April 29, 2016
Dev Tools Digest – Apr 29
This week's Dev Tools Digest features news from Codenvy, Gradle and Pagerduty, with event recaps from Iron.io and Keen IO.
In this episode of To Be Continuous, Edith and Paul discuss ‘crunch time’, work-life balance in the startup world, and the efficacy of the 80 hour workweek.
Paul Biggar: Today we're talking about working hours.
Edith Harbaugh: Yeah, there's a couple of really interesting articles in Venture Beat recently. One from the president of the game developers' conference, talking about how crunch time was an issue.
Paul: A bad thing.
Edith: A bad thing. And then the second response saying that crunch time was something that people should not only expect, but relish.
Paul: Okay, I would be interested to see why should you expect and relish crunch time.
Edith: The second article was basically saying people who work in games should be passionate about their games, and as such, they should expect to be working basically 70 to 100 hours every week.
Paul: Gotcha. I can't help but feel like Marx would have something to say about this that would be particularly relevant.
Edith: Well, it's funny because this guy, the second guy, actually took a very Marxian slant. He's like, "You shouldn't be a wage slave. " And then he twists that argument back up on itself. He's like, "You shouldn't be a wage slave, you should expect to work really hard, because you consider yourself an owner of your company."
Paul: I see, except you're not an owner of your company. As an owner of my company, I feel that when people complain that there's no mission involved, that I disagree strongly.
You should be working at a place that is mission driven, and personally, I'm a fan of work-life integration.
Although, Circle ended up being much more of a work-life separation. But I'm also not particularly sympathetic to the idea that 70 to 100 hours a week is anything that's justified. Reading through that article, it seemed like pure exploitation. And the games industry in general seems like pure exploitation.
Edith: Well, the slant of the article was like, "Well, you love games, so you should love games so much you spend all your time on them." But I think that's very shortsighted for a couple different reasons. It assumes that, let's say hypothetically, you work 100 hours. You're not going to be productive all 100 hours. I have a friend who was working at a small startup, and they said, "We expect you to work 80 hours a week."
Paul: Okay, and then he quit the next day, I assume.
Edith: Well, he said they expect you to work 80 hours a week. And we know that the last 40 will not be as productive, but that's fine. We're still getting value.
Paul: I think that it's ridiculous to suggest that there's no value in the next 40 after the first 40, right? The 41st hour is almost as good as the 40th. The 42nd is pretty good as well, and you're definitely not going to get twice as much.
The problem I have there is the idea that there's an entitlement to your time.
I think there's some entitlement to your time, right? That's what you're paying for. But there's a reason we ended up at 40 hours a week. And in the Valley, there's an expectation just in tech, in general, that developers work 50, 60 hour weeks, and everyone works 50, 60 hour weeks, that's kind of on the upper range of what I think is reasonable. But to suggest 80 hour weeks is exploitative. It's pure exploitation.
Edith: Yeah, and then, you know what happened, my friend quit that job.
Paul: Good. The next day, I hope.
Edith: Well, no, it was sad. He was like, "Let me try this for a little bit," you know what I mean? Because he used to go home and code at night on a side project. So when he's like, "I'll just stop coding on my side projects," he tried it, and he said, "I can't do this anymore." I think these companies that try to do these crazy hours, they just have a different mentality of, "We will grind out people."
Paul: Sure, and you see this in Amazon a little bit. Amazon had that culture document. Generally, the view that I've seen of Amazon is that they grind through people, and some people survive in that. I think they would describe it as high pressure, but let's just call it exploitative environment. And some people don't.
Edith: Yeah, I disagree with that fundamentally. I read a good response, it said, basically, "Great, you want me to work 100% more hours than my salary originally was supposed to be for. I don't get double pay."
Paul: And it's funny because there's already a disproportional leverage that you get from working in software, so often the person producing the code is producing 10 or 20 times the value that they're actually being paid for.
Edith: Yeah, but the thing is, when you're working 80 hours, I used to work those crazy hours.
Paul: Wait. You have a startup. You're not working 80 hours a week?
