WEBVTT 00:00.000 --> 00:10.400 Anyway, as we get to the end of the day, our next speaker is going to cover burnout in open 00:10.400 --> 00:15.520 source, particularly burnout as a systems issue and what we can actually change. 00:15.520 --> 00:20.280 Please welcome Miranda Heath. 00:20.280 --> 00:27.800 Thank you so much for hearing me okay, it's the microphone, great. 00:27.800 --> 00:33.040 Has everyone's energy, it's second to last talk, it's really warm in here, give me a shout. 00:33.040 --> 00:39.600 It's okay, it was okay, I would ask if this sounds familiar to anyone. 00:39.600 --> 00:44.880 My open source success went from a major blessing to a great curse, it was one of the darkest 00:44.880 --> 00:49.760 times of my life, something that started out with hope and such light ended up being 00:49.760 --> 00:53.280 about just getting thousands of emails. 00:53.280 --> 00:58.320 So these words, the words of Mark Grabansky, a friend, ten masters, speak to the issue of burnout 00:58.320 --> 01:03.560 in open source software, I'm Miranda Heath, I study the psychology and the philosophy about 01:03.560 --> 01:08.160 tourism, if I'm not a programmer, and I've been researching what's been causing open 01:08.160 --> 01:11.160 source developer to burn out. 01:11.160 --> 01:15.760 So what this psychology tells us about burnout, burnout is a work related to the exhaustion 01:15.800 --> 01:20.320 of both our physical and our mental energy, and it affects us in three different ways. 01:20.320 --> 01:25.320 It affects firstly our motivation, so if we're fatigued, we can't force ourselves to exert 01:25.320 --> 01:30.560 effort, our work becomes lower quality and it's a really big predictor of quitting. 01:30.560 --> 01:35.160 It affects our cognition, so it affects our thought processes. 01:35.160 --> 01:39.400 We might start thinking about things in a more cynical or a negative way, it affects 01:39.400 --> 01:42.000 our self-worth and our self-esteem. 01:42.040 --> 01:46.200 We can become obsessive over issues at work and unable to drop them. 01:46.200 --> 01:49.960 We can also feel like we've lost that sense of direction, both in work, and our sense 01:49.960 --> 01:54.040 of meaning and what we're doing in our life generally, and we might also indulge in 01:54.040 --> 02:00.520 dark humor and escapism, and in the worst cases this can even need to behave like addiction. 02:00.520 --> 02:05.520 And finally, it affects our emotions, so burnout makes us feel drained emotionally. 02:05.520 --> 02:09.800 It also makes it harder for us to regulate our own emotions, so we might be quick to overreact 02:09.840 --> 02:14.440 or to anger or to panic, and it makes it harder to do with emotions in other people as well. 02:14.440 --> 02:19.720 We have like compassion for tea, and these two factors can make it hard to, you can affect 02:19.720 --> 02:23.640 your relationships outside of work and inside work as well, so it can affect your life 02:23.640 --> 02:25.040 in a broader way. 02:25.040 --> 02:28.160 We also might feel guilty or experienced depression. 02:28.160 --> 02:31.440 In fact, mental and physical health really suffer with burnout. 02:31.440 --> 02:35.720 So if you're wondering why we should care about burnout in open source, you can see that 02:35.720 --> 02:40.240 it's putting developers at risk of real harm, and it's an ethical issue. 02:40.240 --> 02:46.120 Now, there are some things that drain our energy, so things like having a high workload, 02:46.120 --> 02:51.680 really complicated work, and also a lot of responsibility, having unclear expectations, 02:51.680 --> 02:55.800 what our role is supposed to be, conflict with our colleagues, or feeling like we're being 02:55.800 --> 03:02.040 treated unfairly at work, and some of these you might recognize from working in open source. 03:02.280 --> 03:06.320 And there is something that's a thinner energy, so things like having good social support, 03:06.320 --> 03:11.680 having opportunity to use to grow and develop and learn new things, fair pay, or autonomy, 03:11.680 --> 03:16.720 so having freedom and control over what you're working on and how, having joy and meaning 03:16.720 --> 03:19.760 in the work you do and feeling competent at it. 03:19.760 --> 03:25.000 Now, burnout is caused by an imbalance in the things that drain us and the things that sustain us, 03:25.000 --> 03:28.720 so it is also an issue of fairness and injustice. 