have a lot of options with everything. 18:04.000 --> 18:10.000 I have for speeds to text, there are quite decent on-premises models. 18:10.000 --> 18:17.000 And already four years ago, we integrated recognized, which will automatically tag your photos with faces, 18:17.000 --> 18:19.000 and with objects so you can search through them. 18:19.000 --> 18:24.000 And there was a completely on-premise running model. 18:24.000 --> 18:28.000 It was a gigabyte, it was pretty big, but it worked on a CPU. 18:28.000 --> 18:30.000 And so, that was nice. 18:30.000 --> 18:36.000 But when all the LLM came out, we were like, okay, we need to think about how we're going to do this. 18:36.000 --> 18:40.000 So, we came up with what we call our ethical AI approach. 18:40.000 --> 18:42.000 It's fairly simple. 18:42.000 --> 18:46.000 The idea is that each of the AI options that you have, 18:46.000 --> 18:51.000 I think of CHGPT or locally running Lama, et cetera, and everything in between, 18:51.000 --> 18:56.000 gets a rating, and the rating goes from all, rather green. 18:56.000 --> 19:00.000 And it depends on these three factors. 19:00.000 --> 19:05.000 So first is the code that you need to run the LLM and to train the LLM open source. 19:05.000 --> 19:11.000 Second, if the model is available, that means you can run it on-premises, if these two are true. 19:11.000 --> 19:15.000 And then last but not least, is a training data completely open and available. 19:15.000 --> 19:20.000 So if all these three conditions are met, the AI solution is green. 19:20.000 --> 19:26.000 If one of them is not met, orange and fury can figure out the rest, where it goes from there. 19:26.000 --> 19:32.000 So the idea from us is to bring transparency and to give you as user a choice. 19:33.000 --> 19:36.000 So this choice is quite literal. 19:36.000 --> 19:41.000 In the interface, you will be able to choose which AI model, if you have a bunch on the locally, 19:41.000 --> 19:44.000 or which remote service, et cetera, you want to use. 19:44.000 --> 19:47.000 And then in our app store, you see the rating. 19:47.000 --> 19:51.000 We try to make sure that they're all transparent there, and you can then choose. 19:51.000 --> 19:57.000 Yeah, if you have a big BFGPU and you have a nice local model, you can use that. 19:57.000 --> 20:01.000 And if you're not handling any critical data and you're happy sponsoring Microsoft, 20:01.000 --> 20:04.000 you can send your data to chatGPT. 20:04.000 --> 20:06.000 So this way, there's choice. 20:06.000 --> 20:08.000 So we integrate in tonal place. 20:08.000 --> 20:12.000 I already showed you a few things in our summaries in mail, for example, 20:12.000 --> 20:14.000 but there are lots of other places. 20:14.000 --> 20:17.000 We already were using neural network things, 20:17.000 --> 20:20.000 or experimenting with that, as I mentioned, 20:20.000 --> 20:24.000 our smart inbox that is already downloaded before the LLM crash. 20:24.000 --> 20:26.000 We also did a suspicious log and detection, 20:26.000 --> 20:30.000 so we're training on your server neural network on your logins. 20:30.000 --> 20:33.000 And if you normally log in from 9 to 5 from the office, 20:33.000 --> 20:34.000 and suddenly, there's a tree. 20:34.000 --> 20:38.000 A.M. in the morning, a log in from another continent, 20:38.000 --> 20:41.000 then you will get a warning, and maybe we will block the log in, 20:41.000 --> 20:44.000 depending on your settings. 20:44.000 --> 20:47.000 Cold transcripts, so if you have a call in Excel talk, 20:47.000 --> 20:50.000 you can have it if you have the recording on. 20:50.000 --> 20:53.000 After the call, you will get a transcript, 20:53.000 --> 20:57.000 you will even get a summary that there's new in Excel to have nine. 20:57.000 --> 21:00.000 And then, yeah, we'll save you some time, of course. 21:00.000 --> 21:03.000 And you can search through your calls, basically, this way. 21:03.000 --> 21:05.000 And there's, of course, a text generation, 21:05.000 --> 21:11.000 you know, translate translation, dictation, all these features. 