Edith: I mean, crazy hours.
Paul: I need to tell your investors on you.
Edith: I'll tell one story, and then I'll tell a second story. You know, I used to get up, I would go to work, I would work until the BART stopped running, which was midnight. I would go home, I would sleep six hours. I'd go back to work, and then I would come in and I would work on Sundays. I got a lot done, but it was miserable.
Paul: The reason that we're referring to it as exploitation is that it burns you out, it's completely unsustainable. It affects you physically, it affects you mentally, and you're miserable. If people worked 80 hours a week and were happy, then I don't think we'd call it exploitation.
Edith: IBM in Austin was notorious for this. Austin was a place where they actually filmed Office Space. Have you seen Office Space?
Paul: Of course.
Edith: Well, I never know, you're Irish.
Paul: We get movies there, too.
Edith: The joke in Office Space is "Hmm, you're going to have to come in on Saturday." But it was true. I had friends who worked at IBM, and they would get a memo like, "We have mandatory Saturdays."
Edith: Like, "We will check that you're at your desk." Yeah, and my friend said it was just a miserable time because everybody had to be there.
Paul: Were they working or were they just there?
Edith: Well, at the end they were just like, "Fuck this." Because literally they were told, "We will go around and we will count who's at the desk."
Paul: There's a practice that I do agree with, which is similar to this, or sort of looks similar externally to this, and that's like a work-life integration. I think that, especially when you're in the early years of your career, and let's say you're working at a startup, you're probably going to love your job. You're probably going to be like accomplishing things and learning at this incredibly vast rate, and you're probably going to really benefit from working 60 to 80 hour weeks.
Edith: The sad part about me, back when I used to come in and work on Sunday, part of it was that I lost touch with everybody who I didn't work with. I'd literally be like, "I don't have anything to do today because I don't have any friends anymore.
Paul: This is why I refer to it as work-life integration, because I think that the people who work for these companies often, you see a lot of people moving to San Francisco after college, and they make friends with everyone they work with. So they work with them, they hang out with them in the evening, they go out partying with them on the weekends, they go hiking together on Sundays.
Edith: It was like that except for the not working part.
Paul: Okay. You just worked all the time.
Edith: We just worked all the time, literally all the time. It was ultimately very depressing to look back on.
Paul: This is the problem, that there's versions of this which are awesome, and there's versions of this which are not awesome. And whether they're awesome or not is often a function of where you are in your life at that time and how you want to be.
Edith: Yeah, and I think also, as you said, I'll tell your investors, I think there's definitely a time in every company's cycle where they do have a crunch time. When you are like, "Here is when I need to put my foot on the gas and move." But I think every company has that time when they're like, "We need to work a little bit harder." So, that's the original article about crunch time, is then you go into perpetual crunch time.
Paul: Right, perpetual crunch time is ridiculous and soul destroying. It should probably be illegal.
Edith: But I 100% agree that if you're in a startup, you have to have times where you're like, "Hey..."
Paul: Right, so sometimes, there's a deadline that you have to hit. There's a demo coming, there's a presentation, there's a launch. You see a lot of startups that have moved away from the idea of road maps and launches and deadlines and that sort of thing, so it's like, "We'll ship it when it's ready." GitHub is a good example of this, or GitHub in the old days was a good example of this. I'm not sure if they still prescribe to that philosophy. And I think that that sort of thing makes a lot of sense, but sometimes there is a deadline for some external reason, and you have to hit it.
Edith: I think, for me, not to talk about me, I firstly feel the weight very much, you know? Because I'm the founder. I always feel like there's something more I could be doing
Paul: Right, and you feel, you feel like physical pain, or at least I did, at something that's wrong with your product or something that, isn't shipping or a customer emailing you with a problem. I tell you, it hurts to feel that sometimes.
Edith: Yeah, there's always more you could be doing. So it is very hard as a founder to not burn out.
Paul: Right, especially the first few years as a founder are really tough. After a certain number of years, or once you hit certain milestones with the company, it has enough people or you've farmed out your role to enough other people that you can take a little bit of a break. But I know founders who are four or five years in who are like, "Now I only work 80 hours a week."
