October 6, 2015
Ep. #4, Scaling Your Engineering Team
In this episode, Edith and Paul talk about who owns code, how to scale your engineering team, and cowboy coding.
In episode 55 of To Be Continuous, Paul and Edith discuss their conference experiences, as both attendees and hosts, as well their firsthand knowledge of freemium and low-tier pricing models.
Paul Biggar: As I was saying, a good lead is when you're starting in the story already, you're not just starting with a question. Like, "What do you think of conferences?"
Edith Harbaugh: Or, "Who am I? Why am I here?".
Paul: Right, exactly. You don't want to "It was a Dark and stormy night," it was like, "Yes. Someone walks in and, I don't know, does something. Someone gets murdered."
Edith: Yeah. So what do you think about conferences?
Paul: Depends on whether I'm going for work or whether I'm going for pleasure. It dramatically affects--
Edith: You actually go to conferences for pleasure?
Paul: I have done, I would say not recently, but I went to Strange Loop last year.
Strange Loop is amazing, and I would say that's definitely the best conference I've ever been to.
Edith: What did you like about it?
Paul: Largely the audience. It's an un-commercial conference.
There's sponsorships and that thing, but it's just a lot of people talking largely about programming languages, distributed systems, that kind of thing.
Edith: Stuff you can't stand talking about.
Paul: Stuff I can't stand talking about.
Edith: You got to sit down, you like it that much.
Paul: Right, exactly. I don't know if it was quite a goal, but our intent was to talk to a lot of people about Dark, give them a lot of demos.
So there was definitely a work component to going, but between the parties and they have this amazing party in the city museum which has these five-story slides in them, or 10-story slides or something.
It ended up being just a lot of fun as well. We're actually going to time the launch party for Dark with Strange Loop, we're going to do it in St. Louis this year and invite all the people who've seen Dark or who just want to see Dark to the launch.
Edith: That's funny. What do you see as the purpose of a launch party? Because LaunchDarkly never did one.
Paul: Did you do a dark launch?
Edith: I think the closest we got to a launch party was after we closed our seed round we had a meet up here at Heavybit, and we bought T-shirts.
Edith: Up to that point I had considered T-shirts too extravagant.
Paul: I think it's one of those things, and this is related to what you're asking about conferences.
There's a lot of marketing strategies that done well can work, and done poorly can fail dramatically. I think having an official launch is one of those things that it can just be nothing, or you can turn it into something good.
Edith: Well, it was the whole Steve Blank thing of "Don't launch."
Paul: I would say ours takes into the Steve Blank thing. It's like, Airbnb launched three times.
The first two, no one was listening so who cares? Keep launching until it takes off, and their third one was their taking off.
So we've done things that are vaguely launch-like. We had a recruiting poster where we we're looking for an EIR to build stuff in Dark.
That was nearly a year ago, and so that was an opportunity for people to hear about Dark, almost a content marketing opportunity. This is the same.
There's an audience of people who seem to be extremely interested in either Dark specifically, or the same sort of thing.
People are looking for stuff to do, so we're going to have an event and we're going to tie that to--
We're going to release a video of Dark to the world at large, but Dark will not actually be open at that point.
So it'll still be like a pre-release thing. We'll be switching from a private-alpha to private-beta.
Edith: This is a granularity that I cringe at.
Paul: Yeah, it's really not important. It's like you released version 7 of your software.
It's like, "Why did we do that? That gives us an opportunity to have an announcement."
Edith: Yeah. My theory is that the last really successful launch parties went out after Windows 95.
Paul: I remember Windows 95 coming in, I spent so much time reading about Windows 95 when it was being released.
They put so much money into that, the whole fucking world was talking about Windows 95.
Edith: Which now seems so dry.
Paul: Yes, but also it was such a huge deal for the world at large.
Edith: Yeah, why?
Paul: Everyone in the world used this software.
Edith: Were we just so much younger back then that an OS was a big deal?