21:11.000 --> 21:16.000 So, this is all under the name of the next other system, 21:16.000 --> 21:18.000 also the old features. 21:18.000 --> 21:20.000 And the next other system is, well, obviously, 21:20.000 --> 21:23.000 then a large language model, based mostly, 21:23.000 --> 21:26.000 at least 100% open source and on-premise, 21:26.000 --> 21:27.000 so no data sent. 21:27.000 --> 21:29.000 If you use the local LLM, as I said, 21:29.000 --> 21:32.000 this is a choice, this is transparent. 21:32.000 --> 21:35.000 Last year, we basically introduced version 2.0, 21:35.000 --> 21:38.000 so that came with a complete overhaul of the interface. 21:38.000 --> 21:42.000 A lot of work on the back end, so you can now have a separate server, 21:42.000 --> 21:45.000 dedicated for AI, so maybe with a GPU, right? 21:45.000 --> 21:48.000 So, you split it up from your normal application server. 21:48.000 --> 21:52.000 The difference and performance between a GPU or a CPU for AI is, 21:52.000 --> 21:55.000 well, I'm sure I don't need to tell most of you, 21:55.000 --> 21:58.000 it's absolutely insane, so this is quite helpful. 21:58.000 --> 22:03.000 Yeah, they came then with the summarizing of email transfer example. 22:03.000 --> 22:07.000 You also introduced two features, context, right, and context yet. 22:07.000 --> 22:12.000 The context, right, allows you to basically give the LLM an example document. 22:12.000 --> 22:16.000 You can say, like, hey, here's an environmental responsibility report. 22:16.000 --> 22:20.000 If you want to write one yourself for your organization, 22:20.000 --> 22:23.000 you give it to your responsibility report from another organization, 22:23.000 --> 22:26.000 then you give it the notes from a meeting that you had to discuss, 22:26.000 --> 22:29.000 the state of environmental policies and your organization, 22:29.000 --> 22:33.000 and you say, use the meeting notes to write a report similar to the other one, 22:33.000 --> 22:36.000 and you have a pretty good start to get going. 22:36.000 --> 22:38.000 The second thing we did is context yet. 22:38.000 --> 22:41.000 So, cause I can write right for you, context yet, 22:41.000 --> 22:44.000 allows you to ask questions about your own data. 22:44.000 --> 22:48.000 So, basically, all your documents, all your chats, all your data, 22:48.000 --> 22:51.000 in-next loud, is put into a vector database, 22:51.000 --> 22:54.000 and you can ask questions, you can say, like, hey, you know, 22:54.000 --> 22:58.000 what is our internal process for organizing an event, 22:58.000 --> 23:02.000 and it will look through your documentation and make a summary of that. 23:02.000 --> 23:07.000 Yeah, another thing, it's not only able to look through your documents, 23:07.000 --> 23:10.000 it's also able to look for data in other applications. 23:10.000 --> 23:14.000 So, for example, the next other application, called analytics, 23:14.000 --> 23:19.000 and in there, you can, while either via spreadsheet or via an API, 23:19.000 --> 23:21.000 upload all kinds of analytical data. 23:21.000 --> 23:24.000 For example, whether data, if you're tracking whether data, 23:24.000 --> 23:27.000 for you at home, you can hook this into the analytics app. 23:27.000 --> 23:31.000 And the nice thing is, the analytics app can then let this get indexed 23:31.000 --> 23:33.000 by the vector database. 23:33.000 --> 23:36.000 So, you can then go and actually ask a question from the LLM, 23:36.000 --> 23:40.000 and you can say to the LLM, like, hey, what was the warmest month last year, 23:40.000 --> 23:42.000 and then look in the data and answer your question. 23:42.000 --> 23:45.000 So, that's pretty neat. 23:45.000 --> 23:50.000 Also available on your phone, as I already showed twice now. 23:50.000 --> 23:55.000 So, for the last 20 minutes, I've been showing you all kinds of features, 23:55.000 --> 23:57.000 and abilities of next loud. 23:57.000 --> 24:01.000 Everything, you know, connected from apps in next other itself. 