Edith: Yeah, I mean, it was really nice. I've been doing all our sales, and we just have too much for me to handle.
I literally can't take a break because people start emailing me, like, where's our quote? Like, we want to buy. So I was very, very happy because we hired our first salesperson.
I can finally, when I go to a conference, I don't have to basically work a double shift at the conference, then go home and you work another eight hours. So, you wrote a pretty inflammatory article that got a lot on Hacker News.
Paul: This is the NewsTilt one, from six years ago or something? That got a lot of flame, but the article was called "Why We Shut Down NewsTilt," and it was the company I did about six years ago. I was in a space I shouldn't have been, all sorts of problems with it.
Edith: I actually really liked that blog piece.
Paul: Oh, thank you.
Edith: It's about NewsTilt and the lessons that Paul learned.
Paul: Right. Inherently, it was a blog post about failure. "Why didn't we get it? Why didn't we make it work? Why didn't it work? What was wrong with it? And what did we learn from it?" Remember, at the time, a lot of people said, "You shouldn't publish this."
Edith: Did you think it was because now everybody's like,"One of the pinnacles of Maslow's hierarchy is to publish a media piece about your failure."
Paul: Right. I'm a failure hipster. I failed before it was cool. One of the reasons I'm very happy to have published this is I get emails every now and then, they've kind of died down, but at one point I was getting one every few months or so, of people who were saying, "My company's not doing too well. I don't know what to do," just like feeling this massive amount of stress.
They're in perpetual crunch time, but they're sort of banging their head against a brick wall, and one of the things I feel about crunch time in general, and burnout in general, is that burnout is the result of not seeing progress. And the way that relates to crunch time is you are expecting to get to the end of it.
You will have a launch, you will have an event, and then you will take two days off and you will go whitewater rafting or something. And then you'll come back refreshed to start a more meaningful or slower pace or normal pace sort of thing. If you can never see the end of that crunch time, then burnout is inevitable very quickly.
Edith: We had these crazy releases where I was working these crazy hours. I basically lost touch with all of my friends because if somebody asks you three times to do something, every time you say no...
Paul: Yeah, then they stop asking you.
Edith: Yeah. And so, finally, I just literally took some time off. I was like, I need to take some time off. It was very helpful. So I think there's nothing wrong with crunch time. I think what's wrong with it is the perpetual crunch time, and I think that's just management basically not recognizing that they need to invest in people, or not caring, or I don't know.
Paul: I think there's this idea that crunch time is necessary, and I view crunch time as a failure in the release process.
Edith: I view it as a failure of management to plan.
Paul: Right. And the people who get punished are the people who didn't cause the problem.
Edith: That's part of why I'm such a big fan of continuous delivery, because what used to happen is
the reason we would have crunch time is we would have this massive year-long release, and the longer it took, the more stuff we wanted to put into it.
Paul: Right, the more crunchy it gets as a result.
Edith: All our competitors are moving forward, and it's like, "Well, we haven't done a release for a long time, so it needs to be even more." And it got to the point where my manager actually had three schedules. He had a schedule he would show to marketing, where everything was fine. He had a schedule, which was engineering's, which everything was completely fucked up.
He basically kept these parties very carefully separated. I mean, he had a complete mental breakdown. He was trying to show everybody that everything was okay. At marketing, he would say, "Everything's perfectly on schedule." To engineering, he would be like, "We're completely fucked."
Paul: Surely this is a disaster.
Edith: This was a disaster.
Paul: I thought this was being held up as like, he had carefully managed that.
Edith: No, it was a complete disaster because marketing kept trying to squish more stuff in, and engineering is like, "What the fuck?" Eventually, the guy got fired and had a mental breakdown.
Paul: In that order?
Edith: No, a mental breakdown and then fired. It was a bad situation.
Paul: Right. I have a friend who's in the games industry, and she's going through this sort of thing. There's a release coming, and so you're working late, you're spending time, a lot more time, and my sense on it, in some of the things that I've heard is, does that really need to ship the first time? The games industry is moving to, or in many cases has moved to, a model where patches are expected. A lot of work is being done in the server anyway.