Paul: Right. Things didn't turn into this continuous delivery release. It's like, you have a new version of iOS. People care, but not as much.
Edith: Actually, I think people care in a negative way.
Paul: It's just like, "Here's another annoyed piece of--" Yeah, that's true.
Edith: Yeah. It's like when you see the pop up, "Do you want to update your computer?" Like, "No. I'm fine. Leave me alone."
Paul: "I would absolutely like you to not shut down all my windows."
If they were really good at keeping everything working the same, like freeze a snapshot of something and update the kernel, load the snapshot, something like that.
When my phone updates, it's no big deal. Everything stays the same.
But when my computer updates it fucks everything up, and then I have to go reset things.
Edith: One of my worst conference talks ever was I was sitting at my laptop and everything worked just fine, like the projection on the screen was ready to go.
Then one of the AV techs came in and said, "Do you want a clicker?" I'm like, "No. I'm fine, I don't need a clicker." He's like, "Let me just put it in."
I'm like, "Do not touch my computer." He jammed it in and didn't work, and I'm like "Take it out."
He's like, "No, no. Let me just click this button." Then he rebooted the entire machine, and it took 25 minutes to reboot.
Paul: Why did it take 25 minutes to reboot?
Edith: Because he had clicked enough to like -- He was like, "I'll just install this driver."
Edith: Because he'd stuck in one of those--
Edith: It's like, "I need this driver and this needs to reboot to install the driver."
Paul: Jesus fucking Christ.
Edith: It was ten minutes into my talk at this point.
Paul: This was like a minute before your talk, or something?
Edith: Yeah. So he rebooted my entire computer while he's installing all this stuff, 15 minutes into my talk I said "I'm just going to start without the slides."
I was livid. I was like, "Do not touch my computer."
Paul: Yeah. So you're saying you don't like conferences, then?
Edith: I love conferences. I think they're a very effective way to have high bandwidth communication.
Paul: Do you mean from a people hearing about LaunchDarkly sense?
Edith: Or even, here's an example. Right before this we had a lunch for some of our customers and some of our prospects.
Paul: Prospective customers, or prospective buyers?
Edith: Prospective customers. So we have a panel, and then a lunch, and afterwards somebody came up and in five minutes he told me more about what was happening at his team than we had learned in two years.
He's like, "I loved your panel. Let me tell you about this, what's actually happening."
Paul: Conferences are the same sort of thing?
Edith: Yeah. People have an opportunity to think about something, and then talk face to face it.
You just suddenly feel comfortable in a way that you don't feel comfortable in email.
Paul: So, it's about giving your customer base or prospective customer base access to you?
Edith: At least for us. In the early days of LaunchDarkly I got invited to speak at a conference in Sydney, and they paid for my travel so I went.
Edith: Yeah. I gave a really neutral talk on just why feature flagging is good, and after it two people came up and said, "We actually want a demo."
They both became customers, and really valuable customers.
Edith: All because I was there.
Paul: Right. With Circle, we almost never did conferences. I think one of the reasons for that was that one of our major competitors, Travis CI, they were amazing at conferences.
They were just a lot of open source people, they went to all the conferences, and then when they went to the conferences they did workshops afterwards.
Like, "Setup your open source with Travis." We didn't work with open source at the time, but we do now.
So it's like, "Do we want to compete at this where they're already--" We would be going to the same conferences. Or, "Do we want to focus on our own thing?"
And in the end we spent that time focusing on product, which ended up being a differentiator, and more on content in terms of marketing.
Definitely one of them wasn't going to be a successful strategy, but I think with Dark that will be different.
With Dark, a lot of people are going to want to see a demo. "I want to hear about Dark."
Like, I imagine conferences are going to be a much bigger thing for Dark.
Edith: Yeah, I think I want to be clear that I don't think conferences are a cure all for all products.
I think they work well when you're evangelizing or doing something new that people want to feel a connection with.
Paul: There's also the company that you are, so you're a top-down sell, is that right?