24:01.000 --> 24:05.000 Of course, we're a very big ecosystem with lots of other applications 24:05.000 --> 24:07.000 that we are integrated with. 24:07.000 --> 24:10.000 So, if you say, hey, the next I talk, so fine, 24:10.000 --> 24:13.000 I don't need to be able to edit documents in the chat. 24:13.000 --> 24:19.000 I'm using element, for example, or, you know, hack. 24:19.000 --> 24:23.000 Maybe you're using a proprietary solution, like WebEx, for example. 24:23.000 --> 24:24.000 That's fine. 24:24.000 --> 24:26.000 You can connect this to next loud. 24:26.000 --> 24:29.000 We have tons of integrations with these other solutions. 24:29.000 --> 24:33.000 And always working with good friends, like XWiki, you know, project, 24:33.000 --> 24:37.000 and the intervention lots of other cool projects together to make these integrations 24:37.000 --> 24:39.000 better and make it easier for you. 24:39.000 --> 24:44.000 And, of course, because many of you are still stuck to this company, 24:44.000 --> 24:46.000 we also have integrations with Microsoft. 24:46.000 --> 24:51.000 I hope that this helps you migrate away from it, but still, 24:51.000 --> 24:54.000 it is necessary in many cases. 24:54.000 --> 24:59.000 So, the apps for next loud can be built as part of the core. 24:59.000 --> 25:06.000 Next loud is written in PHP with components in Rust and go in other pieces. 25:06.000 --> 25:09.000 But you don't have to build an XWTP. 25:09.000 --> 25:13.000 You can build an XWTP and go in Ruby and Python, whatever language you want. 25:13.000 --> 25:16.000 These apps would then run as part of a Microsoft's architecture. 25:16.000 --> 25:21.000 They use Docker deployment, so you get distributed compute. 25:21.000 --> 25:24.000 And, of course, this also can be a security benefit. 25:24.000 --> 25:29.000 So, if you want to build an XWTP in other languages, this is possible as well. 25:29.000 --> 25:33.000 Now, last year, we introduced one more big thing. 25:33.000 --> 25:36.000 I want to talk about it for a minute, and that's an XWTP flow. 25:36.000 --> 25:41.000 Next up is essentially a business process automation engine. 25:41.000 --> 25:44.000 Honestly, quite boring for a home user. 25:44.000 --> 25:47.000 We have some automation features for private users too. 25:47.000 --> 25:50.000 But for businesses, this can be quite helpful. 25:50.000 --> 25:54.000 It can do a ton of things like connecting, basically, 25:54.000 --> 25:58.000 the different functions between next loud and executing actions on it. 25:58.000 --> 26:04.000 So, the reason we developed this was in Germany, like in many other countries, 26:04.000 --> 26:08.000 they're trying to automate processes as government. 26:08.000 --> 26:11.000 So, that, for example, if you want to request a password, 26:11.000 --> 26:14.000 sorry, passports, not a password. 26:14.000 --> 26:20.000 That is handled in an efficient way at the government. 26:20.000 --> 26:25.000 And to automate these processes, to be able to request this via the internet 26:25.000 --> 26:30.000 rather than having to physically go to an office and give all your papers in current passports, 26:30.000 --> 26:31.000 etc. 26:31.000 --> 26:34.000 Well, they need to make a lot of changes in improvements. 26:34.000 --> 26:36.000 And Germany is really lagging behind in this. 26:36.000 --> 26:39.000 So, are several other European countries. 26:39.000 --> 26:44.000 And all the automation tools for this are proprietary. 26:44.000 --> 26:47.000 So, basically, as your government, if you want to automate these things, 26:47.000 --> 26:51.000 you either have to build something yourself or you buy from a proprietary vendor. 26:51.000 --> 26:52.000 And that's not great. 26:52.000 --> 26:56.000 So, we thought, well, if you already have a ton of your data in next loud, 26:56.000 --> 27:01.000 wouldn't be nice if you can then use that data, just directly from within next loud 27:01.000 --> 27:03.000 to automate these processes. 