Composability with Ishan Anand and Scott Steinlage
A live Remix Conf discussion covering remote work challenges, America-centric web development, Layer0's acquisition and edge computing, and composable architecture.
Episode Description
A live Remix Conf discussion covering remote work challenges, the America-centric nature of web development, Layer0's acquisition and edge computing, and composable architecture.
Episode Summary
Recorded live at Remix Conf, this episode brings together Anthony Campolo, Christopher Burns, Scott Steinlage, and Ishan Anand for a wide-ranging conversation that begins with the excitement of returning to in-person conferences after two years of remote-only interaction. The discussion quickly turns to the complexities of distributed work, with the group exploring time zone friction, the challenge of synchronous collaboration across continents, and Conway's Law as it applies to team organization. Christopher Burns offers a unique perspective as a UK-based developer, highlighting how web development is deeply America-centric — from conference locations and community activity hours to the English-language bias baked into code itself, including spelling differences that can cause real bugs. The conversation shifts to Layer0's acquisition by Limelight Networks and the planned rebrand to Edgio, with Ishan explaining how the merger positions the company to become a major edge platform by combining Layer0's Jamstack expertise with Limelight's massive CDN infrastructure. A nuanced debate about edge computing follows, weighing when computation at the edge makes sense versus centralized rendering distributed via cache. The group also tackles GDPR, data sovereignty, and compliance as underappreciated concerns in distributed architectures before closing with a discussion of composable architecture and the upcoming Composability Summit, exploring when monoliths versus microservices make sense depending on team size and business context.
Chapters
00:00:00 - Live at Remix Conf and the Joy of In-Person Events
The episode opens with playful banter as the hosts and guests settle into their first in-person recording together at Remix Conf. Ishan and Scott are welcomed back for their second appearances on the show, and the group shares their excitement about attending one of the first post-pandemic developer conferences after two years of purely remote interaction.
The conversation takes a lighthearted turn as everyone marvels at finally seeing each other in full — legs and all — after only knowing one another from the torso up on video calls. Anthony discovers he and a former coworker are the same height, something impossible to know from remote work alone. The moment captures the broader sentiment echoed by conference attendees: that in-person connection is irreplaceable and deeply energizing for the developer community.
00:02:48 - The Realities of Remote and Distributed Work
Ishan explains that Layer0 was remote-first before the pandemic, giving them a head start, while acknowledging that COVID forced the rest of the industry to catch up. The group discusses the importance of periodic in-person gatherings to build social capital, with Ishan describing how Layer0's quarterly on-sites were essential for maintaining team cohesion and shipping effectively throughout the year.
Anthony reflects on the double-edged sword of remote work — he thrives working from home but deeply missed the social contact that conferences provide. The conversation turns to the real pain points of distributed teams, especially time zones. Ishan breaks down how Layer0's team spans from the US West Coast to Eastern Europe, and how organizing engineering teams geographically — much like decomposing a monolith into microservices — helps manage the complexity. The group connects this to Conway's Law and the two-pizza team model.
00:09:29 - America-Centric Development and Global Inclusivity
Christopher Burns shares his perspective as a UK-based developer, arguing that programming culture is overwhelmingly America-centric in ways Americans often don't notice. From conference locations to Discord activity hours to community event scheduling, developers outside the US face significant friction in participating fully, with entire time zones experiencing dead periods of community engagement.
Scott pushes back gently, noting that India likely has a massive developer population that's underrepresented simply due to accessibility barriers, drawing a parallel to how Google expanded YouTube's reach by optimizing for slower networks. The group discusses how social media platforms are predominantly American, and Christopher reflects on how being at the conference in person has shifted his understanding of American culture and development norms. The conversation highlights the tension between wanting global inclusivity and the practical difficulty of accommodating all time zones.
00:18:07 - When Code Speaks American English
Christopher offers a concrete illustration of how American English is embedded in code itself, from CSS properties like "color" lacking support for the British spelling "colour" to the dollar sign being universally associated with jQuery rather than currency. He shares a real bug he encountered where he unknowingly created two versions of the same function — one with British spelling, one with American — in different files, caught only by TypeScript.
The discussion broadens to how address formats, date formats, and terminology like "zip code" versus "postal code" create friction for international developers working with American-built platforms like Stripe. Christopher points out a positive example from the Remix workshop where the native browser date picker automatically localized the format. The group considers whether the web should be smarter about handling these variations, and whether non-native English speakers face even greater challenges, before joking that all programs should be written in Esperanto.
00:23:46 - Layer0's Acquisition and the Birth of Edgio
Ishan sets the stage by explaining Layer0's core offering: a Jamstack platform designed for high-stakes, large-scale websites where traditional static approaches don't work. He describes how the company was doing serverless Jamstack before the term existed, even building their own version of AWS Lambda. The conversation then turns to the acquisition by Limelight Networks, one of the largest CDN providers, known primarily for powering major streaming services.
Ishan uses an airline analogy to explain the strategic logic: Limelight already had the infrastructure (the planes and pilots) and needed the software and expertise to enter the web CDN and edge platform space. He announces the additional acquisition of Edgecast and the upcoming rebrand to Edgio, describing the experience of being acquired as going from high school to college — an achievement that simultaneously signs you up for operating at a much larger scale with new cultural and organizational dynamics to navigate.
00:29:08 - Edge Computing: Promise, Nuance, and Data Sovereignty
Anthony expresses enthusiasm for edge-native architectures, arguing they simplify the developer mental model by making deployment location invisible. Ishan offers a contrarian perspective, noting that running server-side rendering at every edge node isn't always efficient — rendering once and distributing via cache to 100 POPs can be more cost-effective for non-personalized content. The group agrees that the right conventions should abstract these decisions away from developers.
Christopher steers the conversation toward data sovereignty and GDPR, arguing that edge computing's biggest impact may be in compliance rather than performance. The group discusses how distributed databases raise thorny questions about where customer data physically resides, with Christopher sharing real-world examples of UK nonprofits requiring data to stay in their country. Ishan notes that compliance is the foundational layer of a developer's hierarchy of needs, and the conversation touches on the legal gray areas of the US Data Shield, cookie consent, and how regulations like GDPR, while designed to protect citizens, can disproportionately burden startups compared to large enterprises.
00:37:30 - Composable Architecture and the Composability Summit
Ishan introduces the concept of composability as an umbrella term for trends including Jamstack, API-first, MACH architecture, and headless approaches — all centered on the idea of building and rebuilding technology stacks like Lego bricks. He cites a McKinsey survey showing nine out of ten executives believed their business model needed to change, underscoring why developer velocity and adaptable architectures matter. The team announces the Composability Summit at composability.dev, a virtual event planned for late June.
Anthony challenges the composability narrative, noting that microservices and micro-frontends can actually slow developer velocity compared to full-stack monoliths. Ishan clarifies with enterprise examples — like Salesforce needing consistent branding across acquired properties built in different frameworks — showing where composable design systems and framework-agnostic approaches become essential. Christopher offers the hot take that architecture choice ultimately doesn't matter to customers, and the group converges on the wisdom that premature optimization is the root of all evil, with the right architecture depending entirely on team size, business context, and current needs.
00:47:48 - Shoutouts, Closing Thoughts, and JavaScript Jam
The episode wraps with the guests sharing their social handles and promoting upcoming projects. Both Ishan and Scott highlight composability.dev and encourage listeners interested in the topic to reach out about speaking at the summit. Ishan also promotes the rebrand to Edgio, asking listeners to watch for the new name by the time the episode airs.
Scott gives a shoutout to JavaScript Jam, the weekly Twitter Spaces show he co-hosts with Ishan and Mark Brocato, describing it as an audience-driven open mic for web development discussion held on Wednesdays at Pacific lunchtime. The group shares an anecdote about Dan Abramov spontaneously joining that day's session from his car, and Christopher humorously notes that he spoke after the creator of React. The episode closes on a warm, personal note as Christopher expresses gratitude for the in-person experience.