Edith: I'd say we're middling around.
Paul: What does that mean?
Edith: Like, we really need to convince a team leader or director to use us.
Paul: OK. So at a conference, who is it that hears you and then-- Like, who becomes a lead for you?
Edith: Usually it's a mix. They've seen our website and I find this out now, they've seen a lot of our content online and they're coming to talk to learn more. Or, they've seen some stuff--
Paul: And they are an engineer, a team lead?
Edith: Engineer, team lead. They're like, "My manager sent me to your booth to check this out."
Paul: Gotcha. Yeah, Circle was a very bottom-up. It was an engineer set up for their team. We're doing a lot more top-down these days.
People can actually reach out and talk to a salesperson, and have someone help them purchase if they're a large company, that kind of thing.
But at the time it was like everyone got started on the $50 plan, or even the $20 plan. Then they grew from there.
Edith: Yeah, that was our original thesis. We had a $49 dollar plan, but nobody signed up for it.
Everybody needed to get by on at least the team lead level, so they just auto-selected to higher levels.
Paul: Do you have freemium?
Edith: We never had freemium deliberately.
Edith: I had been at TripIt, which was a freemium product. It's a marketing channel.
Paul: Of course it is, yeah.
Edith: I think some people don't really understand that.
Paul: If it doesn't bring you in money, then it's got to do something.
Edith: I got into this heated discussion with another CEO who was like, "It just gets you a lot of eyeballs and the VCs give you more money." And I'm like, "Until they don't.".
Paul: That's not how I view the world.
Edith: Yeah, at TripIt we converted good in our class, so around 2-3% users converted to paid.
But if you do the back of envelope math of how many people you have to get on your free tier, and what you have to charge.
Paul: It's interesting how different it is for different industries.
Like SaaS, 9% is unheard of and 3% is pretty good. But then you get things like GitHub for example, every software engineer has a GitHub account, and who knows what percentage that actually turns to.
Maybe it's only like 3%, they've just seen so much GitHub that at this time 3 % is a lot.
But there are companies where everyone in an industry uses them.
Edith: Yeah. So, for conferences I think it's like freemium. I think it is a strategy, but I don't think it works for everybody.
I've seen people-- It's so easy to spend millions of dollars on conferences.
If you have a misaligned sales and marketing team who is aligned, just like "We're incented by how many Beanie Babies we give away ."
Paul: You mean that you need to incentivize people by sales-qualified leads?
Edith: Yeah. For every conference we go to, we look at the complete cost of the conference, including travel.
We cost-in when we're thinking about going to a conference.
Paul: So how much does a normal conference cost you?
Edith: It depends. Like, let's say a regional conference in Seattle. Let's say a sponsorship is $15-20K, and then we also cost-in travel.
Paul: Of course, yeah. From here it's $500 a person. How many people go?
Edith: Let's say four people go, round trip flight is $400 dollars.
Edith: For hotels, and Seattle's expensive in the summer. So, $300 for two nights. It adds up quickly.
Paul: What about things like prep time, do you manage to factor all that into the total cost of ownership?
Edith: When we look at a conference, what we look at is we just look at "Do we man it with the pipeline out of it?" And then closed.
Paul: OK, so it sounds like you're talking $30k for a conference?
Edith: For a smallish one. Like AWS, much more.
Paul: Right. Spend $100K for AWS?
Edith: Depends how big a booth we get.
Paul: Right. How many people would you send?
Edith: I think this year we sent six. So in the early days I was the only booth person.
Paul: Yeah. You had a little tiny booth in the back?
Edith: Yeah, no. Literally the first time I went to Web Summit. It was just me, and I just stood there.
Our first conference as a company was billed-- We were 8 people at the time, so we rotated everybody out.
Including engineers, because it was a three day conference. You can't have one person stand there for eight hours a day, but it is exhausting.
Paul: Yeah, I can imagine.
Edith: So you want to have some sort of rotation where somebody can take a break.