You're no longer putting something on a gold disk and shipping it for a certain date.
Now, something very often still goes on a gold disk and ships at a certain date, but if the last three levels of the game didn't ship on day one, and instead were included in the thing that went out a week later, probably no one would notice.
Edith: That's what we end up doing on the project I was talking about. We had a lot of delegated administration basically allowing people to pass permissions on to other people. We just decided to not ship that at all, the first release. Everybody was a super admin.
Paul: Exactly. You cut things that you need, and everyone who is in the product market fit, minimal viable product part of the world sees that, right?
Edith: This was in 2000.
Paul: Wow, okay. Well, impressive.
Edith: That was kind of my first, that's why I was such a convert to lean and product market. We were killing ourselves to build something before we even really knew the market.
Paul: Games have this strong idea of what they're going to ship, right? They know already, because there are well-defined genres. They do a lot of play testing and they do get one launch event. So that there's a feeling that things have to be probably a lot more perfect than they really have to be.
Edith: I think that's interesting because you could really decompose games, because games take a while to progress through levels. A really interesting lean game would be to just build the first three levels.
Paul: And many games don't even have levels, right? You continually play Bejeweled or you're in a multi-player MMORPG, right? And there's these things are released late, but there's also game dynamics that are released late.
Edith: I mean, Zynga was the king of that.
Paul: Exactly. Every game today has economies that they experiment with, just lots and lots of experiments, and so they could cut a lot of what they're going to ship. And they do, I would argue, maybe not enough.
Edith: Zynga's really interesting. Part of why they were so successful is their first thing was just, "Can we draw people in this game at all? And then let's start experimenting on the game mechanics. Can we get people to show up to play?" I think they took advantage of a lot of Facebook loops that just are not available anymore, mainly because Facebook is like, "Why are we leeching all this money to Zynga instead of keeping it?" But we drifted far away from your post-mortem, "failed before it was cool."
Paul: I think the thing that you're pointing to in that blog post is something I said that was very controversial that people did not like. And that was there was an engineer that we were looking to hire, we had just met, we had just started working with him, we're within a week of starting to work with him.
Edith: So you'd hired him?
Paul: It was either hired or on the cusp of hiring or in that sort of nebulous first employee, are we working together or are we not, sort of period. This was back when we had very little funding. It was much less concrete than things were later. And so I said in the article that it was a Sunday, and he wasn't willing to work or he didn't want to work or something like that.
The point that I had been trying to make was he didn't seem that into it. He didn't seem that into the startup, didn't seem that into the product, and as an example, I sort of summarized to say it was a Sunday where we were all in the same place, and me and my co-founder are working on it. He was there, but he just wasn't really into the product. He wasn't into the startup, he wasn't into the whole thing, and the people who complained about what I said, like, "Oh, you shouldn't be expecting people to work Sundays," and all that jazz, and I agree with those people.
I definitely come sort of down the middle between the people who are staunchly for 40 hours a week and no one will ever think about their employer outside those 40 hours, and the other side, the people who think crunch time is necessary. I'm down a practical middle of you shouldn't have people burn out, but at the same time, I generally prefer people who are on mission. And day one of a startup mission does involve a little crunching, usually.
Edith: I agree with you, as I always do. Crunch time should be for emergencies and should be avoided at all cost. Not all cost...
Paul: Not all cost, yeah. There is a cost. I mean, it's sleep debt, it's burnout debt. It's one of those things.
Edith: Yeah, it's interesting. I was going to say this might be a tangent, which means it is a tangent. I'm kind of burnt out at running. I just decided I was running too much.
Paul: Well, you did 50 miles last week, right?
Edith: I did 50 miles.
Paul: That's a lot of miles.
Edith: It wasn't really the physical cost of it, because I rode my bike to work. It was just like, wow, I've been on my feet for 13 hours. I would rather be spending that time doing something else.