Transcript
00:00:00 - Anthony Campolo
I would say your hottest hot takes.
00:00:02 - Scott Steinlage
Not hot takes. Again.
00:00:13 - Christopher Burns
We are still at Remix Conf. They have just wrapped the main conference, and we're now in between that and the afterparty.
00:00:23 - Anthony Campolo
We can't start the episode like this. Okay, we have to do it the same way we start every episode, which is: Ishan, Scott, welcome to the show. What's up? Thank you. Good job answering at the exact same time. Perfect.
Let's do that again. Okay, so let's go. You first and you first. So, Ishan, Scott, welcome to the show.
00:00:43 - Ishan Anand
Thank you. Glad to be back here a second time.
00:00:45 - Scott Steinlage
Yes, absolutely. Second time for me as well.
00:00:48 - Ishan Anand
And it's really fun to actually be doing this in person, live, totally in the same room.
00:00:53 - Anthony Campolo
Yeah, we had Ishan with Mark quite a while ago now, last year, to talk about Layer0, which has now been acquired. We'll get into that. And then, Scott, we talked to you just two days ago with Will, so that episode will air before this one. We're at Remix Conf, and we're really excited to have you both here to talk about both Remix and what's going on with Layer0 today. We'll have a lot of topics to get into.
00:01:18 - Ishan Anand
Yeah, I think the key thing with Remix Conference is that it's one of the first in-person developer conferences. There have been maybe two or three others so far. Aside from the technical topics, it's just amazing to see everyone. Everyone I've talked to has been so excited and energized by seeing the people they haven't seen in two years. This is my first in-person conference in like two years, and it's been really invigorating. Everyone I've talked to has basically echoed that.
There are people I have only met in the last two years purely online, like you guys. I mean, even my own coworker. And now we get to see each other in person.
00:01:56 - Scott Steinlage
Which is awesome.
00:01:57 - Ishan Anand
Yeah, it's just totally great.
00:01:58 - Christopher Burns
I find it crazy to think we've all got legs.
00:02:01 - Scott Steinlage
[unclear]
00:02:01 - Anthony Campolo
We do. We've only seen each other from the torso.
00:02:04 - Christopher Burns
Yeah, well, it's interesting because not only do we have legs, we're also all different heights. When you're sitting at a desk, you're like, are they taller than me, or am I taller than them?
00:02:11 - Anthony Campolo
Just so everyone knows, I'm the tallest.
00:02:13 - Christopher Burns
Anthony is the tallest.
00:02:14 - Scott Steinlage
I don't think so.
00:02:15 - Ishan Anand
But yeah, Scott seems like he might. Yeah.
00:02:17 - Scott Steinlage
Six foot.
00:02:18 - Anthony Campolo
Yeah. Six-one.
00:02:19 - Scott Steinlage
Okay, well, I guess he's got me.
00:02:20 - Ishan Anand
Oh, yeah.
00:02:21 - Scott Steinlage
I don't know. Or I could be six-one and just not know it.
00:02:24 - Ishan Anand
And just be modest.
00:02:25 - Anthony Campolo
Yeah, I actually met one of my old coworkers here, Sam from StepZen, and we are the exact same height. We're both six-one, and we looked at each other. We were like, right on.
00:02:37 - Scott Steinlage
There you go.
00:02:37 - Ishan Anand
And you hadn't met in person before?
00:02:39 - Anthony Campolo
No, no. I met Anant and some of the team who were in the Bay Area at the time, and I was going into the office like once a week at StepZen. It was not fun.
00:02:48 - Ishan Anand
You know, this is somewhat off topic, but somewhat on topic, just the whole remote-work dynamic. Layer0 had been remote-first pre-pandemic for the longest time.
00:02:59 - Anthony Campolo
We were already ready to go. You were set.
00:03:01 - Ishan Anand
I mean, we were essentially set up for going remote-first already. I was already in a different location than where the headquarters were.
00:03:09 - Anthony Campolo
I think that's a competitive advantage.
00:03:11 - Ishan Anand
I think it has been, and it was. But COVID really forced a lot of companies to catch up to that.
00:03:18 - Anthony Campolo
Which I'm very thankful for. I could not work in an office. I will only ever do remote work.
00:03:23 - Ishan Anand
Yeah, I mean, there's lots of people who say that. And I think part of what's fueling that so-called Great Resignation is, from all the data, people switching for new opportunities that have opened up, partially through remote work, or partially just because their existing position was gone and that forced them to reevaluate and say, oh, maybe I should look for something else.
One of the things we did, though, is we still made sure that once a quarter, or three times a year, we'd all get together in person just to kind of cement and build up some social capital that we would basically be using throughout the rest of the year. It was really invigorating. There were certain things we could only get done when we all had those on-sites together. This kind of feels like that same feeling, except for the JavaScript community as a whole, where you walk in and you're like, oh yeah, I not only enjoy my work and the people I work with, but I really just enjoy hanging out with them.
[00:04:13] It feels like that's the same thing happening here.
00:04:15 - Anthony Campolo
Yeah, and that is kind of the double-edged sword of remote work. Even though I say I wouldn't want to work in an office, I can only do remote work, I'm so invigorated by this conference and getting to meet all the people in person and just hang out in a room with you guys. That's really great.
And it's such a bummer that we all had to spend like two years just not having social contact. Humans are social creatures. We thrive on social contact. For us, not having social contact feels like you've been jettisoned from the tribe, which back in the day would mean you're going to die.
00:04:44 - Ishan Anand
Yeah.
00:04:45 - Anthony Campolo
Your survival used to depend on being in a social group.
00:04:48 - Ishan Anand
Yeah. I will say there's another double-edged sword to it, having done this for a while, and that's two things. One is, if you're distributed and you're remote, the time zones can kill you.
00:04:58 - Anthony Campolo
Time zones are complicated.
00:05:00 - Ishan Anand
Yeah.
00:05:00 - Christopher Burns
Tell me about it.
00:05:01 - Ishan Anand
Yeah, yeah.
00:05:02 - Anthony Campolo
We've been working around that the entire time we've had to do this.
00:05:05 - Christopher Burns
Yeah.
00:05:05 - Anthony Campolo
I had to wake up at 9:00 to do podcasts. 9:00. Who has to wake up at 9:00?
00:05:11 - Ishan Anand
I wake up at 6:30 a.m.
00:05:12 - short split/interjection
You say that. You've literally said this.
00:05:14 - Christopher Burns
It's my favorite activity to do. You're like, this is my favorite activity to do.
00:05:18 - short split/interjection
And so...
00:05:18 - Anthony Campolo
I wake up at 9:00 to do it.
00:05:19 - short split/interjection
[unclear]
00:05:20 - Christopher Burns
9:00 is late.
00:05:21 - Anthony Campolo
I am not a morning person. Night owl. I stay up till midnight every day, at least, sometimes 1 or 2.
00:05:28 - short split/interjection
[unclear]
00:05:28 - Christopher Burns
I feel like I'm really burning the candle here. I go about 1 a.m., 2 a.m. UK time, and then I get up like 8 a.m. UK time, like seven hours, six hours every single day. I need to sleep more. I think that's the answer.
00:05:42 - Ishan Anand
For us, the center of mass for the team is something that's worth calculating and figuring out. What's the best time zone? My back-of-the-envelope on this is, at least for us, the best is somebody who lived on the East Coast and wakes up early. For us, we span all the way from the West Coast of the United States all the way into Eastern Europe. That's a pretty broad swath. We've still got other folks even further around the globe in regions like APAC and so forth.
We have to sometimes be on odd times just to communicate with other teammates. My schedule, even though I work on the West Coast, is an East Coast schedule. I wake up East Coast time and I communicate.
00:06:23 - short split/interjection
I think it's really challenging.