03:29.680 --> 03:35.520 Now, burnout in open source is also a significant problem, so 73% of software developers 03:35.520 --> 03:40.480 in open and closed source worldwide have experienced burnout at some point in their career. 03:40.480 --> 03:44.160 In fact, they wonder how many people in the room put your hand up if you feel like you've experienced 03:44.160 --> 03:50.320 burnout or working in open source, oh my goodness, that's worse than I thought, that's more than 72%. 03:50.320 --> 03:57.520 60% of over 400 open source developers considered quitting open source in 2024, 03:57.600 --> 04:02.960 and they've been at least four globally significant security incidents where burnout has played a role, 04:02.960 --> 04:08.960 so burnout in open source is also an issue of sustainability, the dreaded word, and security. 04:10.000 --> 04:15.600 Now, I wanted to better understand what it is about open source in particular that's putting people at risk of burnout, 04:15.600 --> 04:18.960 and I wanted to understand this in developers' own words. 04:18.960 --> 04:25.120 So I analyzed more than 50 articles, blogbows, podcasts, videos, and online discussions on burnout 04:25.120 --> 04:29.680 by people working in and around open source software. You'll hear their quotes throughout if talk. 04:30.400 --> 04:35.280 I also spoke to eight developers. I start feedback from the community, and then I wrote it all together 04:35.280 --> 04:39.280 into a report on the causes and possible solutions to burnout in open source software. 04:41.120 --> 04:44.960 So what are these causes? What is it about working in open source and makes it dream more than 04:44.960 --> 04:51.840 it sustains? Well firstly, it's difficult to get paid, so many developers felt it's unrealistic 04:51.920 --> 04:57.600 to expect sufficient and reliable payment for the work that they do. This is also backed up by the numbers, 04:57.600 --> 05:02.560 so according to research by the sovereign tech agency and by tide lift, about 60% of open source 05:02.560 --> 05:08.240 developers don't receive any payment at all for that open source, and a further 16% of them are paid, 05:08.240 --> 05:15.120 but it's not enough for them to earn a living. Now, if we feel like we're being treated unfairly, 05:15.120 --> 05:20.400 then this is likely to lead to burnout, and developers feel like that they're being unfairly exploited, 05:20.400 --> 05:24.800 that they're being pressured into free labor for an industry that could actually afford to pay them. 05:25.440 --> 05:30.560 In the words of David Alma Kiever, we've concluded that we'd much rather make people work for free 05:30.560 --> 05:35.680 in their spare time to produce adequate software, and then shame them into supporting it, 05:35.680 --> 05:39.200 then fairly compensate them for their labor and get good software out of it. 05:41.280 --> 05:45.360 Another thing that makes open source training is the workload and the time that it takes 05:45.360 --> 05:50.560 to high workload zero, so associated with burnout, and developers describe open source work as 05:50.560 --> 05:54.880 kind of like a stealth shift, so because it starts out as a hobby, you might not recognize it, 05:54.880 --> 05:59.360 it's also so work, but a labor of love is a labor nonetheless, especially once it becomes more 05:59.360 --> 06:04.400 popular and there's more work to be done, and because it's hard to get paid, many developers also 06:04.400 --> 06:09.920 have full-time jobs alongside their open source, effectively doing this double shift and incredibly 06:09.920 --> 06:13.600 long hours in order to be able to make a living while doing their open source work. 06:14.560 --> 06:20.080 In the words of Jacob Kaplan Moss, I wasn't getting paid to do the work I felt like I had to do, 06:20.080 --> 06:24.880 and I had to do the work that I was getting paid to do because of mortgages and builders and car payments. 