Paul: Right, and I guess when you're running, there's all these work thoughts going through your head. It's like, "Oh, I really have to send that email, I really want to ship that product." You have the idea in your head, and then by the time you get there, it's like 13 hours have passed and you're exhausted.
Edith: Yeah, I was. I would be much happier if I had gone running for four or five hours and then worked for eight hours.
Paul: Yeah, I mean, that's still a lot of hours.
Edith: I guess what I'm trying to say is my work is now more exciting than my running because gosh, I want to get all this stuff done at work.
Paul: Well, it's a good time when that stuff is happening. There's some part of doing a startup where it's like, I don't want to really do anything else other than address the thing I need to address.
Edith: I guess there's nothing like running 50 miles in the rain to make you want to go back to work. I think as always, for an agreement, are there any angles we haven't looked at for this crunch time and how it fits back into continuous delivery?
Paul: I think there are two things. One is bad management.
Edith: There's two factors.
Bad management and bad planning. And if you combine those two, it's a furnace.
Paul: Let me try that again. There's a bad culture, all right? People who have a culture where sometimes we work late to ship a product. I'm totally fine with that. But anyone who's espousing crunch time as a permanent state of affairs, or looks down on the "peons coding in their caves" to produce more product, and there was one that I saw recently. Do you know Alex St. John? Yeah, so he had a blog post, a presentation.
Edith: Yeah, that was...
Paul: Wow, I know, right?
Edith: Well, wow on so many levels.
Paul: I read Alex St. John's stuff in the '90s when he had the back page in PC Gamer or PC Hardwire or something like that, right? He was talking about DirectX and what it did with the games, and I was like, oh my god, this guy is amazing. And he almost made the template for what would become the entire Evangelism space. But when you read that thing, you're like, holy shit. I'm amazed anyone would say this in public. I mean, some people think it, I guess, but to put it in a presentation?
Edith: Well, why did you think it was so retro? Because like, everything about it was stereotypes.
Paul: Yeah, it was just pure stereotypes. I mean, the obvious one is "Get the girlfriend or wife onboard." Assumption one, that the engineer is always male.
Edith: Male and straight.
Paul: Male and straight, so that may be a thing which is still statistically valid, but I think we're certainly at a place in the industry where it's not okay to say that anymore, or to perpetuate that stereotype anymore. And then the idea, "We're going to hire these little aspy kids, and we're going to put them in a cave, and we're going to work them until they quit. Here's how to offset the economics of hiring a more experienced person. Well, we'll just give them five more kids.
Edith: The beginning story, when I talked with a friend who was working 80 hours, he would hire kids out of Waterloo. Waterloo has a co-op program, Waterloo is a university in Canada, so he would hire 10 college kids who really didn't know much, with the theory that with the 10 cheapest kids, he could reproduce one experienced person. And really what you have is kids.
Paul: One experienced person is much cheaper than 10 kids.
Edith: Yeah, and what you have when you have 10 19-year-olds, is chaos.
Paul: It's interesting because I think the games industry is one of these places where kids grow up really being into games.
Edith: That was the point of the article. "Well, if you love games, then you should put the time in."
Paul: If you love games, that doesn't mean that you should be exploited. Because people love games so much, it makes them ripe for basically being exploited. I did some time in the games industry just when I finished college, and I saw the pay is much, much worse. The working hours are fucking terrible, but you do get a lot of cultural benefits. It's a lot more casual and what you're looking for, and I think one of the great things about San Francisco and Silicon Valley is it has all the upsides of that casual culture, without exploiting you.
Edith: For our own company, I've tried to follow the advice. I had an ex-boss named Ann Duane. She's brilliant, just absolutely brilliant. One of the best things I did was work for her. I remember when I showed up to work, it was my first week, and I'm like, " Ann, what hours should I be working?" And she said, "I don't care." I'm like, "What?" And she's like, "I'm very results driven. Increase our revenue by 15% this year, 20% would be better, and I don't care what hours you work." I was like, "Okay, and I went off and I increased her revenue by about 20%." I think actually 25%, but she's like, "I don't care if you get here at 8:30 or 9:30. That's not my bag." She's very formal, so she didn't say "that's not my bag."