00:06:24 - Christopher Burns
And this is obviously a podcast about JavaScript, but I take my hat off every day to my partner, who doesn't care if I work late. Working late is not, I wouldn't say it's a normal thing to do, but when you're a founder, you're always working. It's that thing of, like, when do you make time for your partner, who works 9 to 5, but you're speaking to everyone who is seven hours behind you?
So what? Your workday is 10 till 8. And then how do you make time for your partner? I personally think that remote work is an amazing idea, but the execution is the hardest thing ever. I believe the only time remote working can properly work, in my idea of working now and also my experience of doing the podcast and all these other things, is when you work in teams in close enough locations. For example, I think that's what Prisma does.
[00:07:20] Prisma is in Berlin and San Francisco. All the people around Europe work European time, and all the people in San Francisco work American time. I think that's really good. But then the problem with that idea is that when communication happens, you're speaking to San Francisco. That is dawn in the morning for them. In Europe, that's nighttime. You have people that have had their workday and want to go home, and people just starting. There's no winning, in my eyes.
00:07:48 - Ishan Anand
This actually does have a technical angle to it, which is the ideal is almost like microservices. When you're decomposing a monolith, you want to make sure you cut the lines of responsibility appropriately to the right modular components. If you can organize teams geographically appropriately, then you can kind of get that modularity. You just have to make sure they're composable business units or composable teams, kind of like the AWS model of the two-pizza rule for how big your team is.
One of the things we've done is we've kind of slowly had a shift of a lot of our engineering being US West Coast-based. Now we actually have a locus in Tallinn, Estonia, and that's where one of our engineering teams is. We have another engineering team that focuses on different parts of the data pipeline, and then another part of Eastern Europe. That has actually helped us work out. So it's almost like that same microservices thing. There's that famous law, Conway's Law.
[00:08:39] Right? Your organization can't help but build a product that reflects it. So you have to make those judicious decisions in your HR organization. It's an HR thing, but it actually does impact your product in a lot of ways.
00:08:51 - Anthony Campolo
It's funny you mentioned Conway's Law and the two-pizza team. Remember we talked to Monarch about this all the way back on episode three? We talked about those topics, and it's so true. You really have to be aware of that.
But I think the benefit is, if you have a global user base, if everyone's asleep for eight hours, what happens when something goes down? The benefit of having your support team distributed across a 24-hour time span is so beneficial because then there's always someone there to answer a question and can jump in and help out. That's why I think having people in different time zones is much more of a virtue.
00:09:29 - short split/interjection
But that absolutely...
00:09:30 - Christopher Burns
...is a very specific job. I'm not saying I don't respect support because we need it.
00:09:34 - Anthony Campolo
People are gonna need support for...
00:09:36 - short split/interjection
Exactly.
00:09:36 - Anthony Campolo
...one...
00:09:36 - short split/interjection
...day.
00:09:37 - Christopher Burns
Exactly, exactly. But support can be done without much knowledge of relationships between milestones. Say, for example, you're in America, I'm in England, and we need to pair program. Who's going to sacrifice their time? Am I going to sacrifice my evening, or are you going to sacrifice your morning? I don't think there's an answer to this.
00:09:56 - Anthony Campolo
Pair programming is not support, though. Support is when a customer needs help.
00:09:59 - short split/interjection
Yes.
00:09:59 - Christopher Burns
No, no. Sorry. I was trying to say that support can be very distributed. But when it comes to development, I think it's much harder to be distributed because we are social animals and we can fix problems faster when we have people around us who have similar experiences. And if I have to wait six hours for my friend in America to wake up, is that six hours of my workday gone? It's a really interesting question.
The reason I think this is so interesting is because programming, even JavaScript, is so America-dependent. You don't realize it because you're here, if that makes sense.
00:10:34 - Anthony Campolo
[unclear]
00:10:35 - Christopher Burns
No, no. Most programming languages are just America-centric. The first place any conference has probably ever been is probably America. Remix is in America. Obviously there's some big conferences in Europe, but it's not the norm. It's normal in America because that's where most programmers are. Far more programmers live in the United States than most everywhere else.
00:10:56 - Ishan Anand
Are there other aspects? You're right, I'm a fish in water, so I don't realize that besides the conferences. Is it dealing with the support of some company? Is it that you're getting only the tier-one support, but it's hard to get to the experts?
00:11:12 - short split/interjection
Trying to create community.
00:11:14 - Anthony Campolo
Or dead zones for like eight hours a day.
00:11:16 - short split/interjection
Yeah, yeah, yeah.
00:11:17 - Anthony Campolo
That's a big problem for Discord. People I know will say this, like people who are hanging out or living in India or that kind of area. They're just like, I have no one to talk to when I'm on Discord, you know?
00:11:28 - short split/interjection
Yeah, yeah.
00:11:29 - Christopher Burns
This is actually really prevalent to JavaScript Jam. I would love to sit there and spend time in these Twitter Spaces, but they're all more or less American times.
00:11:39 - short split/interjection
They are.
00:11:39 - Christopher Burns
And for me, that's 9:00 at night, and I'm either...
00:11:43 - short split/interjection
...like...
00:11:43 - Anthony Campolo
You went to bed after midnight.
00:11:44 - Christopher Burns
Well, I do, but I like to give my partner some time. And I hate to say it, but if I didn't have my partner, I'd probably be burnt out by now. She's the...
00:11:54 - short split/interjection
...one that you...
00:11:55 - Christopher Burns
...know. She's the one that says, just turn it off. That's frustration from my side, but I understand why she says that. And I really loved being on JavaScript Jam. But that's the thing. Why don't we host these at UK times or European times? It's because Europe just... we don't work like that. Development is so different on that side of the world in the middle.
Distributed jobs are amazing. The idea to work for an American company and live in my home country is amazing, but there's definitely downsides to it. Personally, I'm a home worker, but I feel like I would like to work in an office. I personally feel like I'd like a balance: go to the office and communicate when I want to, but also go and stay at home when I want to keep my head down. You know what I mean?
00:12:40 - Ishan Anand
Yeah, I've often thought about the right model. I kind of think of it like there are two models in my head for how this could work. One is like an Avengers model. You have everyone distributed around within a five-hour driving radius because, you know, when you've got to ship something, so much gets done when you're all physically in the same room and you're just like, let's bang this out for this week. And so you can say, when that team or that microservice needs to ship whatever their core part of the product is, let's just get together, we'll find a WeWork, we'll get this done, and then you can go back and you can get those on-sites with a more frequent kind of pace.
The other one is this Paul Graham maker-manager essay that he wrote a while ago, which is the programmer or developer needs lots of focus time, a manager's whole job is meetings. Try to do both, you're going to fill up your day with meetings, and you won't be able to get the focus time you need for development.
[00:13:23] I think something analogous happens in organizations. Is your organization synchronous-based, which means things get done in meetings, or is your organization asynchronous-based, where you can do everything through Slack or email? There are some companies like Automattic that are very asynchronous-based. It's been virtual and remote for a long time. I think GitLab is another one, if I'm remembering correctly.
00:13:48 - Anthony Campolo
Yeah, GitLab is 100% async. I think you have to be as much async as possible, and you should only go to synchronous when it has to be something that needs to be communicated face to face very quickly. But I think, Scott, you had something to say on this.
00:14:00 - Scott Steinlage
Yeah. I mean, actually I was going to say something about what Chris was saying with the whole, okay, I'm in the UK, I'd like to get on more of these things, right? But I can't because the majority of these things are hosted at a time that's just not convenient for him. And the funny thing is, when we were running that Clubhouse thing, we had a lot of people from India and stuff like that, and it was like 1 or 2 in the morning for them. They couldn't really talk because it was late, and they could only type into the chat.
So when we introduced the chat, we actually got more conversation going. The interesting thing I just thought about was, think about it like, what does Google say? They're like planning for the next, what is it, million users or whatever. That's what their focus is, right? That's their big thing. I think it's a million. But anyway, that's a number.