Paul: I don't think I've ever had a booth.
Edith: You never had to do a booth?
Paul: Not that I can think of, no.
Edith: I did it all in the early days, and I was so happy when I no longer had to go in regular rotation.
Paul: If I'm thinking about a booth now, that probably means we'd have to put together marketing materials, and we never did any of that either.
Edith: You can be pretty basic. The first time we did Microsoft build, we just made a sign and then we had a little Levi's.
Edith: I don't think we had stickers back then. It was not super elaborate.
It was funny, whenever our engineers wanted to experiment with bookmarks we made bookmarks and nobody took them, but it was a pretty cheap experiment.
Paul: Do you feature flag your experiments in real life?
Edith: Up until the B round, we're pretty ruthless about we would only go to conferences that we thought we would definitely get leads from.
Like, OK we're not going to go to this one, because it's funny.
We didn't want to go to Java one because we'd been there as engineers and we didn't think that was our market.
Paul: Right, OK. You think Java engineers don't use feature flags that much?
Edith: We thought that they were on a different release cadence.
Paul: That was extremely delicately put.
Edith: When we got more money, we had more money and we could experiment more.
So when we raised our C round, which was $44 million dollars, we'd be like "OK. Let's try to go to more conferences. Let's be vigilant about tracking costs."
Paul: Do you find that there is a significant difference in the number leads between whether you're speaking or whether you're just attending?
The best thing is to have somebody speaking, and then the booth can be a driver.
What I do is I give very vendor-neutral talks, but I could say "If you're interested in learning about the company, we have a booth.
Paul: So that sends people down.
Edith: Yeah. If you just have a booth and nobody's speaking, it is harder, because you're hoping that somebody is walking around looking for stuff to buy.
Edith: Then you also have to be careful about T-shirts.
Paul: What specifically do you have to be careful about?
Edith: By the way, thank you for wearing your LaunchDarkly T-shirt.
Paul: I got it at a LaunchDarkly conference.
Edith: A lot of people will go and grab T-shirts to take home to their kids, and they don't have any interest in the company at all.
Paul: Right. So that's going to conferences, but you hosted your first conference recently. Trajectory?
Edith: Trajectory. Has Circle done a conference yet?
Paul: I don't think so, no.
Edith: That surprises me. You have so many customers.
Paul: I don't know. And people, we also have a lot of people.
Edith: Do you publicly say how many customers you have?
Paul: Do you mean only paid customers? Because we have a lot of people who are relying on Circle.
Edith: Do you release how many paying customers you have?
Paul: We do not.
Paul: My last recollection was for our series A, we had 1,000 customers.
Edith: That's awesome.
Paul: We didn't have a freemium tier at the moment, so a lot of them became freemium when we launched freemium later.
Edith: That's terrible.
Paul: No, it's not that much money. Lots of them were on the $20 plan.
Edith: So, they just went to freemium?
Paul: The switch to freemium ate growth for like, two months. It was every dollar that we made was worked against someone not upgrading or downgrading, or something like that.
Edith: That must have been a brutal month for your metrics.
Paul: Two months. Yeah, it was pretty bad. It was like, "Why aren't we growing?" We did not put together that it was the freemium launch the previous month.
Edith: So did you auto move people down, or did they self-select?
Paul: It was all self-selected.
We didn't auto-move anyone down, what we did though was we grandfathered people in, so there was a bunch of people who were getting the $100 dollar value for $20 because they were two grandfathers in, and then they refused to move because the jump was too big.
It wasn't perfectly managed, but the number of customers that were affected by this is like 10 or 20.
Paul: Then the number of people who downgrade, there's only a thousand who could potentially downgrade. Most of them were actually-- so it's probably only half that.
It wasn't necessarily that big a deal, but it's just like--.
Edith: I'm doing some quick math. So, you're charging $20 bucks a month.
Paul: Some people are under $20, some people are under $50.
Edith: That could be a hint.
Paul: Yeah. It was at the time where we're doing a million year, probably around then.