06:26.880 --> 06:32.400 On top of that, developers are increasingly overloaded with issues and requests and bug reports, 06:32.960 --> 06:37.200 and this is often a problem when they're particularly low effort, because low effort reports mean 06:37.200 --> 06:41.840 that it takes a higher effort in order to review them, and this needs developers feeling like their 06:41.840 --> 06:47.200 time is worth less than other people's, and remember this feeling of unfairness is also associated with burnout. 06:48.240 --> 06:53.600 It's is also seen to be getting worse with the rise of AI coding, as AI lowers the effort required 06:53.600 --> 06:58.800 to make the poor request, leading to a greater volume of requests and of a worse quality than if they're 06:58.800 --> 07:05.520 written by a human who understands the code base. In more time we spend working, the last time we have 07:05.520 --> 07:11.520 our friends, our family, for relationships, for interests outside of work, and the psychology 07:11.520 --> 07:16.240 literature tells us that both loneliness and having work as the only meaningful thing in your life 07:16.240 --> 07:22.000 is likely to lead to burnout. Developers are being left, making really painful, existential choices, 07:22.000 --> 07:26.880 as an older loss and puts it, I just can't see how I'll be able to make the time for both 07:26.880 --> 07:34.000 raising a family and doing open source. Another cause of burnout, all of that maintenance work 07:34.000 --> 07:39.600 that developers are doing isn't always rewarding to do. So in psychology we tend to distinguish 07:39.600 --> 07:44.320 between two types of reward, intrinsic reward, which are things that we do for their unsafe, 07:44.320 --> 07:49.280 for the joy of it, and then extrinsic reward, which are things we do for some external good, 07:49.280 --> 07:54.720 like money recognition and job prospects. As we've seen, maintenance is not sufficiently 07:54.720 --> 07:59.840 extrinsically rewarded, for developers are rarely paid, they're rarely recognized or thanked for 07:59.840 --> 08:04.560 doing maintenance in a market society that favors extraction rather than sustaining things. 08:05.600 --> 08:11.440 And developers don't always find maintenance intrinsically rewarding either. So doing triars, 08:11.440 --> 08:17.440 code review, managing people's expectations is not the kind of creative expression or intellectual 08:17.440 --> 08:22.240 challenge that got a lot of us involved in coding in the first place. And this is also being 08:22.240 --> 08:27.200 exacerbated by the rise of AI coding, because in some cases it feels like we have intentionless 08:27.200 --> 08:32.240 machines doing the creative part of writing code, and then people end up doing the machine 08:32.240 --> 08:36.960 like work of looking for hard to spot errors in this deceptively competent looking code. 08:37.360 --> 08:41.200 And if you want work to be intrinsically motivating, this is totally the wrong way around. 08:42.320 --> 08:46.640 I'm pushing ourselves to do work with neither extrinsically nor intrinsically rewarding, 08:46.640 --> 08:52.400 it drains our energy and puts it at risk of burnout. And T. Red sums this up well. 08:52.720 --> 08:56.800 I would rather be remembered as a bad artist than a good programmer, 08:56.800 --> 09:01.520 but now I'm asked more and more to express myself less and to maintain the project more. 09:03.920 --> 09:08.560 Now a big cause of burnout that came up in the research I was doing was that the community 09:08.560 --> 09:12.800 can be toxic. I don't know what the overlap is between open source developers and 09:12.800 --> 09:18.240 Brittany fans, but there she is. I talked about social pressure earlier, 09:18.320 --> 09:23.200 the people act entitled to developers' time, and they demand their help for features and 09:23.200 --> 09:28.160 for fixes, and sometimes resort to anger or insults or shaming when their needs aren't there. 09:28.160 --> 09:31.520 Even if it's for a really specific use case, it's not generally helpful. 