Subscribe
Share
Share
Embed
Honored two years in a row as a top tech podcast by The Webby Awards, CRAFTED. is a show about great products and the people who make them. Featuring incredible founders, innovators, and makers that reveal how they've built game-changing products — and how you can, too.
What trade-offs did they make? What experiments did they run? And what was the moment when they knew they were on to something BIG?
Hosted by Dan Blumberg, an entrepreneur, product leader, and public radio host with chops as both a technologist and as a public radio host. Dan has founded startups and led product releases and growth initiatives at LinkedIn, The New York Times, and as a consultant to big banks and startups. Before getting into tech, Dan produced and guest hosted WNYC's Morning Edition, the most listened to show on the country's largest NPR station.
Listen to CRAFTED. to find out what it *really* takes to build great products and companies.
[00:00:00] Justin Cormack: A lot of companies have a customer obsession value. but we really wanted to make it clear that, the person we think about on a day-to-Day basis, the person who has to love Docker is the developer.
[00:00:11] Dan Blumberg: That's Justin Cormack, the CTO of Docker.
[00:00:14] Over the past decade, Docker exploded in popularity as companies moved to cloud and adopted Docker's software containers as they did... but the company struggled as as a business and, five years ago, made massive changes:
[00:00:28] Justin Cormack: When we restructured people were like, well, uh, I kind of hope this works, but I doubt it will.
[00:00:34] Dan Blumberg: Well, five years in and Docker's developer obsession is paying off... the company is bigger than it ever was...
[00:00:42] On this episode of CRAFTED... We'll discuss developer productivity, how Docker is building products to improve it, and why so many organizations are in an awkward phase, with too many responsibilities being put onto the developer:
[00:00:56] Justin Cormack: we've had this shift left movement of putting more work onto developers they were being asked to do more with less support Plus, we'll look outside the container for new vectors of growth....
[00:01:08] Almost all of our customers are implementing some form of ai and often, for the first time, And Justin will share tips for developers on how they can better communicate their needs. And what CFO-types can do in return:
[00:01:23] the one thing they can do is actually like, listen to the developers.
[00:01:27] Speaker: Welcome to CRAFTED., a show about great products and the people who make them. I'm Dan Blumberg. I'm a product and growth leader, and on CRAFTED. I'm here to bring you stories of founders, makers, and innovators that reveal how they build game changing products and how you can too. Crafted is brought to you in partnership with Docker, which helps developers build, share, run, and verify applications anywhere without environment confirmation or management.
[00:01:51] More than 20 million developers worldwide use Docker's suite of development tools, services, and automations to accelerate the delivery of secure applications. Learn more at Docker.com.
And Crafted is produced by Modern Product Minds where my team and I can help you take a new product from zero to one and beyond.
[00:02:10] We specialize in early stage product discovery, growth, and experimentation. Learn more and sign up for the Crafted newsletter at modernproductminds.com
[00:02:19] Dan Blumberg: Justin, I understand that you know, not only are you CTO of Docker, you also love to ferment things. Would you say that great software needs to ferment?
[00:02:28] Justin Cormack: software does take, time and experimentation, but actually fermentation's kind of interesting 'cause actually a lot of it is actually about, measurement and control and it's actually fairly repeatable. temperature and, and measurement and things are actually kind of quite important.
[00:02:44] I think the most important thing you can have if you're a beginner in cheese, for example, is a,pH meter to measure the acidity it's kind of useful to know what's going on and whether you,
[00:02:52] Dan Blumberg: Mm-Hmm.
[00:02:52] Justin Cormack: whether you've got it right, measurement and metrics are also important, but in a, in a kind of different sort of way.
[00:02:59] 'cause software's. Often less repeatable, but the measurement parts are for different reasons. Like how it, you know, how's this performing? Am I getting there? is this helping the user? Is this, what's going wrong with the software? So it's, it's, it is similar, but
[00:03:13]
[00:03:13] Dan Blumberg: Gotcha. Gotcha. Um, I promise I'll move off this analogy in a second, but I am curious how, how the new docker, uh, is fermenting. It's been five years since Docker, recapitalized, and refocused, and I'm just, I'd love to understand how, how are all the ingredients coming together?
