September 19, 2018
Ep. #48, Resolving Accidental Complexity with Dark
In episode 48 of To Be Continuous, Edith asks Paul what it took to found his fourth startup, Dark.
In episode 57 of To Be Continuous, Paul and Edith explore the vast differences between the devtools community of yesteryear and today, from the tooling to the business ecosystem, and even the developers themselves.
Paul Biggar: So what is devtools? Do we have any idea what devtools is any more?
Edith Harbaugh: Is that grammar?
Paul: I think so. I didn't hyphenate it, is that what's wrong?
Edith: Isn't it with devtools are?
Paul: No, devtools itself is the word and that's a singular. I'm not saying I don't know what developer tools are, I'm saying I don't know what the devtool space is.
Edith: Well, that's why I was trying to clarify your grammar.
Paul: Okay, yeah, that was ambiguous, I see it now.
Edith: Usually, you're very precise.
Paul: I was very precise.
Edith: Okay, for that home audience can you explain what you mean?
Paul: So we're in, I don't know, the fourth or fifth wave of devtools. Or at least in my--
Edith: More if you count Visual Basic.
Paul: Yeah, so I'm thinking in my own memory, so like, I'm counting like the first wave of devtools companies as GitHub and Heroku, which sort of shows my age there's definitely like BA, WebLogic and Pure and that sort of thing in the '90s. And I know nothing about them.
So we'll just say, starting in like the mid 2000s and there's this huge ecosystem and stuff built around Heroku and GitHub.
And then the next generation, I'm going to claim the CircleCI.
And the tools of around that time, and then I know there's been three generations since and there's devtools popping up everywhere and everything is a devtool. And like YC has a devtools category.
What has changed as well as the existence of them is the marketing channel, the people who are developers has radically changed in five or six years, how one sells to developers, how one thinks of a developer, it's all completely changed. And it makes having a devtools company kind of challenging.
Edith: I have many thoughts.
Edith: I'd say first, this is the golden age of having a developer tools company.
Paul: Sure, sure.
Edith: Developers only have access to budget. Developers can use SaaS tools. There's all sorts of awesome supporters like Heavybit that have a ton of great content, just on how to build a developer tool company.
Paul: Is all the content still relevant?
Edith: I don't know if it's still relevant but there's kernels.
Paul: For sure, for sure.
Edith: Like even in podcasts we did four years ago, it's something that I'm like, that's not relevant some of it is brilliant. I still go back and watch occasionally Jason Lemkin did a talk here on enterprise sales.
And I go back and like this is gold. Javier Soltero did a Heavybit talk on board meetings, brilliant.
Paul: But I mean, those probably apply to non developer tools as well.
Edith: I think for a long time, it was kind of a wasteland for developer tools, because there's a belief that developers would never pay for them.
Paul: Right, right. I still have my email from Paul Graham saying developers don't pay for things.
Edith: Have you blown it up and framed it on the side of your wall?
Paul: No, just because I'm depressed about that whole part of the world.
I used to look up to Paul Graham now I don't, and I just find every time he tweets, it's just like fucking eye rolling. And so let's just move on from that.
Edith: Great, we'll just move on. What's really flattering, there was a company which is starting in developer tools company.
And they emailed me and they said LaunchDarkly has shown that one can really build a successful developer tools company.
Edith: And that's part of why we're starting our company is that you've shown us this way.
Paul: And you're like, are we really developer tools? Is that--
Edith: Oh, no, I always just had this forum happy, pardon me, I heard that it had really nothing to do with the chili I just eaten, it was warm.
Paul: It's lovely. It is nice to get those emails.
Edith: Yeah, so I think to put it in perspective, like five years ago, there was this perception that developers will not pay money.
Paul: Yes. I think that wasn't true five years ago. It was true 10.
Edith: There was a perception or it wasn't true?
Paul: I believe there was still a perception. Five seems a little stretch for me, so I might quibble and call it seven.
So I think there was still a perception, but it was untrue. Whereas it was actually true, like 10 years ago.
Edith: Yeah, I think let's agree on 10 years ago. 10 years ago is like, well, it's open source.
There's just this repetition of the word open source, open source, open source ever like this is incantation.
Paul: People were making products out of things, and everyone was claiming that should be a library.
Edith: Yeah. We'll just open source this, so nobody will ever pay.
Edith: And I think what the last, depending on whether you want 10, seven or five years has shown is that people will pay for developer tools.