09:32.560 --> 09:36.800 And developers also experience rudeness from other collaborators, or sometimes described 09:36.800 --> 09:41.840 becoming rude themselves. And remember when we're burnt out, our emotion regulations were, 09:41.840 --> 09:47.680 so we feel worthless, we have low self esteem, so it's easier to lash out if we stretch 09:47.840 --> 09:53.360 them and try to defend the very last bit of our ego. And the fact that a lot of this toxic behavior 09:53.360 --> 09:58.640 happens in public is particularly bad for burnout. It can make us feel powerless or humiliated or 09:58.640 --> 10:04.240 feeding to those feelings of guilt and inadequacy, which makes it a lot worse. All of this makes 10:04.240 --> 10:09.760 it very hard to want to carry on doing maintenance work. As Jamie Kyle puts it, every single day 10:09.760 --> 10:15.520 and reading someone else rant about how awful of a job we're doing. It's been hard to stay motivated. 10:15.520 --> 10:21.920 I've practically stopped looking at issues and at poor requests. Now community toxicity can 10:21.920 --> 10:28.000 lead to what I've delicately called the burnout death spiral. So in communities of toxic, 10:28.560 --> 10:33.120 maintainers will light a burnout that's harder to regulate their emotions and are more likely 10:33.120 --> 10:39.040 to get into conflict. And technologies with toxic communities are seen as riskier investments, 10:39.040 --> 10:42.720 which might mean that businesses and governments are less likely to want to get involved in 10:42.720 --> 10:48.240 investing and open-source, which makes it harder to get paid. And then issues where people 10:48.240 --> 10:53.200 are rude to each other have been shown to take longer to solve and rudeness in a community 10:53.200 --> 10:58.560 to deter as newcomers from getting involved, all of which contributes to having a greater workload. 10:58.560 --> 11:02.720 And the more conflict there is, the harder it is to get paid, the greater the workload, 11:02.720 --> 11:07.360 the more likely we are to burn out. And then the more likely we are to engage in toxic behavior 11:07.440 --> 11:14.960 and further spiral goings on and on. And finally, developers are carrying a tremendous amount 11:14.960 --> 11:20.320 of responsibility. So the majority work either alone or in small teams. This is research from 11:20.320 --> 11:26.560 the sovereign tech agency. And many of them also feel irreplaceable. So I feel like if I stop doing this 11:26.560 --> 11:32.000 so I don't keep doing it who will. And when we have too much responsibility and when we feel like 11:32.000 --> 11:36.480 we can't be replaced, we're less likely to recognize the meaning to slow down and we're much more 11:36.480 --> 11:42.400 likely to burn out. There are three sources of responsibility in particular that we're weighing 11:42.400 --> 11:47.200 on developers. So firstly, responsibility towards the project, the people, the factors that depend 11:47.200 --> 11:51.840 on it. So caring about it, that it's being good quality, it can use to be usable, it continues 11:51.840 --> 11:56.640 to exist at all. And then there is responsibility towards other collaborators on the project 11:56.640 --> 12:02.160 to employ youths. And then responsibility towards the open-source community into uphold open-source 12:02.240 --> 12:08.800 values. And this can come into conflict. Let me illustrate this with an example. So who's been 12:08.800 --> 12:15.360 following what happened at Tailwind, CSS, can you note? People. So in what I thought was a really 12:15.360 --> 12:20.960 admirably candid recording, the creator Adam Weithen described how revenues were down. So he 12:20.960 --> 12:26.320 had to lay off 70% of his staff. He explained how he wasn't going to implement a request that would 12:26.320 --> 12:31.920 make documentation easier for LLM's to read because the documentation helps Tailwind sell premium 12:31.920 --> 12:36.720 features, which would help keep the revenue streams up. And then in response to this post, 12:36.720 --> 12:41.200 some people got angry with him and said he was betraying open source values or just being in it 12:41.200 --> 12:46.560 for the money. Now this was a lose-new situation. It's not possible to juggle these competing 12:46.560 --> 12:52.160 responsibilities. Yet this podcast culminates in and saying that he feels personally like a failure 12:52.160 --> 12:56.800 and like a fucking idiot for making something so popular, but not being able to figure out how to 12:56.800 --> 13:02.880 make enough money to keep people working on it. And it's made me think of this art installation. 