[00:14:41] They're always trying to be more inclusive with things, right? And so is everybody. It's just kind of a big thing nowadays, like really inclusivity, accessibility, all that together. My thought is maybe that's something we should all be thinking about more. Maybe we should host something at a time that's more convenient to those people.
Because here's the thing. Chris said that the majority of people who are engineers and developers are in the United States. But I think you might be wrong. I think you're close, but I think that India has a lot of developers that we just don't get in touch with because there's a lot of people there.
00:15:14 - Christopher Burns
When I say...
00:15:14 - Scott Steinlage
And...
00:15:14 - Christopher Burns
...elsewhere, a lot of developers in the United States because it's a lot of the known developers. It's so much easier to... how do I make this not about UK versus America again? But it's just so different here and so much more approachable.
I think the biggest thing I can say here is that, great, you want to host things at times that I could deal with, but then you're sacrificing your life.
00:15:38 - Scott Steinlage
I get that, I get that.
00:15:39 - Christopher Burns
But I mean...
00:15:39 - Scott Steinlage
But okay, so there's got to be something then. Maybe it's that we help other people in, say, India or the UK or whatever, to take the initiative and help them to build these things up or something. I don't know. I mean, there's probably people that do this already. It's just not as many, I guess, right?
But that's how Google got so many more users. They changed YouTube to make it accessible on these slower-network phones in India, right, and stuff like that. And it increased their user base by hundreds of millions of people, right, which is insane. So it's like, if you could do that same thing, but from the perspective of what you're talking about, then holy cow, there could be a lot of people that you're missing out on.
00:16:31 - Christopher Burns
But it's also that social media, as an internet thing, is very Americanized here.
00:16:38 - Anthony Campolo
The American social media networks. Like, China has their own social networks, which are completely different.
00:16:43 - Christopher Burns
But, like, for example, Europe, we don't really have any social networks that are like, this is a European social network. We use Twitter, we use Reddit, we use Facebook.
00:16:52 - Scott Steinlage
Well, there you go. That's your next thing, Chris.
00:16:55 - Anthony Campolo
Your next startup idea.
00:16:56 - Scott Steinlage
Yeah.
00:16:57 - Christopher Burns
It wouldn't go anywhere. It honestly wouldn't go anywhere. Trust me. Everyone loves Facebook. No, sorry, TikTok. But what I'm trying to say is that I think the idea of distributed manpower is such a powerful idea, but it's not something that can just happen. I think it has to be worked on, and I think that's part of people's jobs.
My job is to make sure the machine of distribution runs smoothly, because obviously this is on the outside. But when I try to think about all these things, I'm like, you don't just say, yeah, we're all distributed, and that's it. You work from home. Yeah. It's never just that simple.
I think there's so much to learn and digest. I feel like today I'm digesting so much knowledge about America as a whole, and it's changing my opinions on everything. People keep saying, "America is not this," you know, because I say, "America is X." "America is not X." And I really get that now.
00:17:54 - Anthony Campolo
Yeah. No, it's a hard lesson to learn. I'm sure it's the same with England. No country is one thing. It's an amalgamation of millions or even billions of people, and they don't all agree.
00:18:07 - Christopher Burns
Yeah. Let me give you an example of how code is Americanized.
00:18:12 - Anthony Campolo
By default it's English, first of all.
00:18:14 - Christopher Burns
Yeah, but it's not UK English. Like colour, C-O-L-O-U-R. I believe that didn't used to work. You just wrote color. But that's how we spell everything, you know.
00:18:28 - Anthony Campolo
[unclear]
00:18:30 - Scott Steinlage
The dollar symbol. Remember that one?
00:18:33 - Christopher Burns
That's jQuery. Okay, the dollar symbol.
00:18:35 - Ishan Anand
Symbol.
00:18:35 - Anthony Campolo
You're talking about when we were asking Chris if we used a pound sign when you wrote jQuery.
00:18:41 - Christopher Burns
And I said it's the jQuery symbol. That's all I know it for.
00:18:47 - Ishan Anand
That is a good one.
00:18:49 - Christopher Burns
But what I mean is, say when you work in a distributed workforce. Say you have British people, you have American people. How should something be spelled? Should the British person write a function called capitalisation with an S? Or should I write it the American way with a Z? It's a really good question.
00:19:09 - Anthony Campolo
I mean, Americans wouldn't even say Z.
00:19:10 - short split/interjection
Yeah, and...
00:19:11 - Christopher Burns
...you...
00:19:11 - short split/interjection
...write it with a Z, and if you're...
00:19:13 - Ishan Anand
And if you're in JavaScript, that could be a runtime error if somebody is totally using one spelling.
00:19:19 - short split/interjection
Versus your entire...
00:19:20 - Anthony Campolo
App could break if...
00:19:20 - short split/interjection
...you don't have any TypeScript. Yeah.
00:19:22 - Christopher Burns
Here's an example. I actually, by mistake, had both versions in my app: the same function in two different files, one the UK spelling, one the American spelling. And I was calling it in different places because I wasn't realizing I was doing it. You know what I mean? Because I'm...
00:19:37 - short split/interjection
Yeah.
00:19:37 - Christopher Burns
I'm not bilingual because it's English, but you don't realize it.
00:19:43 - Ishan Anand
This is really where TypeScript...
00:19:45 - Christopher Burns
Yeah, this was TypeScript. This was TypeScript saying, oh, you got it here. You just don't realize.
00:19:51 - Ishan Anand
Are you conscious of this, or does this build up? Is it forefront in your mind, or do you only realize it after a while, like it's creating little bits of friction, right? A lot of paper cuts.
00:20:01 - short split/interjection
Yeah.
00:20:02 - Ishan Anand
And imagine, I'm sure it's even worse for somebody who isn't even in the English-language sphere. But how are you mentally aware of it? Do you feel it, or do you only subconsciously?
00:20:13 - Christopher Burns
Yeah, I feel it all the time in very unique ways. Let's take Stripe, for example. Stripe is an international platform, but when it comes to writing addresses, the field is called zip code. ZIP codes are a US thing. Everywhere else calls them postal codes. We don't write zip codes; we put postal codes.
The other one, when it comes to addresses, is state. We don't use states in the UK. We have regions, but they're called counties, like Lincolnshire, Leicestershire. There's no answer. There's no answer to how everybody can win.
But let me give you an example where I've seen a win that made sense. This happened actually in the Remix workshop. They had the default native web calendar picker for dates. I looked over at the person next to me's computer and he saw it in the US format, month/day/year, whatever. But in my view, I had it in the UK way, day/month. They're not big numbers, but you know what I mean. I did it the opposite way, the way we do in the UK.
And that seemed right. The web was smart enough to go, I just know what that means. If I was to write CSS, say if I wrote color, surely the web should be smart enough to go, yeah, I know what that means. That means he wants to change the color of the text. I believe that is actually in the spec now, I think.
[00:21:33] But it's a really interesting thing about web development. I think this is a very British thing because we have our own unique spellings of American words. But say, if you were not a native English speaker, you would just default to the American way of spelling things, I believe.
00:21:55 - Ishan Anand
Well, no, maybe not. I mean, look at countries like India where the language has been influenced effectively by both countries.
00:22:03 - Anthony Campolo
I think Canadians tend to lean toward UK spelling. I may be wrong about that.
00:22:07 - Ishan Anand
Yeah. I mean, well, most of the Commonwealth.
00:22:09 - Scott Steinlage
French over there a lot.
00:22:10 - Ishan Anand
The Commonwealth countries, it would make sense. That's really fascinating. I'm walking away from this conversation thinking about accessibility in a different way, in terms of not just the technology but also in terms of community building, as Chris was bringing up. What you're talking about there is in terms of the web, but I think you really meant the browser, because it was localized.
00:22:30 - short split/interjection
Yeah.
00:22:30 - Ishan Anand
And that speaks to all the work around making the platform or your product more accessible, whether it's a browser or whether it's YouTube. Accessibility is about performance, sure. This is another dimension of it that I just haven't considered.
