WEBVTT Kind: captions Language: en 00:00:00.020 --> 00:00:01.500 All right, so I have Jos here 00:00:01.500 --> 00:00:02.760 with me from Nextcloud. 00:00:02.760 --> 00:00:03.760 I wanted to ask you, 00:00:03.760 --> 00:00:05.280 if I was a contributor and wanted 00:00:05.280 --> 00:00:06.860 to contribute to Nextcloud, what are kind 00:00:06.860 --> 00:00:08.860 of the main places that you'd have people 00:00:08.860 --> 00:00:10.960 start if they're not technical whatsoever? 00:00:11.300 --> 00:00:12.120 Yeah, there's a lot 00:00:12.120 --> 00:00:13.270 of ways you can get involved. 00:00:13.270 --> 00:00:14.400 I mean, we have a big translation 00:00:14.400 --> 00:00:15.350 community, of course, 00:00:15.350 --> 00:00:17.080 if you're an English native speaker, 00:00:17.080 --> 00:00:19.080 we do pretty much everything in English, 00:00:19.080 --> 00:00:19.920 so that doesn't help. 00:00:19.920 --> 00:00:21.200 But we could definitely use help 00:00:21.200 --> 00:00:22.460 with Spanish, for example, 00:00:22.460 --> 00:00:23.850 or other languages, of course. 00:00:23.850 --> 00:00:25.100 And we have a lot of people, of course, 00:00:25.100 --> 00:00:26.510 just finding issues on GitHub 00:00:26.510 --> 00:00:28.420 and reporting various problems. 00:00:28.420 --> 00:00:30.580 Of course, in these days with AI, 00:00:30.580 --> 00:00:32.430 like if you want to learn to program, 00:00:32.430 --> 00:00:34.500 it is actually easier to do that, I think, 00:00:34.500 --> 00:00:35.900 than ever before, especially 00:00:35.900 --> 00:00:37.620 for small things and to get into it. 00:00:37.620 --> 00:00:38.420 I think this is also 00:00:38.420 --> 00:00:39.660 where we really see that a lot 00:00:39.660 --> 00:00:41.260 more people are getting involved. 00:00:41.260 --> 00:00:43.210 And I think a contribution maybe people 00:00:43.210 --> 00:00:45.300 really underestimate is to, you know, 00:00:45.300 --> 00:00:46.500 kind of spread the word. 00:00:46.500 --> 00:00:48.050 You know, earlier we were talking 00:00:48.050 --> 00:00:50.150 about how it's sometimes mind blowing 00:00:50.150 --> 00:00:52.320 how many people that you think 00:00:52.320 --> 00:00:54.790 should know that there is an alternative, 00:00:54.790 --> 00:00:56.550 that there are other solutions, right? 00:00:56.550 --> 00:00:58.480 People who are sometimes even aware 00:00:58.480 --> 00:01:00.580 of the problems with the solutions they're 00:01:00.580 --> 00:01:02.680 using, that they are dependent on a single 00:01:02.680 --> 00:01:04.800 big tech player, then it stops because 00:01:04.800 --> 00:01:06.800 they don't know that actually they could 00:01:06.800 --> 00:01:08.890 do something, that there is another way. 00:01:08.890 --> 00:01:10.700 And so telling other people 00:01:10.700 --> 00:01:12.620 about that or even teaching them 00:01:12.620 --> 00:01:14.520 to use it, like there are really quite 00:01:14.520 --> 00:01:17.000 a few people who have a real impact 00:01:17.000 --> 00:01:19.280 by just helping other people, training 00:01:19.280 --> 00:01:21.580 them, showing them there's a different way 00:01:21.580 --> 00:01:23.110 and explaining the problems. 00:01:23.110 --> 00:01:24.760 I think this is actually a really 00:01:24.760 --> 00:01:27.580 undervalued contribution in a way. 00:01:27.580 --> 00:01:29.960 Like it's really impactful because well 00:01:29.960 --> 00:01:31.960 what you don't know you can't do, right? 