13:02.880 --> 13:07.600 I thought about not putting this in because it makes me want to cry every time. So it's a robot 13:07.600 --> 13:12.720 that slowly leaks this blood-like hydraulic fluid and it's programmed when too much of it leaks 13:12.720 --> 13:18.160 out to stop and scoop it back in. And at first it can stay on top of this scooping and it uses 13:18.160 --> 13:23.920 its downtime to do a little happy on waving dances. But the more that it leaks out, the more time 13:23.920 --> 13:28.480 it's then scooping, the less time it spends doing it's happy dancing until eventually it's 13:28.480 --> 13:33.840 all blood out and switched off completely. And I can really see parallels between this and some of 13:33.840 --> 13:38.880 the developers that I've researched descriptions of what it can feel like working in open source software. 13:41.280 --> 13:44.160 So congratulations, we made it to the low point of my talk. 13:46.240 --> 13:48.880 So now I'm going to try and let a little bit of light locked in with them in. 13:48.880 --> 13:55.840 So the open source community excels at recognizing the importance of collaboration that the value 13:55.840 --> 13:59.840 of your work can't be attributed to you and you alone and that we can do better things if we 13:59.840 --> 14:04.880 work together out in the open. But it's not so good at applying this same logic to burnout. 14:04.880 --> 14:10.400 So individuals are being left to meet these impossible sacrifices to keep working in open source. 14:10.400 --> 14:15.040 And individuals are left trying to cope with conditions that according to psychology really create 14:15.040 --> 14:21.200 the perfect storm for burnout. The burnout already affects our self-worth. And if the only fix is 14:21.200 --> 14:26.800 available to burnout are individual fixes, then the risk that people will like Adam feel like they 14:26.800 --> 14:31.360 failed personally when things fall apart. And I don't know about you, but I don't feel like that's 14:31.360 --> 14:37.360 an individual burden for us to bear. So in order to solve burnout for good, we need system level fixes 14:37.360 --> 14:42.880 and system level fixes require us to do what open source does best and to collaborate. So before 14:42.960 --> 14:47.440 we wrap up, I'm going to argue for three such system fixes and I'm going to argue against 14:47.440 --> 14:52.000 the position that we should keep treating burnout as an individual issue only and then highlight 14:52.000 --> 14:56.080 some things that are important to consider from the perspective of psychology if you're someone 14:56.080 --> 15:02.880 who's working on implementing them. So firstly, we made it a bit taboo to talk about money in the 15:02.880 --> 15:07.680 room. I'm going to say that we need to make it easier to get paid. So the fact that developers aren't 15:07.680 --> 15:12.400 getting paid is making everything else worse. It's making the workload greater. It's making 15:12.400 --> 15:17.840 maintenance less rewarding, extrinsically, and it's leaving individuals juggling impossible competing 15:17.840 --> 15:22.720 responsibilities. I've heard people say it is a conference that open source isn't about money, 15:22.720 --> 15:27.520 it's about people, but I don't think there's a dichotomy here. I think that if we're leaving 15:27.520 --> 15:32.800 people struggling to cope, to make a living just because they've decided to work to maintain 15:32.800 --> 15:38.320 the comments and that's not respecting the people of open source. Now if we approach burnout as an 15:38.320 --> 15:42.800 individual issue, it might be tempting to say, well, didn't you sign up for this when you made 15:42.800 --> 15:48.640 your code open source? Why do you feel like you're being treated unfairly? But I don't think this 15:48.640 --> 15:54.080 is true. I think the psychology tells us that social behavior is organized by group norms, which 15:54.080 --> 15:59.520 are kind of like unwritten rules for the group. There's a norm in the open source