Paul: I actually think that people won't pay for developer tools.
I think that people pay a lot less than you think. And the people will pay for services but not the tool.
So CircleCI gets paid basically for infrastructure. And there's a subset of the money that we make where we sell to enterprise who do pay for the software itself.
But I think that for the most part, people are paying for a service and not so much for tools.
Edith: I think we're kind of dancing on the head of a pin right now.
Paul: Yeah, I mean, it's a delicate thing of like, what is the thing? But people, I don't think would pay LaunchDarkly for the LaunchDarkly library.
What they pay for is the fact that you run and operate this service for them and it is cheaper to pay you to operate the service than it is to operate the service themselves.
Edith: Yeah, absolutely. I mean, I agree with you and I'll even expand.
I mean, so what people pay us for is that we have an always up API, that it just works.
That we have libraries for the languages they want. And that we're continuing to invest in that platform.
Paul: Right, and I think that's a really important part to continuing to invest in that platform is having your tooling be sustainable.
We talked about this in the open source episode. It's very hard to have your tooling be sustainable.
But the people don't want to pay for a thing where they don't actually get the value.
So it has to be that the tools that we make are packaged in a way where people feel there's actual value being provided such as hosting.
Edith: Yeah, absolutely. I mean, as an engineer, I love selling to engineers, because they're very honest.
Paul: That's very applicable.
Edith: You know, so they'll say stuff like, and I take this as a compliment, they're like, "Oh, we could build about half of what you have. "But it would take as much time "and then we know you're going to keep building other stuff "that we can't keep up." And I'm like, great.
Paul: I mean LaunchDarkly is like if statements as a service, right?
Edith: Yeah, but a lot of really fancy if statements are layered together a trillion a day.
Paul: Yeah, a trillion a day now?
Edith: A trillion with a t.
Paul: Wow, that's amazing.
Edith: Yeah, so sure the LaunchDarkly that we started off with six years ago, somebody could build that. But the LaunchDarkly we have now, that is very hard to build.
Paul: Yeah, have a very similar feeling about Dark. That like I mean, I think actually it would have been hard for people to build what Dark was even several years ago. But there's a lot of pressure.
People want it to be like open source and they want to avoid vendor lock in, and I think those are very, very valid concerns. But of course, the thing that people need more than anything is a sustainable ecosystem. And sustainable ecosystem takes revenue.
And so revenue has to be like a business model has to be attached to anything that you want to keep alive.
And more than anything we want Dark to succeed, our customers want Dark to stay around.
And it is entirely plausible that we choose a business model which just mobilize that. And so one has to be very careful that they choose a business model that does.
Edith: Yeah, I remember the early days of the company, people would ask for an on prem version.
And they didn't really understand it 'cause it didn't make much sense. And they said, well, we're worried you're going to go out of business.
And we don't really get that questioning anymore.
Paul: Yeah, we had a similar thing at Circle. And just when we grew, it went away.
Edith: Yeah, I mean, so now, I'm sure it's still a checklist somewhere of, "Do they seem legit? Yes or no?"
Paul: It's funny I talked to a company last night that's called TwoDesk.
And what it does is, you put in a URL and you get a downloadable electron app of your SaaS app. And it seems like a hugely valuable thing except that like, what if they went away.
So we were talking about how they could align their business in a way that avoids that problem, avoids that question is like more adoptable.
But the value of it is so clear. And yet at the same time, I would never buy it in its current form because of that fear.
Edith: I think everybody has gotten burned.
Paul: Yeah, yeah. I mean, people will not stop mentioning that Google shuts down services.
Edith: Yeah, I think Google, I don't know if they intentionally shoot themselves in the foot, but they shoot themselves in the foot.
Paul: They really do. They must be aware of the Zeitgeist about a Google thing.
And certainly for Google like Google Cloud is really suffering because of what the mothership has done in the past.
Edith: Well, something leaked at an internal meeting and they said this needs to get to this team or we're going to shut it down.
I think they meant to be motivational internally, like you need to get to here we're going to shut you down but to leak out to the market is this horrible ripple.
Paul: Well, I looked at the wording and I think that that was a little bit of extrapolation.
My recollection of what happened is that they said, you need to get here and with this sort of unstated or else.
Edith: It's kind of like, nice story you got there so even if anything happens to it.
Paul: It is obviously a thing where a company like Google must talk about when are we going to stop investing in this thing?