Paul: Yeah. I have a friend who has mandatory office hours in his company, and the hours are something like 10-7 Or 9-7 something like that. Yeah, I know. I know what you're thinking, right? That's, like, everyone should be in the office noon to two, so we can collaborate and have meetings too and that's all right.
Paul: Right, and I'm amazed, but it appears to be working for him.
Edith: In terms of?
Paul: He's able to hire people, the people are high morale, they're delivering a lot, they're enjoying themselves.
Edith: I mean, I personally don't know if I would enjoy that.
Paul: I don't think I would either, but I think he's also looking to hire a persona that he's looking for, which is sort of the Bain analyst kind of persona. And that's his background as well, and I think they work well in that kind of culture, at least that's my understanding. But I would be out the door.
Edith: For my company, we have people who have kids, and sometimes they take every Wednesday morning off, which is great. And they work Sunday afternoon. You might say, "Oh we're exploiting them, because they're working Sunday afternoons." It's like, no, you get some time with your family on Wednesday.
Paul: There is a down side with that, in that most people I know, if you ask them to do something on Sunday evening, it's like, "Oh, I can't. I've got a bit of work to do." And so that's like this refrain that I hear everywhere. Because, certainly, people take a little bit of time off to do their personal stuff, but the work becomes all-encompassing. Even just the fact that you have email or that people speak to each other on Slack; You check your board, you check your work email, you start to think about work, you know?
Edith: I'm really bad.
Paul: Yeah. People don't get as much time off, they don't get as much time to clear their heads and think about things that aren't work. And some people enjoy that, and even the people who enjoy that feel stress, and burnout as a result of it. Or it contributes, at least.
Edith: I think Slack was supposed to save us, but it's made things worse.
Paul: I don't think it did, yeah.
Edith: It's this always-on thing.
Paul: It's this always on thing that you feel that you constantly have to keep up with, as well. You feel you have to read the back trace.
Edith: And it's not threaded. That's the thing that kills me. For us it's like, we're discussing a support ticket, and then there's a bunch of gifs, then we start talking.
Paul: Yeah, there's no end to the conversation. So I think threading is a reasonable technical solution, but there's just a problem, not just in our industry, but in our lives, the fact that we're carrying a computer around with us all the time. And if it's not work that's pinging you, there's just lots of messages that you have to reply to and lots of things you have to do.
Edith: Your co-host telling you it's Tuesday.
Paul: I turned off my email notifications about three years ago, on my phone.
Edith: Oh, that must have felt so good.
Paul: On my phone, I turned off all notifications about a month ago.
Edith: I've just started to try to wean myself into turning on "do not disturb" at night.
Paul: There's a setting on your iPhone where you can turn on do not disturb automatically, and so I have it go from 12:01 a.m. until 12 a.m. so it's just a permanent do not disturb. And if I turn off do not disturb, I'm waiting for a call, waiting for a text, you know, whatever. It'll be able to get turned on again at night. Yeah, it's great. I highly recommend it.
Edith: This is why you look so relaxed.
Paul: I think it is why I look so relaxed. I mean, I also gave up sugar for April.
Edith: I thought it was the LaunchDarkly shirt that made you look so cool.
Paul: Well, thank you.
Edith: But it's really the do not disturb.
Paul: It's either the sugar or the fact that I'm sleeping well or the fact that I turned off the notifications. I'm not sure which.
Edith: Yeah, well, all of it. All of it. Or the haircut.
Paul: That could be it, too. Or the fact that I'm not in perpetual crunch time.
Edith: To tie it all back to continuous delivery, let me go away from continuous delivery.
I think the combination of poor management, poor culture and poor planning, sometimes cause a bonfire of crunch time.
Where it's basically this culture of like, well, people work 80 hours a week. What if they did that all the time?
Paul: Right. And, I generally condense it to planning as the problem here. That with better, I don't want to say PMing, but just choose better what you ship. You probably don't need to spend all that time.
Edith: Or refuse to have a culture.
Paul: Or the two work together nicely.