community that 15:59.520 --> 16:04.080 yes, we give away code for free, but in return, we expect the community to collaborate and then 16:04.080 --> 16:08.240 everyone shares in the benefits and we produce something good. So it starts with developers 16:08.240 --> 16:12.400 sharing code with other developers, but then when your open source becomes more popular, 16:12.400 --> 16:17.280 more people outside the group come to depend on it, and this also creates more work, and these 16:17.280 --> 16:21.600 people who are further from the middle of the group don't always abide by those group norms, 16:21.600 --> 16:25.760 and these aren't the terms that a lot of developers signed up for when they made the code available 16:25.760 --> 16:32.240 open source. Plus, while making money is a taboo topic in open source, I think leaving the 16:32.240 --> 16:36.560 videos on their own figure out how to make a living while doing open source is arguably putting 16:36.560 --> 16:41.840 open source values at greater risk. So individual developers don't have much power to make a living 16:41.840 --> 16:46.320 out of their open source beyond restricting access to participate or making free immune features, 16:46.320 --> 16:51.680 which limits the possibilities for open collaboration. If we work together, we don't need to 16:51.680 --> 16:56.320 sacrifice those open source values, so together we have the influence to encourage, for example, 16:56.320 --> 17:00.960 companies that rely on open source to subscribe to those group norms and to give something back 17:00.960 --> 17:07.120 to maintain it. Now I did this research long with centuries open source pledged team, and I think 17:07.120 --> 17:11.360 this is a great example of trying to get companies to give something back to the community, 17:11.360 --> 17:16.080 so companies can become members if they pledge at least £2,000, sorry, $2,000 per year, 17:16.080 --> 17:20.800 put a developer at the company, and this aligns with what psychological research tells us about 17:20.800 --> 17:25.520 the organizing power of group norms, so encouraging companies to make their commitment to open 17:25.520 --> 17:29.840 source norms public makes them accountable to the group, so they're more likely to stick to them, 17:30.560 --> 17:34.480 and in turn these companies have then recognized and celebrated by the community for the 17:34.480 --> 17:40.000 part of playing in it, and they've raised over $3 million to maintainers in the last year alone, 17:40.000 --> 17:43.840 in fact they have to keep updating this slide because I look on the website and it's snuck up 17:43.840 --> 17:50.240 since last time. Secondly, we need to share the responsibility that's associated with open source 17:50.240 --> 17:56.240 software, so I think it might be tempting to apply this individual fix of learning to say no 17:56.240 --> 18:00.640 to taking on more responsibility than you can cope with, and this is really important, we should 18:00.640 --> 18:04.560 know how to draw boundaries, there are some things that were pressured to doing that we should feel 18:04.560 --> 18:09.520 no qualms about saying no to, but I think we also need more people who are willing and able to 18:09.520 --> 18:14.000 share in the responsibilities that can't be avoided, so we need to be able to grow the maintenance 18:14.000 --> 18:19.920 community, so we sometimes talk about this in terms of the bus factor, I don't know if anyone 18:19.920 --> 18:24.480 if there's overlap between the Mean Girls fan community and I'm just a developers, but there it is. 18:25.840 --> 18:30.800 I have a word of caution about the bus factor, which we used to talk about how to increase the 18:30.800 --> 18:34.480 number of people working on a project so that it's less likely to fail if someone played the 18:34.480 --> 18:40.400 methodically and gets hit by a bus. The bus factor treats developers as if they're interchangeable, 18:40.400 --> 18:45.520 which misses out an important qualitative information about those individuals and about team dynamics, 18:45.520 --> 18:49.520 so the team might have many people who have the skills to step up, but one person still ends 18:49.520 --> 18:54.400 up doing all of the