[00:03:27] Justin Cormack: Yeah, I mean, it's, it's, been
[00:03:29] a, an amazing success. I think that when we restructured people were like, well, uh, I kind of hope this works, but I, I, I doubt it will. And, um, you know, it's actually been an incredible success. We're, larger and stronger than we were before.
[00:03:43] when Docker started, it was an open source project that just exploded and led to this huge, huge amount of innovation across the whole, um, the whole space. but it took a few years before Docker, trying to build a business out of that.
[00:04:01] And it was very early back then, and, we started building a product around deployment and operations of, uh, Docker of Docker Swarm and later Kubernetes. that was a really crowded market and at the time it was quite small, relatively, it's huge now. it was also kind of interesting because it was a, um. Our, our core original customers were developers, and this was an operations product, so the people who had made Docker successful were not really the buyers for this product. Um, so we,pivoted back to those original people who had made Docker successful, who were so excited by it, which was the developer community.
[00:04:41] Um, and we. Built a business around them. that was the real change, was really going back to our roots, going back to our core audience and, and, and building a building products for them.
[00:04:54] Dan Blumberg: Docker Liss is one of its values. Developer obsession. What, what does it mean to be developer obsessed?
[00:05:00] Justin Cormack: Yeah, so we, we spent a long time thinking about, you know, our values and, and why we were doing things and what we were doing, particularly as we restructured the company and, A lot of companies have a customer obsession value. Uh, Amazon is a famous one. but we really wanted to make it clear that,
[00:05:20] the person we think about on a day-to-Day basis, the person who has to love Docker is the developer. And there's a bit of nuance between the, you know, the developer and the customer. 'cause the, when you think of the customer, you think of the buyer, versus the end user.
[00:05:34] And we really wanted to focus on the fact that, you know, it's the developer who. It loves Docker. It's the developer who is gonna be the, using Docker every day, and,cares about performance, cares about how it works, cares about usability and all those things. and they're the people who tell their manager and they tell the organization that they want Docker and that they should.
[00:05:55] buy Docker. And also, we're a software company, we're full of developers and, part of it is just channeling your, you know, your, the customer. Zero focus on like, do, am I shipping something that I want to use as a developer? as a developer?
[00:06:08] Am I proud of this? Am I thinking about the, the quality aspects that developers find important?
[00:06:14] music button here?
[00:06:15] Dan Blumberg: Docker's growth really tracked the, the move that companies made to the cloud. Um, can we just step back for a second? what is the, the value of containers and why was it that that was such a critical piece of the move to the cloud?
[00:06:29] Justin Cormack: Docker grew up really with the, with the microservices movement.
[00:06:33] And if you think about microservices, you know, microservices were about, organizations suddenly got a lot more developers, like development became more important because everything was turning into software. And so if you've got more developers, you have to organize them, they've got more teams.
[00:06:47] and microservices was around like, how can we split responsibility between teams to work, effectively without having to understand a gigantic context of a, you know, huge applications. And so containers will let you, deploy Pieces of your application as microservices, separate out the differing responsibilities, um, separate out things like, databases from code and, and, and so on.
[00:07:12] And let people work with smaller domains, separate out things like APIs and do API driven development and so on. So there was a lot of breaking down of the problem into pieces, combined with a lot around, the kind of the cloud native way of working around repeatability, automation.
[00:07:30] being able to understand what was going on.
[00:07:33] probably the biggest workflow change with the containers is you build something, you test it. and then you deploy it, and if you want to change it, you start again At the beginning of that pipeline, you build a new version, you test it again, you deploy it again.
[00:07:46] And so there's a, there's a straightforward pipeline and you, you can record metadata on exactly what went into production. you know, what went, when you can revert things easily
[00:07:56] and so everything about what's running in production is going through the same version, controlled container, controlled pipeline. And that gives you that repeatability, it gives you the understanding of, um, when there's something goes wrong, you know what's in production. You don't have to go back and spend your time triaging, well, what, what did change?
[00:08:16] why is this broken? what changes actually are in production versus what we had in, production last week, you know, when it was working fine. And so those, those kinds of pieces around repeatability and, teams working together, separation of concerns, and around, um, understandability and which can also help drive the whole observability movement of trying to collect better data so you can understand better what's going on.
[00:08:41] Dan Blumberg: So a lot of what you just described, you know, solves a lot of true pain points for, for engineering teams, for product teams. A lot of those products are not, are not new. I'm curious how in the last five years you've taken those products and obviously built others that ladder up to a much better business for Docker.