27:03.000 --> 27:07.000 And that is what the integration of flow was designed to help with. 27:07.000 --> 27:11.000 So, you can connect pieces from next loud. 27:11.000 --> 27:13.000 I like, I don't know, in next, that forms app. 27:13.000 --> 27:15.000 People fill in the form. 27:15.000 --> 27:18.000 And then that goes into the next autables app. 27:18.000 --> 27:22.000 Maybe your notification is sent to somebody to check if, you know, the content of the 27:22.000 --> 27:23.000 form is correct. 27:23.000 --> 27:25.000 You will touch a file, etc. 27:25.000 --> 27:28.000 It can also handle, like, an incoming email. 27:28.000 --> 27:33.000 If there's PDF attached to that email, it can take the PDF parse it. 27:33.000 --> 27:35.000 Put the content in the table. 27:35.000 --> 27:37.000 Let the human double check it. 27:37.000 --> 27:41.000 You can even run it through an LLM for certain operations if you want. 27:41.000 --> 27:43.000 I would then definitely let the human check it. 27:43.000 --> 27:47.000 And then, for example, send an email back or send a ping to a payment system. 27:47.000 --> 27:52.000 And then handle the payment generated PDF and set it out again. 27:52.000 --> 27:56.000 So, all these things are possible within next loud, using flow. 27:56.000 --> 27:57.000 All connected. 27:57.000 --> 28:01.000 Do you have enough government sets and nice things about that? 28:01.000 --> 28:03.000 Which I think is really cool. 28:03.000 --> 28:08.000 And I'm going to close it here so that you also have some time for questions if there are any. 28:08.000 --> 28:12.000 Of course, all of this, we're 130 people at next loud. 28:12.000 --> 28:16.000 That's just a small part of our entire community of thousands of people. 28:16.000 --> 28:20.000 So, for those of you who are contributing to next loud one way or another. 28:20.000 --> 28:21.000 Thank you. 28:21.000 --> 28:23.000 You might all of this stuff possible. 28:23.000 --> 28:26.000 In our next release, which is coming quite soon, 28:26.000 --> 28:29.000 there's really a ton of contributions, including some features. 28:29.000 --> 28:32.000 I'm personally very happy with from community members. 28:32.000 --> 28:38.000 So, this is not just a community, but actually just a company doing itself. 28:38.000 --> 28:40.000 And throwing the code over the wall occasionally. 28:40.000 --> 28:44.000 We're a bottom up community where community members have real impact. 28:44.000 --> 28:46.000 And personally, that makes me quite happy. 28:46.000 --> 28:50.000 I think it also shows from the full bedroom every year here. 28:50.000 --> 28:52.000 So, yeah, thank you. 28:52.000 --> 28:54.000 Come to our conference, by the way. 28:54.000 --> 28:55.000 We will announce the dates. 28:55.000 --> 28:57.000 Hopefully, the coming week. 28:57.000 --> 29:00.000 It'll be in September, in Berlin. 29:00.000 --> 29:02.000 And yeah, we'd be really cool to see you there. 29:02.000 --> 29:07.000 And of course, as I mentioned, soon, this is coming in just a few weeks. 29:07.000 --> 29:10.000 And there'll be a ton of good stuff there as well. 29:11.000 --> 29:13.000 That's the end of the show. 29:13.000 --> 29:15.000 Any questions? 29:15.000 --> 29:16.000 Thank you. 29:19.000 --> 29:22.000 I don't think that they do. 29:22.000 --> 29:25.000 So, now I can repeat the question. 29:25.000 --> 29:27.000 I think that's easiest. 29:27.000 --> 29:28.000 So, yes. 29:36.000 --> 29:37.000 All right. 29:37.000 --> 29:38.000 Yes, question. 29:40.000 --> 29:47.000 So, the question is, can you talk a little bit more about the technical part of the Federation 29:47.000 --> 29:48.000 feature? 29:48.000 --> 29:50.000 Yeah, a tiny bit. 29:50.000 --> 29:54.000 But I think the techie is, from our team, can go a little deeper. 29:54.000 --> 29:57.000 But basically, it depends on what you're federating. 29:57.000 --> 30:01.000 So, I think it's a little bit more about the technical part of the Federation 30:01.000 --> 30:02.000 feature. 