00:22:45 - Christopher Burns
I think the best thing always is: you will never fix it, but be mindful of it. I think that's true about most problems in the world as a whole.
00:22:54 - Anthony Campolo
I mean, we could fix it if we wrote crazy, complicated logic to handle all these edge cases. It's just a matter of, is it worth investing the time to make both spellings correct? Or does it make more sense to have social coordination to decide on a spelling?
00:23:11 - short split/interjection
[unclear] My guess is it's worth it. There's something really cool...
00:23:15 - Scott Steinlage
...that I think could come out of that.
00:23:17 - Christopher Burns
But for example, surely it must be crazy for someone who's French. They just speak French and they're like, yeah, but if I want to write some HTML now, I have to write in English.
00:23:27 - Anthony Campolo
So we should write all of our programs in Esperanto, I bet.
00:23:33 - short split/interjection
Yes, Esperanto is the...
00:23:34 - Ishan Anand
...universal language.
00:23:35 - short split/interjection
And...
00:23:35 - Ishan Anand
...we do it in Java.
00:23:36 - short split/interjection
So it's...
00:23:37 - Ishan Anand
Esperanto. Yeah, you need a virtual machine that will run anywhere, and you can only program in Esperanto.
00:23:45 - short split/interjection
Oh my gosh.
00:23:46 - Anthony Campolo
This is a really interesting conversation, but we're already halfway through our episode, and we should talk about some other stuff. Beforehand it was like, are we gonna... what are we gonna talk about? We're gonna talk about plenty of stuff.
So I would love to really get into what the state of Layer0 is today. Because the last time we talked, you had not been acquired yet, and now you've been acquired. So that's a very big change.
00:24:06 - short split/interjection
Yes.
00:24:06 - Ishan Anand
Let me first set the stage for the audience who hasn't heard about Layer0. Layer0 is basically a Jamstack platform for, I'd like to say, I don't use the word enterprise, but high-stakes websites. Large sites that have lots of pages, or they change frequently, where you want to have the performance, the scalability, the reliability that you get from the Jamstack, but where traditional static techniques don't apply.
We've actually been doing Jamstack, in a sense, before it was called Jamstack, but it wasn't static. We were doing serverless Jamstack, including, at one point, building our own version of AWS Lambda in order to effectively do a serverless-first approach to it.
So that's Layer0. We've got a lot of clients, a couple household names. I always have to check myself which ones we can say, but if you just go to our website, Layer0, you can see the list of customers. That's Layer0. When we last talked to you, that was our focus, and it remains our focus.
[00:24:59] What happened more recently is this: Limelight Networks is one of the largest CDNs out there. They have, I think, the second-largest CDN network out there, but they've been very focused on video for the last five or ten years.
00:25:12 - Anthony Campolo
It's interesting because I think most people who are web developers today, when they hear CDN, they've been kind of trained to think of it in the Netlify/Vercel kind of sense: a CDN is a place where you put static HTML, and that's what it's for. So that's not what a CDN is. There's a lot more to CDNs than that.
00:25:28 - short split/interjection
Yes.
00:25:28 - Ishan Anand
And there's a lot more it can do. In the case of Limelight, what they've been doing is powering most of your streaming services. In fact, most of the people listening to this have used and consumed video content. If you've gone to almost any streaming service that was powered through Limelight, that's why they've got such great bandwidth and infrastructure.
00:25:45 - Christopher Burns
If Limelight is number two, is Akamai number one?
00:25:48 - Ishan Anand
On the size of the network.
00:25:49 - short split/interjection
Because...
00:25:50 - Christopher Burns
They're CDNs.
00:25:50 - short split/interjection
Yes, they are.
00:25:52 - Christopher Burns
But that's the thing. They're classic. Like, 30, 40 years.
00:25:56 - short split/interjection
Yeah. So...
00:25:57 - Ishan Anand
Well, it depends how you're actually drawing the rank. The rank could be by revenue. It could be by number of POPs. It could be by capacity.
00:26:04 - Anthony Campolo
Number of bits.
00:26:05 - Ishan Anand
Number...
00:26:06 - Scott Steinlage
Oh my...
00:26:06 - short split/interjection
...gosh.
00:26:07 - Ishan Anand
I mean, there's a lot of ways, but they have a massive network that's been built for video, and that's where they're most known.
What they wanted to do was two things. They had a new CEO who came in and he's like, it's clear where the hockey puck is going. A couple of years ago, he was like, CDNs have to become edge platforms. Even before we were hearing here at Remix Conf a lot about the edge, the CEO, Bob, clearly realized that's where the hockey puck was going. He believes not only that Jamstack is the future, but that edge is really key to the future.
The analogy I like to make is, it's kind of like an airline. Limelight wanted to move into the web CDN space and the edge platform space in a very bold manner and enter that market. When you think about it like starting an airline, you need to have a certain amount of planes. You need to have pilots.
[00:26:49]
00:26:51 - short split/interjection
There's a lot of expertise.
00:26:52 - Anthony Campolo
I've thought about starting an airline...
00:26:53 - short split/interjection
...before, but...
00:26:54 - Ishan Anand
Imagine you're FedEx or you're UPS. You already have the pilots. You already have the airplanes. What you need to do is kind of change out the software, but you could use the same infrastructure you already have. You've got essentially the moat already built.
That's what was behind us joining up with Limelight, to create not just the next great web app CDN platform, but actually the next great edge platform. To that end, we actually have publicly announced, although SEC warning, I suppose this is not officially closed.
00:27:24 - Anthony Campolo
This won't go out for like a month.
00:27:25 - Ishan Anand
Oh well, it might have closed by the time this is out. We're doing two things. First, there's an acquisition of Edgecast, which is another CDN that's just going to continue to increase our scale in both capacity as well as headcount and the ability to get development velocity out there. And it's also going to... you know, we're going to rebrand ourselves as Edgio to emphasize the new direction.
00:27:47 - short split/interjection
I like that name.
00:27:48 - Anthony Campolo
That's a good name.
00:27:49 - short split/interjection
Yeah.
00:27:49 - Scott Steinlage
E-D-G-I-O, yes.
00:27:52 - Ishan Anand
Thank you. Yeah. So the webpage is up already. That's going full scale ahead right now. So we're at a very exciting and interesting point.
I like to tell people, when they ask what it's like when you've been acquired, it's kind of like going from high school to college. Getting in was an accomplishment, but you've just signed yourself up for four years that are actually going to be operating at a much larger scale than you were before, which is both exciting...
00:28:19 - short split/interjection
Yeah, I've...
00:28:19 - Anthony Campolo
...heard it's like a marriage.
00:28:21 - Ishan Anand
Yeah, it absolutely is.
00:28:22 - short split/interjection
For sure.
00:28:22 - Ishan Anand
In a lot of different ways, culturally and organizationally. There are some things that help. They were already distributed and remote as well, but now we're distributed and global. They're even more distributed and global. I mean, they're a publicly traded company. They're very large.
They've got more presence historically than we've had in regions like APAC, which is Asia Pacific. You were talking about support. It's not just that you need to have the support around the clock. You have customers in Asia and they want their customer support representative, that they build a relationship and a rapport with regularly, and their project managers on the other side should also be on their time.
00:28:55 - Anthony Campolo
Even speak their language.
00:28:56 - short split/interjection
And...
00:28:56 - Ishan Anand
...and can speak their language. So you need to make sure your documentation, your planning, localization, and things like that.
00:29:01 - short split/interjection
[unclear]
00:29:02 - Ishan Anand
Yeah. A lot of what you were talking about actually kind of resonated, but that's also the opportunity, and it's been really exciting.
00:29:08 - Anthony Campolo
Yeah. We're all about the edge here. I've been pushing the edge and edge-native architectures for a really long time, so super happy to hear that. I've been loving all the Remix stuff about the edge, like Deno Deploy and Cloudflare and all that stuff. I think the real benefit is that you just don't have to think about where your thing is anymore. You can assume it's everywhere.