But at the same time, I can't imagine a world in which Google shuts down Google cloud.
Edith: I think the issue with Google is also a strength is that they throw off such an amount of cash from their advertising business.
Paul: Yeah, Google is weird and fucked as a result of that.
Edith: So for every other business, they're like, oh, you're only making $2 billion well, we make 800 billion on this thing. Everything else was so dwarfed.
Paul: I mean, it's like when Facebook bought Parce, they bought this little strategic thing. And then they're like, "Oh, yeah, we decided to go a different direction. Kill it."
Edith: Yeah. But I do think the developer tools now are in a better place than they ever were before.
Paul: I think that's interesting about it is the people who are developers and the communities of developers, it's like in 2011, when we started Circle, we got customers by going to meetups.
Meetups aren't really a thing in the same way anymore. Or they're much more like spread out across different topics. It is harder to go to a meet up and get customers out of it.
Edith: Well, it's like any marketing channel. It gets saturated, and it loses it's effect. Like, there was a time when you could have a meetup and everybody's like, oh.
Paul: Yeah, people were excited to come to a meet up.
Edith: We tried them too and it was just like, wow, it was a time in San Francisco where every devtool company is like, let's have meetups 'cause this works for Y company. So at any given night, there's probably 20 different meetups going on in the city.
Paul: Yeah, all with the same bad pizza.
Edith: Yeah. And then you also saw the rotation of people who would come just for the pizza.
Paul: Yeah, so I think like that doesn't happen that much anymore as far as I can tell. But then you have new things like twitch streaming of coding. Who knew that people would like watch people code on the internet?
Edith: Well, if you've ever done pair programming you're not new to that.
Paul: Not new, I mean there's quite a difference between like--
I mean, pair programming is like everyone waiting their turn to be the fucking driver again, 'cause this person is like typing wrong.
Edith: How can you type wrong?
Paul: I mean I'll pair program with someone, everyone you pair program is like, no, you know how to keystroke for that one, what are you doing? I have a shortcut setup for that, how could you not? Do you just like type this shit every time like a peasant.
Edith: Like a back seat. I can just picture you next to somebody like.
Paul: So we had a fun experience on Friday night, where just when we were just towards the end of work, it was like kind of pm or something like that.
These two developers who just got access to Dark and this is a Avdi Grimm who's a small talker like, you know, known in that sort of community and JessiTron, who's a well known developer tried at Dark for the first time on Twitch.
And so like we're watching this, and there's other people watching it and we're like, oh, there's 20 or 30 people that are watching this thing and hearing about Dark and then we see there's a bug to everyone like, oh, no, oh, my God, don't type that. No, no, no, no.
It was like really tough watching. And at the end of it, it was all good at the end. There's some bugs, but they got ran the bugs there.
And they spent an hour building something with Dark and then they went to fix some docks and then they spent an hour trying to set up a Docker container. And we felt very good about the relative value of that they got from those two segments of programming.
Edith: You've invented the usability test?
Paul: I mean, usability tests via people Twitch streaming to their fans or whatever.
Edith: That's what a usability test is, you watch somebody do it.
Paul: Yeah, but they're not usually streaming to their fans.
Edith: Yeah, but I mean people have been doing that for the streaming I grant you is--
Paul: It is the streaming that is new, the usability test for sure isn't.
Edith: Like sitting there and the most painful thing is when you have built this thing and you're so proud of it you're like, it's right there, it's right there, just click there. You just rolled over, roll back.
Paul: Yeah (laughs), did you not see, oh my god, did you not see?
Edith: And especially when you encourage, do you encourage them to talk aloud?
Paul: I mean, we don't encourage them, it's their stream. They're not on our channel, this is their channel.
Edith: Yeah, but when you encourage somebody to talk aloud when they're doing something and they're like, I don't know where this is, you know, go up to this menu, you're like, "It is!"
Paul: This especially frustrating one is like hitting a bug that has been fixed but hasn't been merged yet.
Or like, you know, the integration tests are failing for some reason, so we haven't pressed merge and this thing is like, no they could have had this problem go away.
Edith: Yeah, there's nothing more fun, a heat map and a usability tests are worth like 50 sometimes internal discussions.
Paul: How many programming languages do you think have ever been usability tested?
Edith: I know that Logo was meant for small children.
Paul: Yeah, I believe academic languages might have been.