00:01:31.960 --> 00:01:33.960 And a lot of people these days are aware 00:01:33.960 --> 00:01:35.820 of the problems with like for example 00:01:35.820 --> 00:01:37.840 social media and if you tell them 00:01:37.840 --> 00:01:39.660 about Mastodon that's super helpful 00:01:39.660 --> 00:01:41.300 because there is a way out 00:01:41.300 --> 00:01:43.400 and not knowing that is really a shame 00:01:43.400 --> 00:01:45.500 and the same as for Nextcloud. 00:01:45.500 --> 00:01:47.940 You don't have to keep using it and have 00:01:47.940 --> 00:01:50.920 to keep reliant on Google or Microsoft 00:01:50.920 --> 00:01:52.960 or all the other big tech players. 00:01:53.500 --> 00:01:57.020 And even when you have heard of Nextcloud 00:01:57.020 --> 00:01:59.200 and you think it is only usable if you 00:01:59.200 --> 00:02:01.150 self-host, well that's not true, right? 00:02:01.150 --> 00:02:03.360 There are lots of hosting providers, 00:02:03.360 --> 00:02:05.580 companies that offer you a Nextcloud 00:02:05.580 --> 00:02:08.420 solution for private use for a charitable 00:02:08.420 --> 00:02:11.330 organization or a soccer club, 00:02:11.330 --> 00:02:13.720 you know, or a sports organization 00:02:13.720 --> 00:02:16.800 or a church, etc. And again, 00:02:17.040 --> 00:02:18.860 people should know about that, you know. 00:02:18.860 --> 00:02:19.910 I think that's really 00:02:19.910 --> 00:02:21.170 a thing people can do. 00:02:21.170 --> 00:02:23.340 Yeah, something that I know it was a very, 00:02:23.340 --> 00:02:25.210 very short part of the conversation, 00:02:25.210 --> 00:02:26.610 but there was a little thing 00:02:26.610 --> 00:02:28.060 about AI code of conduct. 00:02:28.319 --> 00:02:30.640 So do you mind maybe adding a little bit 00:02:30.640 --> 00:02:32.740 of nuance that you might someone, 00:02:33.140 --> 00:02:35.140 you might want people to hear if they're 00:02:35.140 --> 00:02:36.940 going to start doing AI contributions 00:02:36.940 --> 00:02:38.710 and your guys' kind of policy on that? 00:02:38.710 --> 00:02:40.600 Yeah, I already said, you know, AI 00:02:40.600 --> 00:02:42.810 can lower the barrier and help you learn. 00:02:42.810 --> 00:02:45.080 I mean, what was also said, right? 00:02:45.080 --> 00:02:47.510 Like, ideas have always been cheap, 00:02:47.510 --> 00:02:48.990 we say in open source often. 00:02:48.990 --> 00:02:50.500 Like, ideas are cheap. 00:02:50.870 --> 00:02:53.210 The ability to do it is the hard part, 00:02:53.210 --> 00:02:54.160 the expensive part. 00:02:54.160 --> 00:02:55.520 We have lots of ideas how we 00:02:55.520 --> 00:02:56.620 can improve Nextcloud. 00:02:56.620 --> 00:02:58.180 It's just, do you have time 00:02:58.180 --> 00:02:59.160 to actually do it? 00:02:59.160 --> 00:03:00.880 That's often a limiting factor. 00:03:00.880 --> 00:03:02.380 It's in life with many things. 00:03:02.380 --> 00:03:03.830 With AI, some of these things 00:03:03.830 --> 00:03:06.620 become easier in a way. 00:03:06.860 --> 00:03:09.380 But of course, there is a quality impact. 00:03:09.380 --> 00:03:11.710 And what was said on the stage also, 00:03:11.710 --> 00:03:13.260 we want to maintain, of course, 00:03:13.260 --> 00:03:14.500 quality of Nextcloud. 00:03:14.860 --> 00:03:16.760 And so for us on one hand, we don't want 00:03:16.760 --> 00:03:18.540 to close the door on AI contributions 00:03:18.540 --> 00:03:19.580 because again, it genuinely 00:03:19.580 --> 00:03:21.180 can lower the barrier for people 00:03:21.180 --> 00:03:23.460 to contribute to make a difference. 