30:02.000 --> 30:03.000 Yeah, a tiny bit. 30:03.000 --> 30:07.000 But I think the techie is, from our team, can go a little deeper. 30:07.000 --> 30:10.000 So, basically, it depends on what you're federating. 30:10.000 --> 30:15.000 So, for files, this is using WebDev and the Open Collaboration Services. 30:15.000 --> 30:18.000 So, that's an Open API. 30:18.000 --> 30:22.000 I think it also uses Open Cloud Mesh API. 30:22.000 --> 30:24.000 So, this is standardized. 30:24.000 --> 30:26.000 For next time, talk. 30:26.000 --> 30:30.000 It's effectively using also the Open Collaboration Services API, 30:30.000 --> 30:33.000 mounting the room in the other instance. 30:33.000 --> 30:36.000 So, you will get the room locally, but you're actually on the room of the other server. 30:36.000 --> 30:41.000 So, the way we do Federation, by principle, is both for the files, as well as for talk, 30:41.000 --> 30:46.000 is that the data stays on the owner's server as much as possible. 30:46.000 --> 30:49.000 So, when you share a file with Federation in X-Loud, 30:49.000 --> 30:51.000 the recipient gets access to the file. 30:51.000 --> 30:54.000 But the file is not sent to them and stored on their server. 30:54.000 --> 30:57.000 Only when they click on download, they don't sync it, et cetera. 30:57.000 --> 31:00.000 So, the file stays where it is, and the moment you un-share it, 31:00.000 --> 31:02.000 then they don't have access to it anymore. 31:02.000 --> 31:07.000 I think that's better modeling some ways for, 31:07.000 --> 31:10.000 at least from a privacy point of view, and a control point of view. 31:10.000 --> 31:13.000 Then, like, syncing all the data to all the servers, 31:13.000 --> 31:16.000 as a site from the overhead that causes, of course. 31:16.000 --> 31:19.000 It's the same with the chat room effectively. 31:19.000 --> 31:20.000 Yes? 31:20.000 --> 31:23.000 At the land, it's very, you know, health value condition. 31:23.000 --> 31:27.000 For details about the rat spiked of the next cloud, 31:27.000 --> 31:30.000 on modular reviews, and how it's built, 31:30.000 --> 31:32.000 and how to restrict knowledge of use, 31:32.000 --> 31:35.000 like databases, and kind of bugs. 31:35.000 --> 31:37.000 So, the LLM part of it. 31:37.000 --> 31:39.000 Well, we're pretty simple. 31:39.000 --> 31:42.000 So, the question is, can you share more details? 31:42.000 --> 31:43.000 Thank you. 31:43.000 --> 31:45.000 I have the memory of a squirrel. 31:45.000 --> 31:48.000 Can you share more details about, like, the AI, 31:48.000 --> 31:50.000 basically, the integration with that? 31:50.000 --> 31:53.000 So, yeah. 31:53.000 --> 31:56.000 We're essentially not, we're not building our own LLM, right? 31:56.000 --> 32:00.000 I think for the speech to text engine, 32:00.000 --> 32:05.000 we did build a model using some universe of Helsinki data. 32:05.000 --> 32:09.000 But, of course, an LLM running is not really a feasible thing. 32:09.000 --> 32:14.000 So, we're basically supporting pretty much all of the models out there. 32:14.000 --> 32:17.000 I don't know exactly how this works on the back end, 32:17.000 --> 32:21.000 but we're nearly any model, basically, that you will find on hugging phase. 32:21.000 --> 32:25.000 You can run locally, and who can to next cloud. 32:26.000 --> 32:28.000 I think the best answer I can give you. 32:28.000 --> 32:31.000 I'm sorry. 32:31.000 --> 32:33.000 Yeah. Next question. Yes? 32:33.000 --> 32:34.000 Yes. 32:34.000 --> 32:36.000 Welcome to your favorite, sir. 32:36.000 --> 32:37.000 I'm sorry. 32:37.000 --> 32:39.000 Ah, my favorite feature, I have done. 32:42.000 --> 32:43.000 All right. All right. 32:43.000 --> 32:44.000 All right. One hint. 32:44.000 --> 32:48.000 I use talk all day, because I'm, I'm, you know, 32:48.000 --> 32:50.000 chatting with my team, et cetera. 32:50.000 --> 32:52.000 And while the universal search is really nice, 32:52.000 --> 32:55.000 it immediately searches in all chat rooms, documents, chats, emails. 