Edith: What made you decide to do the freemium at the time?
Paul: I don't really know.
It's just one of those things I hold to be true, that if people can get in easier then people can like the product more and give us more money.
I guess partially with CI there's a lot of expansion revenue, and expansion revenue will often be larger than acquisition revenue on a particular month.
So, the easier it is for them to get in and we had our first plan as $50, then we dropped the plan to $20 dollars, and now it's free. We're not losing that much revenue.
Edith: What we discovered at TripIt was the amount of friction between $29 and $49 was the same.
Edith: People were just as likely to pay us $49 or $29.
Paul: Yeah. TripIt didn't have significant competitor, am I right?
Edith: It had many competitors.
Paul: Really? OK.
Edith: Yeah. We had a ton of competitors at the time.
Paul: Gotcha. It feels like there was a whole movement around CI at the same time, so I saw a new competitor every month.
There was a bunch of established competitors-- We're the biggest now, I'm not sure that we were the biggest for the life of the company, certainly in the early years.
Probably Travis CI were roughly the same size. There was a bunch of competitors that have since died off or got acquired, or--
Edith: I mean, Travis is kind of no more, really.
Paul: It makes me very sad. I don't know anymore than people know publicly, but it didn't end in the way one would hope.
Edith: What made you sad about it?
Paul: Travis sold to Idera, I think is the name of it.
Edith: Some private equity.
Paul: A relatively small private equity thing, so the sale price was like $10-20ish million, probably.
It feels like it was a better company than that, it feels like it was worth more and feels like it had a better future.
I just personally have no idea what went wrong, and then they came in with the laying people off.
It felt like-- I'm quoting people on Hacker News here rather than giving my own opinion, but it felt like the story became extracting at the revenue streams, and lowering costs rather than actually continuing to reinvest in the space.
Edith: That's the story of private equity.
Paul: Right. It feels like it is not the time for that to happen to CI, there is so much more to do.
Edith: The other reason we never did freemium at LaunchDarkly is there's a ton of perceived value about having to pay for something.
At the beginning days, Circle was a really early customer of us. It wasn't about the money, it was whether you trusted us.
Paul: Over time you build up that trust naturally by just having a brand, so people will try you out and people will connect their whatever to you without having a conversation, in a way that they would never do when it was just the Edith and John show.
Edith: Yeah, it is so much easier these days. In the early days, there are two huge hurdles. I'm going to tie this back to conferences.
The two huge hurdles were, "Do you believe in feature flagging at all? Do you believe that this is a valid way to build software?" And there is a ton of that.
Paul: Right. What year were we?
Edith: This is in 2014.
Paul: It feels like this podcast didn't exist, there wasn't a lot of stuff out there about continuous delivery.
Paul: That was a joke. There certainly wasn't as much by like an order of magnitude.
Edith: Feature flagging, I would still get invited back then to sites and they would say "Convince us about why we should be feature flagging."
Edith: Like, convince us that this is--
Paul: Does that happen anymore? Or are they significantly larger customers for whom that happens, or more conservative customers?
Edith: There was two hurdles in the early days, "Do you believe in feature flagging? And do you believe that our company is a reliable vendor?"
And those are both two huge hurdles.
Like, I would go and I remember pitching these big e-commerce sites about why they should feature flag, and then the second step was "Why us? Do you trust us?"
Paul: We all have the second one. So for Circle, there was this movement happening around CI and already a bunch of parts of it had happened and had moved to the cloud.
Heroku, the hosting had become cloudy and the source control had moved to GitHub so that had become cloudy.
There was this step in between that was not yet in the cloud.
It feels like around 2012 they just started to be this movement where everyone's like, "Yeah, we need CI and we need it to be in the cloud."
We didn't have to do anything for it, so the only thing that we had to trust people is like, "Why would we give you our source code?"
It was at a time where giving people your source code seemed crazy.
Edith: Yeah, crazy talk.