Edith: Logo, Scheme I know were meant to be more teaching languages.
Paul: I'd be shocked the scheme was usability tested, I could be wrong.
This is one of the fun things about Dark that we're showing it to like strangers and then they're trying and then we see all the problems that everyone has and then we're like, all right, let's change this.
Edith: Did you know they were going to livestream it or did it?
Paul: No, no, we saw them tweet about it. We didn't know. We did tell people to like, feel free to do that.
And make videos of it. And we wanted them to send us videos so we could watch the videos of them using it.
Which I guess we could record in the product but like we don't have that setup at the moment. So we're looking for a lighter way of.
Edith: You know it's exciting in the early days of LaunchDarkly when people started doing unboxings.
Paul: Unboxing of LaunchDarkly?
Edith: Yeah, like setting up their first feature flag.
Paul: Interesting. I've never heard of an unboxing of a devtool before.
Edith: You know what an unboxing is, right?
Paul: I have seen the seven year old who makes like 24 million unboxing toys or whatever.
Edith: Yeah, they're like, here I am putting together this feature flag, look at me. And it's really cool.
Paul: On Twitch?
Edith: No, not on Twitch but they'll record a demo video and then tweet it.
Paul: Okay, okay.
Edith: It's funny the early days of LaunchDarkly were incredibly hard. We had a really hard time getting adoption.
Paul: Yeah, I remember.
Edith: And I remember, CircleCI was a super early customer. And one of the designers posted a video for his parents.
Paul: Was it Daniel?
Edith: Yeah, and it was what it felt like to release a feature. And it was like, so picture this, he is the nicest guy on the planet, like a designer.
And he posted this video with him like this is what it feels like when you push something live. And he pushes it live, and then he kind of makes these happy gestures.
And I got so much joy out of seeing that video.
Paul: Yeah, I could totally see that.
Edith: We had so few users at the time so that we had users--
Paul: Ones of users.
Edith: We had hundreds but somebody posting a video of how happy they were was a big deal for us.
Paul: Yeah, we've just gotten to the, we're in the hundreds of users thing at the moment.
Edith: It's a fun stage.
Paul: It's fun stage when you spend a long time getting there and suddenly you're there, then it is a good place to be. But like, you forget the slog that it took to get to the hundreds of users stage.
Edith: What was your slog?
Paul: I mean, Dark is like three years old now. And we're just gotten to hundreds of users.
Edith: Yeah, I'm really repressing a couple jokes.
Paul: What would the jokes be Edith?
Edith: Well, Dark launching.
Paul: Yeah, no, we had a launch party at Strange Loop in September, and I think you guys had a booth at Strange Loop.
And someone from the booth complained that everyone was coming up to the booth to ask about Dark.
Edith: They didn't complain so much it was like, "Hey, we're getting a lot of unexpected traffic."
Paul: Yeah, we want to come to your launch party. No, that's the other people.
Edith: That's the other people. So to go back to the original thing, your original statement was.
Paul: Devtools has changed.
Edith: But for the better, in my opinion.
Paul: I think it has changed for the better. I think one of the main ways that it has changed for the better is the people who are developers, which isn't to say that, like, the people who were developers before were necessarily a bad bunch. But certainly we're a bit more homogeneous than there are now.
Edith: There used to be this perception as we said before, that developers can't or won't pay for tools. That everything has to be some crazy open source.
Paul: So that's not the only thing that I'm talking about. If you think of the stereotype of a developer from eight or 10 years ago, it was you know, sort of a basement dweller to a certain extent.
They're like an angry asshole who preached about open source and the GPL and that sort of thing.
And that certainly wasn't like the whole industry, but like, there were many variations of this stereotype. And then there were also other people. And I think that persona is significantly less of the developers out there.
And there isn't a new dominant persona, they're just like, just many different people. It's just a wider, much, much wider variety of people. And I think it's great.
Edith: Yeah, so, one of the things we do at LaunchDarkly is we sponsor CodeNation.
Paul: What's CodeNation?
Edith: CodeNation is a program for high schoolers, for high schools that can't afford a computer science class. So we have people from the company who go and teach them.
Paul: Oh, that's cool.
Edith: And then we also sponsor a class or office. So every Wednesday, like 20 people come in, and we teach them computer science.
Paul: That's cool.
Edith: And it makes me really happy. I initially felt kind of guilty about it. I'm like, are we just training people to be the next wave?