32:55.000 --> 32:57.000 It's a bit much. 32:57.000 --> 33:01.000 And a community member actually basically build a very simple, 33:01.000 --> 33:04.000 there's a little search clause in the top right now, 33:04.000 --> 33:08.000 and you click it, and you just search, you filter in this single chat. 33:08.000 --> 33:13.000 So, tiny thing maybe, but it makes your life sometimes a lot easier 33:13.000 --> 33:15.000 that it just searches and won't chat. 33:15.000 --> 33:17.000 And again, you can use it to the universal search. 33:17.000 --> 33:20.000 You can just say, only this chat, but that is still three more clicks. 33:20.000 --> 33:23.000 And, you know, it's little things that make the difference sometimes. 33:23.000 --> 33:24.000 Yes. 33:24.000 --> 33:43.000 So, so, you mean the Federation part? 33:43.000 --> 33:44.000 No, I mean the Federation part. 33:44.000 --> 33:47.000 No, I mean the Federation part. 33:47.000 --> 33:48.000 The single part is fairly good. 33:48.000 --> 33:51.000 You have to sort of drive up to take care of it. 33:51.000 --> 33:55.000 Right. So, basically, the question is, how can you run 33:55.000 --> 33:57.000 next to an higher availability environment? 33:57.000 --> 33:59.000 Yes. Yes, absolutely. 33:59.000 --> 34:03.000 Yes. Because it's in the core, although, as I said, 34:03.000 --> 34:05.000 there are some components in, in, in, in, in, in, 34:05.000 --> 34:07.000 rust and go, et cetera. 34:07.000 --> 34:09.000 And, call a borough, of course, a C++. 34:09.000 --> 34:12.000 So, not all components scale exactly the same, 34:12.000 --> 34:16.000 but the basics of it is largely still very easy horizontal scaling, 34:16.000 --> 34:18.000 because it's a PHP web app. 34:18.000 --> 34:24.000 And PHP means that every request that goes to the servers is almost completely independent. 34:24.000 --> 34:27.000 So, if one of the application servers goes down, you're good. 34:27.000 --> 34:30.000 If you run multiples, so that's pretty easy to do. 34:30.000 --> 34:32.000 Data bases can be done redundant. 34:32.000 --> 34:34.000 These data file systems can be done redundant. 34:34.000 --> 34:38.000 So, it's not really something that happens at an exot layer, 34:38.000 --> 34:41.000 but running a high scale, highly redundant. 34:41.000 --> 34:44.000 Nexot instance is not terribly difficult. 34:44.000 --> 34:48.000 Our largest instance has over 20 million users. 34:48.000 --> 34:51.000 They're not running it on a single-est Raspberry Pi. 34:51.000 --> 34:52.000 Let me put it that way. 34:52.000 --> 34:56.000 This is spread across data centers, across continents, even, 34:56.000 --> 34:59.000 and yet it feels as a single instance to the users. 34:59.000 --> 35:04.000 So, Nexot is, well, scaling is a solved problem in that regard. 35:04.000 --> 35:06.000 More questions. 35:06.000 --> 35:09.000 Yes, in the back. 35:37.000 --> 35:42.000 So, our engineers have a lot of, like, decision-powered themselves. 35:42.000 --> 35:46.000 That, I mean, we structure a company very much like an outsource community. 35:46.000 --> 35:49.000 So, a lot of the decisions are low. 35:49.000 --> 35:53.000 So, yeah, the question is, how do you balance features for sustainability? 35:53.000 --> 35:54.000 And, et cetera. 35:54.000 --> 35:58.000 And, of course, there is a grey area there, of course, 35:58.000 --> 36:00.000 when is it polishing and existing feature? 36:00.000 --> 36:02.000 When is it completely new, et cetera? 36:03.000 --> 36:07.000 By and large, often big features don't need a ton of people. 36:07.000 --> 36:11.000 For example, AI team is less, about two FTE. 36:11.000 --> 36:15.000 So, all the AI stuff that you've seen is two people. 36:15.000 --> 36:18.000 Yeah, and we're a company with 130 people. 36:18.000 --> 36:20.000 So, but this is very fancy. 36:20.000 --> 36:23.000 And we talk about a lot because we're actually competing 36:23.000 --> 36:27.000 and a lot of these capabilities with, well, Apple, Google, Microsoft, 36:27.000 --> 36:30.000 and, as a functionality, you can do that with Nexot. 