work because they're the expert that the senior is the path of least resistance 18:54.400 --> 18:59.200 to leave it to them, or they might have difficulty trusting junior people to step up, and these 18:59.200 --> 19:04.480 individuals are at risk of burnout. So what's the solution? We've talked a lot in this room today 19:04.480 --> 19:09.040 in a really good way about mentoring relationships, so these help collaborators build the confidence 19:09.040 --> 19:14.080 that they need to contribute and help senior developers build the trust that they need to pass on the 19:14.160 --> 19:20.320 reins. And secondly, larger projects could benefit from governance than the next everyone's roles 19:20.320 --> 19:25.440 and responsibilities clear, so unclear expectations lead to a diffusion of responsibility and they 19:25.440 --> 19:30.960 are themselves associated with burnout. But this is also something of a balancing act. We don't want 19:30.960 --> 19:35.840 roles and responsibilities to be too rigid either because you remember that autonomy, so freedom and 19:35.840 --> 19:42.320 control over how we work protects us against burnout. So here's a really useful thing from the 19:42.320 --> 19:48.720 psychology literature to bear in mind. When work has hardly any autonomy and lots of responsibility, 19:49.280 --> 19:53.840 or it has lots of autonomy and vital responsibilities, so the pink danger zones in this graph here, 19:53.840 --> 19:59.200 then there's a risk of burnout. Open source maintenance just is a big responsibility, so we need to 19:59.200 --> 20:04.960 balance it out by making sure we preserve autonomy. We want to be here. So if you're working on 20:04.960 --> 20:09.680 models of governance or if you're working on different models for funding, then it's important 20:09.680 --> 20:13.200 to make sure the model you're working on doesn't prevent developers from having freedom and 20:13.200 --> 20:18.400 control over how they work. I think this is a risk if you have a model that's like a BDFL or 20:18.400 --> 20:22.480 if you're too reliant on one source of funding so they have power over the direction of the project. 20:23.440 --> 20:29.680 And finally, I'm getting close to any out of time, we need a kind of community dynamic, 20:29.680 --> 20:33.680 and I think this will get better, the more better paid and the less overwork developers are 20:33.680 --> 20:38.640 remembering the burnout spiral. However, I think they're still an underlying culture and open source 20:38.720 --> 20:42.960 that's permissive or rudeness in public under the guise of Frank and Open Dialogue. 20:43.680 --> 20:46.800 If we take an individual approach to this problem, we might say, well, you need to develop 20:46.800 --> 20:51.680 thick skin in order to work in open source, you need to be resilient. And I think this is important, 20:51.680 --> 20:56.640 but also it's somewhat fatalistic because it's asking us to adapt rather than believing that 20:56.640 --> 21:01.440 changes possible. And secondly, I think this would mean asking some people to develop thicker skin 21:01.440 --> 21:06.000 than others because groups that are already marginalized in open source are already bearing the 21:06.080 --> 21:10.560 brunt of more of this toxic community behavior. So I don't think this way of thinking encourages 21:10.560 --> 21:17.360 diverse participation in open source. Sorry, so I've clicked on those earlier. So a less individual 21:17.360 --> 21:22.720 solution would be one where we build a sense of community that helps us feel solidarity with one another, 21:22.720 --> 21:27.040 and where we respect each other rather than disrespecting each other because we've become 21:27.040 --> 21:33.680 used to being treated with this respect. This is an easy thing to say, a harder thing to do, 21:33.760 --> 21:37.520 but psychology tells us that there's a role that community leaders can play in this. So if you 21:37.520 --> 21:42.160 or someone people in the community look up to, you can model the sort of behavior that you'd like to see 21:42.160 --> 21:50.480 in open source. So you could be like Adam Wathon and be courageous and candid about your experiences, 21:50.480 --> 21:56.800 talk about your