But they really want to be there. Like there are people from poor communities who do not want to be left out.
Paul: Nice, nice.
Edith: They are like, we are seen this wave and we do not want to get left out. So I was really happy one of the teachers just posted that our current classes 100% attendance.
Paul: Oh, wow.
Edith: Of 15 year olds. Because it's affiliated with the school, but they don't really get a grade.
Paul: One of the things that I feel was like really associated with developers like 10 years ago, was this idea that you had to like code in your spare time and that if you were like in it for the money or for the career or something like that, you were in it for the wrong reasons.
Is this ringing any bells? You're looking perplexed.
Edith: I guess I'm from the wave before that.
Paul: Okay, so what was the wave before that?
Edith: The Dark commerce.
Paul: Right, where everyone was in it for the money. I wonder if that was just a reaction to the .com boom.
Edith: I was in it 'cause I love it software.
Paul: No one is judging you if you don't love software.
Edith: No, I mean, it was funny actually, I really thought, so I got an engineering in econ degree.
And I thought really hard about going into Wall Street because that was the other path I could have gone with. The Wall Street firms recruited out my college.
Paul: You could have gone to Wall Street, you could have been "The Wolf of Wall Street."
Edith: The wolf fit. And I honestly, like, looked at the salary and I was like, well, I can make, and this is my new, this is back then. I could make 700 or 800 K a year on Wall Street.
Paul: That's a lot of money.
Edith: But I want to go build stuff.
Paul: Yeah, I interned at Barclays Capital, which is like in London, but similar sort of thing. And they had a huge like technology section.
And anytime we'd go to talks about technology, like internal talks, the questions would be like, you know, do people from your group get promoted to traders?
And I'm like, what the fuck is this? What are they even talking about? They just gave a lovely talk about database technology, and like you want to become a trader, what are you doing here?
Go be a trader, why did you come in this door?
Edith: Well, because they want to do that. Because the traders is who makes all the money.
Paul: Right, right and were the highest status and all that sort of thing.
Edith: And the developers are the back office.
Paul: I mean, that was certainly the case here. And this was a bank which I think had a higher opinion of developers than most. They were higher status than I think most banks.
Edith: But still there they were considered back office.
Paul: They were definitely lower status than traders.
Edith: Yeah, I do you think a lot of people in software are in it because they like making things.
Paul: I think so. But I think the idea that developers are in it for a nice career and a place where they treat you well, and where as well you also get to make cool stuff.
I think that's like, I don't want to say totally fine because I don't want to say that it even needs me to say that it's totally fine, it's just like, those are reasons to do a thing.
And I think that the idea that the only valid reason to be a developer is the love of code.
And because you love your algorithms and stuff like that, I think that's just kind of crap.
And I used to believe in that, I very much don't anymore.
Edith: I'm really interested because I actually really dislike the word hacker.
Paul: Yes, yes, I have a very similar emotion connected to just a thought.
Edith: Why have you transitioned from like I just want to put heads down and code all the time to?
Paul: Well, I mean, I like to put my head down and code, but I think it is a very privileged position that I'm able to code for the love of it.
And I don't think that the industry should be restricted to people who have that.
Edith: We've talked about this before, I think building software is a lot more than just coding.
Paul: Is that your problem with the word hacker?
Edith: Yeah, it leaves out design, it leaves out quality.
Paul: Developer experience, technical writing.
Edith: And it leaves out even traditionally things that are kind of shunned as like project managers. Like who needs a project manager?
Paul: Yeah, exactly. We sit on our Unix machines and snark at anyone who isn't as enlightened as us.
Edith: Well, the whole Dilbert comics of like.
Paul: Well, I'm thinking of Richard Stallman here and like the sort of maybe a little bit Paul Graham as well.
I talked before about rolling eyes, this is a good example of like, anytime someone uses the word hacker now I'm just like, what?
Edith: I mentor other startups and are like, how do I hack fundraising? Like build a good company with a good product.
Paul: Do you think Facebook killed the word hacker? Or was it the GrowthHackers that killed the word hacker?
Edith: It's funny 'cause I actually really like Sean Ellis. He's a really super guy.
Paul: He's the dude that coined GrowthHacker.
Edith: And he actually, he's told me he's backed away from it.
Paul: That's interesting. Does he still use the word growth or is he just backed away specifically from GrowthHacker?