36:30.000 --> 36:32.000 Of course, we're not building our own LLMs. 36:32.000 --> 36:34.000 That's not possible with two people. 36:34.000 --> 36:38.000 But we're using, we're standing on the shoulders of giants here, right? 36:38.000 --> 36:41.000 We're using all the technology that the people here at Foss Dam, 36:41.000 --> 36:44.000 and many other people all over the world are building. 36:44.000 --> 36:45.000 And we're taking advantage of it. 36:45.000 --> 36:47.000 We don't do office ourselves. 36:47.000 --> 36:50.000 We integrate what our friend said, call a board at us, et cetera. 36:50.000 --> 36:53.000 So, that's one part of the answer where you think, like, 36:53.000 --> 36:55.000 oh, they are spending all over their time of features. 36:55.000 --> 36:59.000 Yeah, we do features, but this is not always as resource 36:59.000 --> 37:01.000 intensive as you might think. 37:01.000 --> 37:03.000 And we're using a lot of tech. 37:03.000 --> 37:05.000 Now, the priorities that the engineers got, 37:05.000 --> 37:08.000 we try to basically just say these are priorities. 37:08.000 --> 37:11.000 And the first is basically fixing issues, mostly for customers, 37:11.000 --> 37:13.000 but also from the community. 37:13.000 --> 37:15.000 So, backfixing is the number one priority. 37:15.000 --> 37:18.000 And I would say 70% of the time is actually spend on that. 37:18.000 --> 37:19.000 Yeah. 37:19.000 --> 37:21.000 Optimization performance. 37:21.000 --> 37:22.000 I mean, do you call that a feature? 37:22.000 --> 37:24.000 I didn't really talk about performance at all, 37:24.000 --> 37:27.000 but each of the releases had a ton of performance work. 37:27.000 --> 37:29.000 And that's always very important, of course. 37:29.000 --> 37:33.000 It needs to be fast and stay fast. 37:33.000 --> 37:37.000 The second priority is then features, 37:37.000 --> 37:39.000 because we need, you know, our customers ask for things. 37:39.000 --> 37:42.000 There are tenders that have certain requirements, et cetera. 37:42.000 --> 37:45.000 So, if you want to keep hiring people and grow, 37:45.000 --> 37:49.000 then you need to also develop the features that are needed for new customers. 37:49.000 --> 37:53.000 And the third priority is, of course, working with the community, 37:53.000 --> 37:57.000 getting features from the community merge, reviewed, et cetera. 37:57.000 --> 38:01.000 So, bug fixing is priority number one, effectively. 38:01.000 --> 38:05.000 But, of course, I'm not going to list here, 38:05.000 --> 38:10.000 the 6,845 bugs we fixed because nobody gives you shit. 38:10.000 --> 38:14.000 At least, I don't think most of you want to hear that, right? 38:14.000 --> 38:16.000 So, I talk about the features. 38:16.000 --> 38:18.000 Doesn't mean most of the work goes in that. 38:18.000 --> 38:20.000 Any more questions? 38:20.000 --> 38:21.000 Yes? 38:21.000 --> 38:24.000 You're talking about governments integration. 38:24.000 --> 38:27.000 How much adaption is there already going on? 38:27.000 --> 38:31.000 Is it like, did you cut some or just try to build a future political government? 38:31.000 --> 38:34.000 Or are they already using it? 38:34.000 --> 38:36.000 So, tons of governments are using it. 38:36.000 --> 38:38.000 For example, and I always love this example. 38:38.000 --> 38:41.000 The European data protection supervisor, 38:41.000 --> 38:44.000 so the organization that, as in your EU level, 38:44.000 --> 38:47.000 supervisors, all the local data protection authorities, 38:47.000 --> 38:49.000 they're using next slide. 38:49.000 --> 38:51.000 And they're recommending everybody to do so. 38:51.000 --> 38:54.000 So, listen to your data protection supervisor, please. 38:54.000 --> 38:56.000 Europe, there will be nice. 38:56.000 --> 38:58.000 But, I don't know. 38:58.000 --> 39:00.000 Many of them I cannot talk about. 39:00.000 --> 39:03.000 I mean, a lot of customers are twitchy about, I don't know. 39:03.000 --> 39:05.000 There's probably a lot of lobbying. 