It simplifies the mental model for the developer in such a great way because you don't have to worry about, okay, I'm going to deploy it here, and then that means it's going to have good performance here but bad performance here. That's overhead that you shouldn't have to think about as a developer.
00:29:49 - short split/interjection
I have...
00:29:50 - Ishan Anand
...a little bit of a contrarian view on...
00:29:51 - short split/interjection
...that. Okay.
00:29:52 - Ishan Anand
So I think we're still sorting that out.
00:29:54 - Anthony Campolo
Because of the database.
00:29:55 - short split/interjection
Right.
00:29:56 - Ishan Anand
Because, well, a couple of things. There's the database, but let's leave the database out of it for a second because that is a big area that still hasn't been solved yet in the ecosystem.
Let's just take server-side rendering as an example. Does it make more sense to server-side render your page once and distribute it out onto a cache to 100 global POPs around the world?
00:30:14 - Anthony Campolo
That's still the edge to me.
00:30:15 - Ishan Anand
Well, that is, but the edge isn't actually doing the computation. The edge is doing the caching and maybe the cache key normalization, which is one of the things.
00:30:21 - short split/interjection
[unclear]
00:30:21 - Anthony Campolo
I don't care what's happening where. The fact is that something ends up on the edge.
00:30:26 - Ishan Anand
What I'm seeing in the ecosystem is a lot of interest and excitement about doing everything at that edge node. Imagine you did server-side rendering on the edge node. That may be the right thing to do. It may not be.
So let's take your homepage. If your homepage isn't personalized, or maybe only parts of it are, it may make sense to process the server-side rendering at one node and have that one node give it to all the other nodes, rather than say, I'm going to have all those 100 nodes run server-side rendering. That is more efficient, and it's probably going to cost you more CPU time collectively. There needs to be a judicious balance of when it makes sense and when it doesn't. I still think the ecosystem is learning that.
00:31:03 - Anthony Campolo
And this is why we have to create the right conventions so that the things that need to be done at the edge are done at the edge. The things that don't need to be on the edge don't need to be on the edge. The developer should not have to think about that.
00:31:13 - short split/interjection
I agree.
00:31:14 - Christopher Burns
I was gonna say, a really good example of why the edge will be useful is because we live in America-centric land. Whenever most startups we use, we're like, ah yeah, you know, we'll just host in US West or US East. And then you think about it and you're like, oh yeah, every customer outside the US, every single request is going all the way to the United States, then all the way back to the country of choice. And you're like, yeah, but speed of light, it's all like milliseconds and time. If I've got to get data from the United States to process something, and then I've got to add my processing overhead, and then say, if I've got to send data back.
If we're looking outside the sphere of influence of America, edge computing will be very successful in terms of, I won't have to take that data to America and back. But there are so many questions. How does that change?
00:32:03 - Scott Steinlage
GDPR for you?
00:32:05 - Christopher Burns
This is what I'm gonna say. For example, we talk about distributed databases, but customers in certain countries want their data in a certain country. Yes. In my area, I know nonprofits in most countries will want that data to be hosted in that country. That means I've got to spin up a database in each country. How do you run all that management? You know, that's something I don't do right now.
But when I speak to PlanetScale, they kind of go, it's just one database everywhere. And you're like, yeah, but how are you gonna explain that to your customers who are not technology-savvy? And I go, I want my data in the UK. Oh, yeah, but your data is also in 50 other nodes around the world.
I think this is a really interesting question. Like we were saying about whether computation could be in a select few places that then gets distributed. You know, we work in areas like the Americas, Europe and Africa, Asia.
00:32:59 - Scott Steinlage
Well, you can turn POPs on and off, right, for different...
00:33:01 - short split/interjection
...things.
00:33:01 - Ishan Anand
You can. I mean, it depends on the platform you're on. This is actually the use case for the edge that I don't think most people pay attention...
00:33:09 - short split/interjection
...to.
00:33:09 - Christopher Burns
The legal...
00:33:11 - short split/interjection
...the...
00:33:11 - Ishan Anand
Legal and compliance. Somebody put out, it's like the hierarchy of needs for developers, and compliance is actually the foundational one. You don't think about that as much.
I think we're going to talk about the Composability Summit. We're actually looking for a speaker specifically to talk about composability and compliance for exactly this reason. When you're no longer dealing with the monolith and your data is going to be in different places, is it a disadvantage or is it an advantage? It could be either or both. Now your data is distributed, it might be easier to handle GDPR. But then you have this management overhead of making sure those things don't get violated.
It gets down to even the routing level to make sure that even if you've got an edge node in one place and all it's serving is cached data, can you legally serve the data from that? When are you a processor or not in CCPA or GDPR is a really important question.
00:33:59 - Christopher Burns
It's not only that, it's the biggest gray area that I still don't have an answer for. For example, have you heard of, I think it's called the US Data Shield? Is that even legally binding to an American company? Who knows?
A great question is when you work with enterprises in the UK, they'll have a DPO. They'll go, show me your services, and you go, oh, we use three services that are in America. What do you mean? They're not all in the UK? And then I think, but you're not asking Facebook to show you their services, and they are going to do a lot worse things than I'm going to do. But that's because of the size and scale.
00:34:33 - short split/interjection
I'm...
00:34:34 - Ishan Anand
...pretty sure they're asking...
00:34:35 - short split/interjection
...Facebook. How do you...
00:34:36 - Anthony Campolo
...know they're...
00:34:36 - short split/interjection
...not asking? Yeah, they're calling them up too. Yeah.
00:34:40 - Ishan Anand
This is... I remember, was it the Data Shield? There was this policy before GDPR.
00:34:45 - Christopher Burns
The Data Shield was proven to be not legally binding.
00:34:47 - Anthony Campolo
GDPR was invented because of Facebook. Let's be real.
00:34:49 - Christopher Burns
It is. Yeah.
00:34:50 - short split/interjection
The bigger...
00:34:51 - Anthony Campolo
Because of Facebook, because of Google, because the reason why that law was written is because of Silicon Valley, because...
00:34:58 - short split/interjection
...of the...
00:34:58 - Scott Steinlage
...amount of...
00:34:58 - short split/interjection
...data.
00:34:59 - Ishan Anand
But I think it's actually because, going back to Chris's point, everything's America-centric. Yes. It's like, imagine you don't control your currency. Imagine you don't control your citizens' data. This other entity that you have no legal arm to affect can be scary to your sovereignty. So it totally makes sense. Chris, you were going to say something.
00:35:27 - Christopher Burns
Yeah. I think these are the bigger questions. Some people would say, yeah, but worry about that stuff when you scale, and it's like, yes, but also we should all be talking about it.
To Americans, it's like GDPR is a tick box. But to Europe, it fundamentally changed how marketing worked. If you want to send me an email, I need to tick a box. If you send me an email, you could say, oh, it was out of good faith, and we could go, okay, but then if you send me another email, I can be like, this is a violation of GDPR and you should not be doing this. I could not sue you, but I could get you fined by the regulation.
Also, each country could implement GDPR separately in certain circumstances. And the perfect one, this is going to be amazing to Americans, is: what's your opinion on cookie bars?
00:36:14 - Anthony Campolo
We hate them. They're the worst. It's the reason why no one likes GDPR.
00:36:17 - Christopher Burns
But that's the thing. Cookies aren't all about GDPR. Cookies were before GDPR. GDPR is about marketing consent.
00:36:25 - Anthony Campolo
But GDPR is why everyone decided to add the cookie banners. Am I...
00:36:29 - short split/interjection
...correct?
00:36:29 - Ishan Anand
I actually thought there was a mandate on cookies before that.
00:36:33 - short split/interjection
But...
00:36:33 - Christopher Burns
There was a mandate, but no one followed it.
00:36:35 - Anthony Campolo
No one paid attention to it.
00:36:36 - Christopher Burns
Yeah, because they just went, we're a US company, you're coming...
00:36:39 - short split/interjection
...to us?