00:03:23.910 --> 00:03:25.960 But on the other hand, what does not work 00:03:25.960 --> 00:03:29.060 is if we have to, you know, we get 00:03:29.060 --> 00:03:32.840 like a 5,000 line of code AI slop change 00:03:32.840 --> 00:03:34.790 that changes things all over the place. 00:03:34.790 --> 00:03:36.290 And like, who can review that? 00:03:36.290 --> 00:03:38.510 Who can check that for quality? 00:03:38.510 --> 00:03:40.670 You know, it's not fair also 00:03:40.670 --> 00:03:41.920 to the developers, right? 00:03:41.920 --> 00:03:42.970 Because of the people 00:03:42.970 --> 00:03:44.040 who contribute to Nextcloud, 00:03:44.040 --> 00:03:45.460 a lot of them do it in the free time. 00:03:45.460 --> 00:03:46.550 It's not like they're paid 00:03:46.550 --> 00:03:49.280 to review 5,000 lines of code either. 00:03:49.280 --> 00:03:50.690 It's not fun for everyone. 00:03:50.690 --> 00:03:52.640 And if you then ask follow-up questions 00:03:52.640 --> 00:03:55.640 and the answer is basically copy-paste it 00:03:55.640 --> 00:03:58.380 from ChatGPT, that's not cool. 00:03:58.380 --> 00:03:59.500 You know what I mean? 00:03:59.500 --> 00:04:01.340 We should have a human touch. 00:04:01.340 --> 00:04:04.180 So it's really important that the human 00:04:04.180 --> 00:04:06.600 is in charge, that you understand the code 00:04:06.600 --> 00:04:08.500 that was written, that you contribute, 00:04:08.500 --> 00:04:10.700 that you have thought about it, 00:04:10.700 --> 00:04:12.720 where it fits, how it should fit. 00:04:12.720 --> 00:04:14.320 And of course, you don't 00:04:14.320 --> 00:04:16.380 know the architecture. I get that. 00:04:16.380 --> 00:04:17.930 But at least you should have done 00:04:17.930 --> 00:04:21.019 an effort to find it out or maybe ask 00:04:21.019 --> 00:04:23.420 in advance where and how to do something. 00:04:23.740 --> 00:04:25.240 Even if you use AI for coding, 00:04:25.240 --> 00:04:26.840 you know what I mean? It's a tool. 00:04:27.200 --> 00:04:29.060 It's not the thing that you use 00:04:29.060 --> 00:04:30.580 so you can stop thinking. 00:04:30.580 --> 00:04:31.930 That's not what AI is for. 00:04:31.930 --> 00:04:33.580 AI aside, I think what it made me 00:04:33.580 --> 00:04:35.130 appreciate, and it's not something I think 00:04:35.130 --> 00:04:37.170 about from my end, right? 00:04:37.170 --> 00:04:39.030 As someone who's a little less technical, 00:04:39.030 --> 00:04:40.980 I'm not contributing to Nextcloud code, 00:04:40.980 --> 00:04:41.890 but I'm like, why don't 00:04:41.890 --> 00:04:43.210 they just do this thing? 00:04:43.210 --> 00:04:45.060 But what I'm not appreciating is even 00:04:45.060 --> 00:04:46.820 if AI does the thing perfectly, 00:04:46.820 --> 00:04:48.220 it doesn't take into account 00:04:48.220 --> 00:04:49.910 the maintenance that you're talking about. 00:04:49.910 --> 00:04:51.580 If you just have 10 new features, 00:04:51.940 --> 00:04:52.760 but they're not maintained, 00:04:52.760 --> 00:04:54.350 it doesn't really matter long term. 00:04:54.350 --> 00:04:56.830 So AI won't solve, I guess, every problem. 00:04:56.830 --> 00:04:58.280 It can help with some things. 00:04:58.280 --> 00:04:59.940 Like it's very good, for example, 00:04:59.940 --> 00:05:01.740 certain bugs, you know, like there's 00:05:01.740 --> 00:05:03.640 a race condition and it's caused 00:05:03.640 --> 00:05:06.370 by a piece really deep down the tree here, 00:05:06.370 --> 00:05:07.560 interacting with something 00:05:07.560 --> 00:05:09.160 really deep down the tree there. 