Edith: I don't recall but he said he was saying it to be provocative.
Paul: That's interesting, it really got taken at face value.
Edith: I think also there was a ton of, it was taken as well marketing is terrible, what we need is growth hacking. And what is this growth hacking? It's like putting out content to that SEOs, that's marketing.
Paul: I think it was fair to consider that there's a different strand of marketing which is growth, which is product facing, like internal to product.
Edith: There's always been product marketing.
Paul: Yeah, I guess there has.
Edith: Like it was just this attack on this.
Paul: On marketing. I didn't realize that that was how it was intended to be.
Edith: I think it was an attack on what was perceived as some of the bloat in marketing. Marketing just buys us pens with colors on them.
Paul: Right or like focus on brand instead of focus on usability.
Edith: But then that also circles back like if all you do is growth hacking, you can get into some very dark patterns that destroy your brand.
Paul: Oh, yeah, that's real. I've definitely used those products.
Edith: Like it doesn't matter if everybody hates us, as long as we have users, it's like no, no, no, no, no, this is my brand and so.
Paul: The Xfinity problem.
Edith: Maybe we should talk another time about how you feel developers have changed.
Paul: I remember in like 2012, seeing a talk by Joel Spolsky, where he said that Stack Overflow had like 50 million users, 50 million monthlies.
And at the time, the world believed that there was like 20 million developers or so but there was much more people than that using Stack Overflow.
So it's like, who are these people, what are they doing? And a certain amount of them are business analysts or people who were like SQL for a living.
But you know, a large number of them were hobbyist developers or developers that weren't accounted for because they didn't go to college, or they're in countries that wasn't taken into account in their statistics that people were coding. And there's question today like how many developers are out there?
And like with the proliferation of boot camps, the frequency with which developers don't go to college and are like self taught.
obviously massive rise both computer science degrees, but also like in categories of things that are sort of developerly like data science and that kind of thing.
It's like, how many people are there who code for a living? How many developers are there? Do we know, do you have a number?
Edith: I looked at these numbers for own market studies and it comes up back to what you said, there's no good hard number.
Paul: Right. Yeah, so I mean, numbers are doubling every five years is a thing that is known.
Although I've heard 10 years so I actually don't know what's known there.
Edith: And there's also like, the question, you touched on what is considered code? Like you glossed over a sequel and it was like, well, I might consider that code.
Paul: I consider that code. Yeah, no, I'm completely down with that. The reason I was sort of like separating that out a little bit is that there's a lot of people who use SQL but don't write code for a living. Or no, that's not the right thing.
Edith: Oh, we just came up with a divide sequel, SQL.
Paul: Which one do say?
Paul: You say sequel, yeah, I've always said SQL.
Edith: We can still be friends.
Paul: People who are studying data versus people building products, I guess, would be the split there. So like SQL is often used for studying data.
It also has a use for creating data. That's probably like one of the first splits before there was like front end and back end. And now there's even more of that.
Edith: There's also and we feel there's a huge amount of coding that's used in science. Like in MATLAB.
Paul: Yeah, Matlab is coding for sure. There's a lot of people who like claim certain things aren't coding.
Edith: Which I don't understand, it seems like a weird hill to--
Paul: Yeah, well, I mean, people always want to protect their status as like they are the one true whatever it is and so therefore, it's important to say that someone who's writing a Monte Carlo method in Excel is not coding even though they clearly are.
Edith: They clearly are. And then, you know, I think people 30 years ago would be like, these children of the sun, they don't know garbage.
Paul: They only use garbage collected language, this isn't real. What is this object thing, I lay at memory?
Edith: Yeah, they don't know how to deal with memory efficiently. They just assume somebody else takes care of that. So it's like, well, yeah.
Paul: Yeah, they do, in fact, 'cause that's the better way to do it.
Edith: Yeah. And they're like you're not programming a chip so you're not actually programming.
I guess my final statement on this is that if you're like, when we're just talking through it, developer tools are just moved and I said this before, further and further and further up the stack.
The stuff you used to have to care about, you don't have to.
Paul: That has all changed.
Edith: Yeah. Like now if you're like, well, I'm really good at managing. Like, I remember learning an assembly language. I'm really good at assembly language and how people were like--
Paul: But web assembly is back in vogue now. So maybe your assembly language skills are like--
Edith: Are you saying I'm old enough to be relevant again?
Paul: Yeah, it's all cycled back.