00:36:39 - Ishan Anand
Well, I thought they felt they were fine under the Data Shield agreement. But this is not the...
00:36:43 - Christopher Burns
We're not legal. We're not...
00:36:46 - short split/interjection
This is not legal. None of...
00:36:47 - Anthony Campolo
...us are experts...
00:36:48 - short split/interjection
...on this.
00:36:48 - Christopher Burns
I feel like...
00:36:49 - short split/interjection
...not.
00:36:50 - Ishan Anand
Legal or financial advice.
00:36:51 - Christopher Burns
And this is actually a really good question. I think there's a startup here that could make a lot of money answering tech-lawyer questions.
00:36:58 - Anthony Campolo
I bet there's a dozen.
00:36:59 - short split/interjection
I'm sure.
00:37:00 - Christopher Burns
You stick in a box: I'm a UK company. I'm going to be operating in America. What do I need to tackle? You know, these 20 things?
00:37:08 - Ishan Anand
I think that startup is called in-house counsel at the FAANG companies. Regulations, although they're designed to help people and they do help, I do worry they create a burden for startups, and it's going to make it harder for them to compete. The larger companies will figure out a way to be compliant. It's actually harder for the smaller companies to have that.
00:37:27 - short split/interjection
[unclear]
00:37:30 - Anthony Campolo
Make sure we get into composability before we close it out here.
00:37:33 - Christopher Burns
We quickly spoke about that. But what is it? Tell me more. I do not know much. I do not know anything. Give me the pitch.
00:37:39 - short split/interjection
Yeah.
00:37:39 - Ishan Anand
So composability...
00:37:40 - Anthony Campolo
The term and the conference.
00:37:41 - short split/interjection
The...
00:37:42 - Ishan Anand
...term and the conference. So let's talk about the term first. There's a number of trends we're seeing in the ecosystem that are all centered around this idea that you can build and rebuild your stack like Lego bricks, or even your whole company. We're used to calling this Jamstack. Other folks call it API-first. Forrester says API-first. The MACH Alliance calls it a MACH architecture. RedMonk has called it the smokestack. Gartner calls it a composable business or composable architecture. But it's this idea that you're now getting this ability to build and rebuild as business needs change and evolve.
00:38:14 - Scott Steinlage
And what's the new one we heard today? Netlify. What'd they say?
00:38:18 - Ishan Anand
Headless is getting rebranded in some contexts, and that's getting rebranded in a lot of places as composability. What we said is, look, they've all got this composability concept. That's our name to kind of put that umbrella on this broader concept of, hey, can I adapt my stack as needs change?
What's really interesting is this isn't just for developer happiness. There's a clear business need. When you look at... McKinsey did this survey. They surveyed, I think, 2,000 executives from a variety of companies and a variety of verticals. It was one in ten. Only one in ten thought their existing business model would last from now till 2023, which means nine out of ten think they've got to redo some part of their business. A digital experience is either going to be a central pillar to that or a buttress to that, and a really important part. So developer velocity is really paramount and time to market is paramount. That's why I think composability has such value.
[00:39:07] So we're launching this virtual summit. It'll be entirely online, called Composability Summit, at composability.dev at the end of June. I don't know, Ishan. Anything else you want to add to that?
00:39:18 - Scott Steinlage
First of all, thank you so much for explaining all that. If you go to composability.dev, there's a little bit of information there. We are going to continue updating it as time goes on. More speakers are coming on board. We are expecting to have 30-plus speakers. It's going to be really good. We have many, many already. And Anthony, yeah, for sure, we're going to talk about that. But we got some pretty good names. You want to do some shout-outs to people we have for sure on there already?
00:39:41 - Ishan Anand
We've got Kelly from Commercetools. We got Daniel from Nuxt. We got Justin from Universal Standard, who also spoke at Jamstack Conference last year. We've got a few other folks we'll be adding to the website soon.
00:39:52 - Anthony Campolo
So something I'm curious about is when Scott had pitched me on this, he used the term microservices and micro front ends. And so I'm curious, because when I hear those terms, I hear the opposite of developer velocity to me. Like, full-stack monolith is what gets you developer velocity. Then that's why we are trying to go more toward the full-stack Jamstack. But breaking things up into different pieces is because it actually is more resilient and does have a lot of knock-on effects. But it can slow down developer velocity because you have to piece together all these different bits and pieces.
So can you help sort out in my mind, because I think I'm not quite understanding this correctly, how do microservices and micro-frontends fit into the concept of composability?
00:40:34 - Ishan Anand
Okay, so you're absolutely right. In some cases microservices are a bad idea, and in some cases the monolith is the bad idea. We actually have a planned talk for one that's called I Heart My Monolith.
So the situation you have to think about is in an enterprise context. Once you get to a certain scale, having everybody in the same monolith is going to have people stepping on each other's toes. You can't, for example, have a team deploy. I want to do an HTML change. I just want to change the text on some button. Now I need to go through this whole long DevOps process and make sure the server gets populated and the cache gets cleared.
In our previous company, I've been on these calls with large companies and they're trying to deploy and it's like 12 people on the call. So everything has to get bundled up in these releases, and your whole velocity gets slowed down in some sense.
Jamstack is kind of the on-ramp to microservices because what you do is you say, well, first let me decouple the back end from the front end. This is one of the first things we found out when we were doing, I like to say, Jamstack before Jamstack at Moovweb, which was a company that was mobilizing sites. What we saw was customers who were pushing out changes to their mobile version of their site ten times faster. There were ten more deploys per week than their desktop site, and that's because what we'd said is we gave them a platform where they could change their HTML without changing the backend.
We had decoupled the front end from the back end, and that's the first example of where... But hey, if you're in Ruby on Rails and you're one person, that decoupling might just add extra overhead, back to kind of Conway's Law.
00:42:03 - Anthony Campolo
I do think decoupling the front end and the back end makes sense. This is the whole point of Redwood. Redwood does have decoupled front end and back end, but it gives it to you in a monolithic kind of package that allows you to kind of split the difference there.
So I think if you can take a microservice architecture and figure out a way to present it to developers as if it was monolithic by a lot of really intelligent developer tooling and smart architecture within your project, you can get that same kind of velocity even though it is actually decoupled.
00:42:33 - Ishan Anand
Let me give you actually a totally different use case to just broaden what we mean by composability. If you are a very large public company, you don't have the luxury of all of the software that you are hosting, all your different properties, on the same identical stack. Imagine you're Salesforce.
00:42:49 - Anthony Campolo
Needs to be multi-cloud, or you need to be across different services.
00:42:52 - Ishan Anand
Well, let's imagine you have a bunch of very large companies that you've just acquired. Salesforce has the Salesforce site, but then they have Salesforce Commerce Cloud, which was Demandware they acquired. They just acquired...
00:43:02 - short split/interjection
Slack.
00:43:02 - Anthony Campolo
You're dealing with legacy.
00:43:03 - Ishan Anand
You're dealing with legacy. You've got a lot of different teams. And so this is one of the talks we've got planned.
Let's say you want to have the same consistent branding, not just the same logo. It has to say Salesforce. You want it to have the same UI components, but one's in React, one's in Angular, and one's in regular HTML. It's got no front end.
00:43:21 - Anthony Campolo
And that's the micro front end pitch right there.
00:43:23 - Ishan Anand
Well, that's what... I don't think... There's actually a name for it that's a little bit micro front end. But it's like, how do you maintain what we've been starting to call a composable design system, one that's totally framework-agnostic. We've got a speaker, the guy who created Agnostic UI, going to be speaking on that.
So that's again an enterprise use case of how do you create a certain level of composability where, let's say you're doing a new tiger team to launch a new product very quickly, but somebody says from corporate, it's got to have the same branding, and they want to go off and use a different framework none of the other teams have done. How do you make that work? That's another element to a composable organization.