Paul: You could not do it. Then as the company got bigger, we had to convince people about the start, but a year or two in questions stopped coming up.
People had heard of us, and that was enough.
Edith: Yeah, the shift happened over last two years, where now it's our DevEvangelists are much more on the road than me.
They'll ask questions like, "Are you currently feature flagging?"
Edith: It used to be that two or three hands would go up, and the rest of the people had just come to the talk to learn about this.
Now, I gave a talk in Chicago at the CTO Summit and I said, "Who here is currently feature flagging?" Every hand went up.
Edith: So it had switched from "Should I do this," to "How should I do this the right way?"
Paul: Right. Are you competing against people rolling their own feature flags?
Edith: I don't think of that as competition at all, I think of that as an on-ramp.
Paul: Nice. Have you practiced that well? Is there are a lot of VC pitches where you've got asked that question?
Edith: No, I honestly believe it. We put out a ton of content about how to build your own system, because if you built your own system you don't want to do it again.
I was just at a lunch and learn with the VP of engineering from Roll Bar, and he talked about how he'd had to build his own system at Cisco.
He's like, "Never again." Then the VP of engineering at Jern Deploy, he had to build his own system at ExactTarget.
He's like, "Never again." If you've built your own you don't want to build it again.
Paul: Right. I imagine that there's a lot of orchestration things that they're all switching to Kubernetes, obviously.
But there was, for a time, there was a bunch of things that are like Kubernetes, and everyone is like "I built one of these. It was fucking horrendous."
Because everyone built one before they existed. Even if you didn't call it an "Orchestration system," it was like you wrote a bunch of software that orchestrated a lot of Cloud or Containers or VMs, or something like that.
Edith: So a lot of our conference strategy is just telling people there's a better way, and if you want to build it yourself, here's how to do it and have fun.
Paul: At the start of Dark, we were keeping quite quiet about what we were doing.
Edith: You were--
Paul: Keeping it Dark. Thank you, I got there first. But now we're just writing it up, so I've a bunch of blog posts.
We wrote one called "What is Dark?" It's on our blog, and it's just lays out the way that we think about the world.
I had previously been very nervous about other people getting this idea and running with it, and now it's just a challenge to even convince people that this is a good idea or that it's even possible. So, just tell everyone.
Edith: Yeah, that's always been my theory. We shout and hope somebody listens.
Paul: Right. Does that lead to competitors at all?
Edith: Of course.
But competition is good, it makes you focus on what you think you're good at.
We have such massive scale now, we helped Intuit at tax day. I am very proud of our customers and what they're accomplishing.
I think we went heavy on conferences, but we were always very careful about not just throwing money in the wind, trying to account for where it was generating results.
And then we did our first conference, and that was so much fun.
Paul: How was Trajectory?
Edith: It was a blast.
Edith: What I hadn't realized is that customers came, people who were interested in us came, and I hadn't realized that people who were thinking about joining us came.
Paul: That's obvious now that you say it out loud.
Edith: Yeah. We had an offer out to a salesperson, and he came to that conference because he wanted to get a feel for how excited the customers were.
He went home that night and he signed the contract.
Edith: But we had people who were thinking about joining us who wanted to get a feel for--
The exact quote from the sales person was, "Do people actually care about this? Or are they just here because their boss sent them?"
Paul: Right . I went to the Honeycomb user conference.
Edith: I was there, too.
Paul: Right. I'd never seen this before, but they barely had a schedule at all.
Edith: It was an un-conference.
Paul: Right, and that was-- They had two talks or something, and then everything else was these "Write your name on the thing. There's a facilitator, there's 12 people sitting around talking about a thing." It was really good.
Edith: It was fun. The interesting feedback we got on our conference was we were trying to make it again, very vendor neutral.
So it was all about talks from our customers, and even talks from non-customers. One of the most popular talks was Josh Wils, who's at Slack.
He talked about the feature flagging system he'd built at Google and the challenges there. So, not a customer at all, just somebody who is in the field.