feelings. Secondly, it's important to have kind of grassroots community spaces 21:56.800 --> 22:03.600 where people can come together and talk about the problems in open source. So, oh dear. And 22:03.600 --> 22:08.640 I'm going to go over by like one minute if that's okay just so I can finish up. Thanks. So this gives us a 22:08.640 --> 22:12.880 greater opportunity to build a sense of community and build solidarity with each other rather than 22:12.880 --> 22:19.520 fighting because we're all suffering by lack of funding and burnout. And to sum up, we should care about 22:19.520 --> 22:23.680 burnout because it's an ethical issue and issue of fairness and justice and an issue of sustainability 22:23.680 --> 22:29.120 and security. Open source culture is built on recognizing the value of collaboration and there are 22:29.200 --> 22:33.600 many actions we can take the last year get to the root causes of burnout that can only be 22:33.600 --> 22:38.480 achieved by us working together. So thinking back to that leaky robot, we need to stop leaving 22:38.480 --> 22:43.440 individuals to try and cope as all the energy drains out of them and we need to get back to whatever 22:43.440 --> 22:48.240 our equivalent to happy waving dancing is. So get back to that creative coding that gives us life. 22:48.960 --> 22:52.320 I thank you so much for listening quickly before we go to the Q&A because there's a sort of 22:52.320 --> 22:57.600 been a call to action. I've started making a list of people working on system fixes to burn out 22:57.600 --> 23:01.920 in open source. That's the top QR code here. So please take a look. Tell me if there's anything 23:01.920 --> 23:06.240 I could add and see if you can help any of these people out. And secondly, this was just a bit of 23:06.240 --> 23:10.960 research I didn't my own free time alongside my PhD and I'd love to be able to expand it. So if you'd 23:10.960 --> 23:15.440 like to talk to me about your experiences of burnout or if you'd like to collaborate with me on 23:15.440 --> 23:19.440 similar work in the future or if you know someone who's willing to fund this kind of research, 23:19.440 --> 23:22.960 you can come talk to me at my website at the fewa here. Thank you so much. 23:27.680 --> 23:35.680 We might have to have for one question. Does anyone have a question from a rander? 23:42.880 --> 23:46.960 Thank you, Miranda. It's a really important subject address. I think actually it even applies to 23:46.960 --> 23:51.680 people who work commercially in software as well. One thing you talked about responsibility 23:51.680 --> 23:56.960 house really important, but people are made differently and you can't turn people into what they are. 23:56.960 --> 24:01.840 I wonder if there's a specific role in communities for if you like the salesperson, 24:01.840 --> 24:06.560 the marketing person, just like in my company, my engineers get beaten up, they come to me 24:06.560 --> 24:12.320 because I can put on my salesman hat and beat the customer back up again. And there are people who 24:12.320 --> 24:17.200 look good at that and whether you need that or that's the sort of champion you need. Someone who's 24:17.200 --> 24:21.920 not just setting a role model but is actively keeping people off other people's backs who aren't 24:21.920 --> 24:26.960 so good at that stuff. I do think that would be incredibly useful and I don't think everyone who 24:26.960 --> 24:31.760 is involved in maintenance has those kind of skills or finds that something that's easy to do 24:31.760 --> 24:35.760 or that they enjoy doing. Somehow if we could bring in more people who are good at that 24:35.760 --> 24:39.520 people management side of open stores, even if they're not so strong on the coding, 24:39.520 --> 24:43.360 bringing in people who aren't just programmers to help out. We had a little bit of talking 24:43.360 --> 24:46.960 about at the end of the book. I think that would be really helpful. I think that's what I 24:46.960 --> 24:52.480 meant by like being careful not to treat programmers as interchangeable, like we will have the same 24:52.480 --> 24:58.240 kind of skills set or find the same things motivating, such as isn't true. Thank you, Miranda.