00:43:55 - Anthony Campolo
Yeah, I think it's starting to make more sense. That was extremely helpful. For me, I'm more startup-biased versus enterprise-biased. That's why this stuff is kind of hard for me to wrap my mind around. But that was a very good pitch, and it's starting to click a little bit.
00:44:07 - Christopher Burns
I think that we're kind of going crazy deciding if it needs to be a monolith or it needs to be decoupled. As you were saying, it really depends on the people. The hottest take is that it really doesn't matter. Your customers won't care. It's not going to stop you making money. It's the product and the idea that will matter.
00:44:27 - short split/interjection
[unclear]
00:44:27 - Ishan Anand
In some cases it may be more important, not how happy your development team is, but maybe the other stakeholders who have to push out content. I have this saying that half your site is built by the non-developers populating out the content.
One of the areas Jamstack has some work to do, although it's been improving, is the tooling for non-developers in order to be able to use these platforms and not have to write Markdown and things like that.
00:44:51 - short split/interjection
That.
00:44:51 - Christopher Burns
Not even Markdown. You use a headless CMS, don't you? Like, grey forms to define everything. Every block is a grey form of grey-ness. You don't know what grey form one is, or grey form one makes grey form two. And you connect these components together and you get grey form three, and it's like, headless.
That's why visual CMSes are starting to take off. Yeah, I really think they have their place, but there's still so much there and there's never going to be an easy answer for all of it.
What I was meaning by my question and my hot take was, are we always trying to over-engineer it? Like, my product failed because I made it a monolith. No. You start off. When is the right time to actually swap models to a decoupled system? Or you may want to go the other way. I bet there's tons of enterprises that went from decoupled to monolith today.
00:45:42 - Ishan Anand
I think you have to do it based on your current context and what makes sense. I remember in the early days when Twitter was taking off and they built on Ruby on Rails, and people were like, oh, that was a mistake. And one of the engineers wrote a really good blog post. He said, no, it wouldn't have become Twitter if we didn't have the developer velocity. In order to get to the place that we got traction, the ideal path may not be the one that delivers you traction. It's path-dependent. You should go with whatever is going to make the team as productive as possible.
It may be a monolith in the early days, and definitely in a startup case where you're limited in headcount and you want to be able to hire somebody who will know those conventions. Let's say you're on Redwood: hire for Redwood. This person's going to know this and they'll be able to work across the stack.
But after the team gets larger and larger, to hundreds, or you're one of these larger enterprise companies that have thousands, tens of thousands of developers, you don't want to have everyone held up because they're trying to make a change and all the deploys have to happen at once.
00:46:41 - Christopher Burns
It's a very good point, and I think the whole ethos of this episode has been: everything depends, and it's a gray area of everything.
00:46:48 - Anthony Campolo
And that should be the ethos of any tech conversation.
00:46:52 - Christopher Burns
What do you mean? We're all drinking the Kool-Aid over here. I thought we were going to re-code everything in Remix tomorrow, and money doesn't matter.
00:46:58 - Anthony Campolo
Wait, you mean you're not done migrating Everfund already?
00:47:01 - Christopher Burns
You know, I was waiting for the codemod to do it for me.
00:47:05 - Anthony Campolo
The Redwood-to-Remix codemod. You'll be waiting a while for that one.
00:47:09 - Christopher Burns
I don't know. I hear Redwood doesn't hate Remix yet.
00:47:13 - Ishan Anand
It goes back to that saying, I don't know, something about premature optimization...
00:47:17 - short split/interjection
...the root of all evil.
00:47:18 - Ishan Anand
Yeah, exactly. Don't optimize for the problems you don't have. You don't know if you're going to have them. Wait till it happens and then optimize.
00:47:26 - Christopher Burns
Yeah. Literally you could optimize it and have the best app ever and nobody wants to use it.
00:47:31 - short split/interjection
Yeah, right.
00:47:32 - Scott Steinlage
Don't fix it if it ain't broke.
00:47:33 - Christopher Burns
For example, let's just take a moment to speak about everything that is terrible about web development. Banks, airlines. These people make lots of money. Their websites are terrible, and they're not optimizing it.
00:47:48 - Anthony Campolo
Or working on getting rid of banks. Don't worry.
00:47:51 - short split/interjection
[unclear]
00:47:54 - Scott Steinlage
Web3.
00:47:54 - short split/interjection
Drop everything. Yeah.
00:47:55 - Anthony Campolo
All right. Well, thank you so much for being here. This is a really fascinating episode. I think we're going to probably call it here, but if there's anything else you guys want to shout out, I'll make sure this episode gets out soon because Composability Summit will be something that we want people to know about. So feel free to drop some links, your own socials, anything like that.
00:48:13 - Scott Steinlage
Awesome. Yeah. Just remember composability.dev. You can follow me at [unclear] on Twitter.
00:48:22 - Ishan Anand
Yeah, I'm going to duplicate the composability.dev shout-out. Check it out. Well, I guess this will air maybe just before. If the topic excites you, talk to us. We'd love to have you either at this one or a future one. We're really excited about this area.
We think it's really interesting. Another thing that we're calling AppOps is basically that the application becomes the unit of scale, and it makes it a lot easier. It takes kind of DevOps, and what Jamstack has done, and kind of applies it broader throughout the whole app. So we're really excited about that.
The other shout-out I'll just give is Edgio. We're rebranding, so we need to make sure we get the word out. So look for that by the time this drops. And if you want to follow me on Twitter, you can follow me at [unclear]. And then the other shout-out is JavaScript Jam, which is the podcast that myself, Scott, and Mark Brocato run. We're on Twitter Spaces now.
[00:49:13] Once every week we have a live open mic. People can come and just bring up anything on web development.
00:49:18 - short split/interjection
Whether...
00:49:19 - Ishan Anand
...you're a...
00:49:19 - short split/interjection
...beginner or whether...
00:49:20 - Ishan Anand
...you're an expert, we want to hear from you. Very much audience-driven, and we do it at lunchtime in the Pacific time zone on Twitter Spaces right now on Wednesdays. So definitely tune in.
What I really like is that it's audience-driven. I get a lot of feedback about the ecosystem through it.
00:49:36 - Scott Steinlage
It's a blast. You gotta at least check it out at least once if you haven't.
00:49:39 - Ishan Anand
I mean, today we had Dan Abramov.
00:49:41 - short split/interjection
Yeah, on Twitter.
00:49:43 - Scott Steinlage
He just popped in, guys. We didn't invite him. It was pretty...
00:49:46 - short split/interjection
...cool.
00:49:46 - Ishan Anand
That was really interesting to get his take on stuff.
00:49:47 - Scott Steinlage
So it was like, how'd you find it? It just kind of popped up.
00:49:50 - short split/interjection
Yeah, he was in his car, actually.
00:49:51 - Ishan Anand
We lost reception.
00:49:52 - short split/interjection
[unclear]
00:49:52 - Anthony Campolo
I would guess because you can see Spaces that pop up on people you follow. So Michael Chan was in. He follows...
00:50:01 - short split/interjection
Oh, I'm sure.
00:50:01 - Ishan Anand
It was the algorithm.
00:50:02 - Scott Steinlage
Yes, it was.
00:50:03 - Christopher Burns
So are you saying that I spoke after Dan from React? Wow. I must be more famous.
00:50:08 - Ishan Anand
Ah, you followed him and then...
00:50:11 - short split/interjection
Yeah.
00:50:12 - Scott Steinlage
Yeah, it was a good time, though. Yeah, it was great.
00:50:14 - Ishan Anand
So those are my shout-outs.
00:50:15 - Anthony Campolo
Awesome. Well, thank you guys so much.
00:50:16 - Scott Steinlage
This was great too.
00:50:17 - short split/interjection
Thank you.
00:50:18 - Anthony Campolo
Yeah, of course.
00:50:49 - Christopher Burns
Thank you. This will always be a special memory for me. Thanks.
00:50:56 - short split/interjection
Don't worry.
00:50:58 - Anthony Campolo
All right. Cut. All right. Awesome.