00:05:09.160 --> 00:05:11.210 And as a human, that's very hard to find. 00:05:11.210 --> 00:05:12.160 And these AI models 00:05:12.160 --> 00:05:13.790 can be very good at finding it. 00:05:13.790 --> 00:05:15.140 So there are certain things 00:05:15.140 --> 00:05:16.880 where AI is incredibly helpful. 00:05:17.120 --> 00:05:20.200 But then often when you ask it 00:05:20.200 --> 00:05:21.980 to build indeed a whole feature, 00:05:22.180 --> 00:05:23.980 it architects it in a way 00:05:23.980 --> 00:05:27.040 that just ignores coding conventions, 00:05:27.040 --> 00:05:29.260 ignores the architecture of the app, 00:05:30.039 --> 00:05:31.640 yes, exactly the future. 00:05:31.640 --> 00:05:33.740 And if you build on a couple of these 00:05:33.740 --> 00:05:35.540 things, yes, each of them looks good 00:05:35.540 --> 00:05:37.940 at first sight, seems to work fine 00:05:37.940 --> 00:05:40.090 at first sight, but as soon as you add two 00:05:40.090 --> 00:05:41.990 more things or you have this like yeah 00:05:41.990 --> 00:05:43.760 but did you think of that thing? 00:05:44.090 --> 00:05:45.680 Oh this completely falls apart 00:05:45.680 --> 00:05:47.760 and this is a challenge because often 00:05:47.760 --> 00:05:50.540 people have big ideas and I get 00:05:50.540 --> 00:05:53.860 that but there is often many more aspects 00:05:54.350 --> 00:05:55.700 to you know to each feature 00:05:55.700 --> 00:05:57.200 and to each improvement right? 00:05:57.200 --> 00:05:58.740 Most people look at a problem 00:05:59.220 --> 00:06:00.880 from their specific perspective 00:06:00.880 --> 00:06:02.680 but when you build software it has 00:06:02.680 --> 00:06:03.780 to work for everyone, 00:06:03.780 --> 00:06:06.040 not just for one single person. 00:06:06.240 --> 00:06:07.390 And this is much harder 00:06:07.390 --> 00:06:09.000 than it seems at first sight. 00:06:09.000 --> 00:06:10.820 I've once did a poll about that on 00:06:10.820 --> 00:06:13.660 Mastodon asking like, in terms of effort, 00:06:13.660 --> 00:06:15.760 how much more work it is to make a feature 00:06:15.760 --> 00:06:18.640 that works for one organization, one use 00:06:18.640 --> 00:06:21.890 case versus making that same feature, 00:06:21.890 --> 00:06:23.640 but so that it kind of is a generic 00:06:23.640 --> 00:06:25.840 solution for other organizations. 00:06:26.360 --> 00:06:27.970 And the replies were really like five 00:06:27.970 --> 00:06:29.880 times, six times, seven times as much. 00:06:30.160 --> 00:06:31.100 So that's significant. 00:06:31.100 --> 00:06:33.000 You have an idea for your specific use 00:06:33.000 --> 00:06:35.440 case, but we can't build Nextcloud 00:06:35.440 --> 00:06:37.140 for each individual user. 00:06:37.140 --> 00:06:38.540 It has to work for everyone. 00:06:38.800 --> 00:06:40.650 And that just requires more thinking, 00:06:40.650 --> 00:06:41.540 more discussion. 00:06:41.800 --> 00:06:43.620 And if you just throw AI 00:06:43.620 --> 00:06:46.720 at that, that does not solve it. 00:06:46.720 --> 00:06:47.980 It solves your problem, 00:06:47.980 --> 00:06:50.020 but doesn't help everyone. Right. 00:06:50.020 --> 00:06:51.360 Well, thank you so much, Jos. 00:06:51.360 --> 00:06:52.780 You're welcome. Always a pleasure. 00:06:53.120 --> 00:06:53.920 Likewise.