The feedback we got was that people wanted a little bit more from LaunchDarkly.
Edith: Because like we'd held--
Paul: Isn't that the feedback that you got before when people wanted demos and you had you had kept it vendor neutral?
Edith: I think it's always better to leave people--
Paul: Asking for more?
Edith: Yeah, I think we went a little too slim. Because I gave a five minute opener and John Kodumal, my co-founder, gave a five minute closer.
People were like, "We actually wanted a roadmap."
Paul: Right. Was it multitrack?
Edith: We did two tracks.
Paul: Right. So you could've had a LaunchDarkly-specific feature set, roadmap, fireside chat with the VP product.
Edith: It's funny, because I think it was a reaction to people who'd been at companies where their conferences were very vendor-specific.
Paul: For sure.
Edith: So we'd gone way over to the other side, where people were asking for more.
Paul: I've definitely been at conferences where they sell a vendor slot and then you have the worst talk ever, because it sold they did not have to qualify.
Edith: We usually don't buy slots, but we do buy slots at QCon.
Paul: I didn't know QCon sold slots.
Edith: They do. But you are on a specific vendor track, and they label it as vendor track.
It's very clear this is the vendor track, but I was very happy because this year they keep track of how many people are--
Paul: Attending each one?
Edith: Pre-registering. So they moved me to the second biggest room.
Edith: Because even though it was a vendor talk, people wanted to see LaunchDarkly and what we were up to.
Paul: That's cool.
Edith: A final thing we should talk about is, what conferences do you like to attend besides the one in St. Louis?
Paul: I like language-specific ones, or stack-specific. If there's something where there's, let's say closure, Closure Conference was one I went to before.
Not only do you get to talk to people who have a broad set of the same problems as you, you get to see a lot of what's happening in the space and what's the advanced front.
Apart from that, I don't really go to much else.
I'm not a significant conference-goer, but the best conference I ever went to was the history of programming languages conference.
Edith: I just like to save the best for you.
Paul: The best for me, it was wonderful, but it was in 2007.
They hold it every 15 years because there's just not that much history, but it happened while I was doing my PhD.
I convinced my advisor to let me go, and it was next to all the other program language conferences like PLDI.
It was really great, and all the people who wrote programming languages were there.
Edith: Are you going to go in 2022?
Paul: I hope to go in 2022.
Edith: 15 years later.
Paul: Right. Can speak in 2037, hopefully. Fingers crossed.
Edith: Do you go to CTO conferences?
Paul: I did go to CTO Summit, that was really good.
Edith: Yeah. Somehow I get sucked into CTO summit.
Paul: A lot of the comments are just like, "You go to hang out with your friends."
Edith: Yeah. That's the way you think of Heavybit events, and I really enjoy them because I get to see people who also care a lot about developer tools.
Paul: Right, but at CTO Summit I know 15-20 people who go to it.
Edith: Something I try to be very careful of, and I love CTO Summit because I think it's the perfect size of around 150.
So you still feel comfortable walking up to somebody and saying, "What do you think of that talk?" At AWS I would not just walk up to some stranger and say--
Paul: Yeah. For Dark, the launch party, we had a planning session for today.
We talked about icebreakers and that sort of thing, make it easier for all the introverts to actually enjoy themselves.
Edith: Yeah. I think up to about 100 and you could still feel comfortable, but beyond that it starts to break into clicks.
Paul: Right. It's so big that you can just pick a random stranger to talk to.
Edith: Yeah, or just feel comfortable that they're not going to be like "This is weird."
Paul: What is your favorite conference to go to?
Edith: I love the Heavybit events.
Paul: They're wonderful.
Edith: I like the DevGuild stuff, because people come up from LA or in from out of town. I love the CTO Summit events, and I'm very fond of GlueCon and I couldn't go this year.
Paul: Yeah. GlueCon is cool. You should go to Strange Loop, come to our party.
Edith: Do get a T-shirt?