39:05.000 --> 39:10.000 There's a certain company from Redmont that has more lobbyist and Brussels than we have employees. 39:10.000 --> 39:13.000 So, I cannot always name the names, 39:13.000 --> 39:17.000 but the European Court of Auditors, there are several parlaments. 39:18.000 --> 39:21.000 Half a dozen ministries are foreign affairs. 39:21.000 --> 39:25.000 Hundreds of cities in Europe use next slide. 39:25.000 --> 39:28.000 But again, they don't make noise about it because, oh my God, 39:28.000 --> 39:30.000 if something breaks, then you have Munich all over. 39:30.000 --> 39:32.000 So, a lot of it is under the radar, 39:32.000 --> 39:34.000 but we have thousands of customers, 39:34.000 --> 39:37.000 and we don't sell anything under a few hundred users. 39:37.000 --> 39:41.000 So, they're all on the bigger side, so to say. 39:41.000 --> 39:43.000 And that's federal government, 39:43.000 --> 39:45.000 state, strategic whole-stein. 39:45.000 --> 39:48.000 Thankfully, it's very willing to talk about the use of next slide, 39:48.000 --> 39:49.000 and open-source, etc. 39:49.000 --> 39:52.000 So, they use next slide all over the state, 39:52.000 --> 39:55.000 and there are many other places. 39:55.000 --> 39:57.000 Yeah. 39:57.000 --> 39:59.000 Yes, next question. 40:14.000 --> 40:23.000 Yeah, so we don't have any repeat the question. 40:23.000 --> 40:25.000 Sorry, squirrel. 40:25.000 --> 40:30.000 The question is, how do you manage this community versus company balance? 40:30.000 --> 40:32.000 Yeah, it's a tricky one. 40:32.000 --> 40:36.000 So, we don't have any formal things, 40:36.000 --> 40:39.000 at least not like on the outside in. 40:39.000 --> 40:43.000 We have a, what's it named? 40:43.000 --> 40:46.000 The name relates, well, there's a Frank, of course. 40:46.000 --> 40:49.000 But, no, there's a group of people who handle, 40:49.000 --> 40:50.000 like, conflicts and such. 40:50.000 --> 40:52.000 I kind of think of the name. 40:52.000 --> 40:55.000 There's some kind of complaining, 40:55.000 --> 40:57.000 board thing. 40:57.000 --> 41:01.000 But otherwise, like, our approach is simply to work in the open. 41:01.000 --> 41:04.000 So, like, if you want to know our roadmap for the next release, 41:04.000 --> 41:06.000 I mean, there are boards that are in GitHub, 41:06.000 --> 41:08.000 that are completely open. 41:08.000 --> 41:10.000 All the issues are in the open. 41:10.000 --> 41:11.000 All the pull-to-quest are open. 41:11.000 --> 41:15.000 And we treat community pull-to-quest equal to employee pull-to-quest, 41:15.000 --> 41:17.000 at least mostly. 41:17.000 --> 41:21.000 And I think, by and large, that has resulted in fairly little, 41:21.000 --> 41:26.000 you know, friction, at least on a development level. 41:26.000 --> 41:30.000 Of course, when you say, well, how do you treat future requests from home users, 41:30.000 --> 41:33.000 versus businesses, that's a different thing. 41:34.000 --> 41:37.000 I mean, the businesses pay as money in the home users, of course, 41:37.000 --> 41:38.000 don't. 41:38.000 --> 41:41.000 So, as management, we have to make sure that every, you know, 41:41.000 --> 41:44.000 release, there is at least a number of things done for private users, 41:44.000 --> 41:46.000 as well, because, after all, back in the day, 41:46.000 --> 41:47.000 we started next slide. 41:47.000 --> 41:49.000 It's, we personally care about privacy. 41:49.000 --> 41:52.000 And privacy, as I think, people have not companies. 41:52.000 --> 41:54.000 So, if we want to succeed in a mission, 41:54.000 --> 41:56.000 we actually have to build features for home users, 41:56.000 --> 41:59.000 because you're the ones who we do all of this for, 41:59.000 --> 42:00.000 fundamentally. 42:00.000 --> 42:02.000 Yeah, all personal level. 42:02.000 --> 42:05.000 I hope that answers the question and times up. 42:05.000 --> 42:06.000 I have to leave it here. 42:06.000 --> 42:08.000 Thank you all very much. 42:08.000 --> 42:10.000 APPLAUSE