[00:08:55] Yeah, absolutely. when we started looking at the restructuring, we looked at what we had, we looked at what had been growing organically over the last few years. We looked at where we spent a lot of time, you know, talking to developers and understanding where their pain points were on the development side.
[00:09:10] Justin Cormack: a lot of what we were just talking about were, it was, and where a lot of the innovation happened, had happened in, in the kind of deployment and production side.
[00:09:19] we decided to focus our piece on the, on the developer side and kind of further downstream from that. And that was for a number of reasons. One is that there've there've been way less investment in that area. Like developers have been kind of left to work out how to deal with this, this their staff.
[00:09:36] And there was actually a big complexity gap between the developer experience and the operator experience that, got messier over in a way, over some the, the time. Kubernetes is a complicated technology and the, the set of technologies around it are complex.
[00:09:50] And so the kind of the operation side of businesses had decided that they would keep that stuff mostly away from developers. and so developers deploying containerized applications, had a kind of different view of the world from what had been going on. And, um, fewer, fewer tools and support for them.
[00:10:07] But at the same time. we've had this shift left movement of putting more work onto developers and they were being asked to do more with less support and there was less investment in, in that developer ecosystem. So we decided that would be a productive area to focus.
[00:10:23] ?
[00:10:23] Justin Cormack: We had Docker hub, which is, the biggest container, image repository. It's where a lot, you know, a lot ofthe base images that people are building their applications come from. we can help, help them use that more effectively. We had Docker desktops, so we had a presence on millions of developers, you know, desktop machines.
[00:10:39] So we, we had thesepieceswhen we restructured and we, we started, working out how to put them together and build them and turn 'em into a business and build on them. So,we started, charging for Docker subscription, which included Docker desktop and Docker Harb.
[00:10:54] and although like it's always difficult making something that was free. Paid. but it was actually very successful. Like people, in general understood the value they were getting from it. And it's, become a good business. But we are really building on this now.
[00:11:11] Docker continues to invest in the speed and reliability in Docker desktop. We've, we've invested a huge amount in getting it better, but we know that that alone isn't sufficient for great developer experience combined with the control that enterprises want So we're building this multi-product strategy to really focus in on this.
[00:11:29] We've launched scouts and docker, build cloud as part of that, both speeding up the inner loop. Docker build cloud saves you time by speeding up builds, and then you iterate faster. And Scout provides insights. Early on into security vulnerabilities that can be avoided before you go out to Alza loop or to production.
[00:11:49] So combining these with the verified images on Docker hub and the trusted content, this enables, you know, far more than just portability and consistency. The Docker pioneered it, reduces the time for dev teams, speeds up, build time, frees up devs to iterate faster, focus on the inner loop and makes building secure code easier.
[00:12:07] While giving you support with policies and remediation guidance from your DevSecOps teams
[00:12:12] Dan Blumberg: you mentioned Shift left a second ago, and I just wanna make sure like everyone understands that's, I think that's a key concept of what developers are being asked more and more to do. Can you explain what Shift left means?
[00:12:22] Justin Cormack: Yeah, so shift left is really, the initial observation was just that, the later you fix something, the more expensive it is to fix. if you have a bug and you, uh, it gets into production and it causes issues for customers and it's, months after you wrote the code, then that's expensive to fix.
[00:12:43] If you have a, the bug and you find it while you are on your laptop before you've even pushed it to get, and you go, oh yes, I shouldn't do that. Let me fix it.
[00:12:52] so the sooner you can do things, the quicker it is, and the cheaper it is and the more efficient it is. And that's also definitely very much true of security, you know, security issues much worse when they're in production than if you fix them early on. and that's why we talk about the inner loop as a loop as well.
[00:13:09] The inner loop is that time when you are in flow of writing code and getting things, you know, fixing things and you want things to be fast. And that's, that's the point that if you can fix things then you know, that's the furthest left. The outer loop is the kind of ci push if you find issues there. Um, for many people the kind of ci push is still, too slow.
[00:13:35] And that's, again, one of the things with doer bil cloud that we're trying to speed up isthrough the caching and so on is the outer loop. But often for people, you know, you talk to developers and they'll be in a cycle that they don't run most of the tests locally because of configuration and other reasons.
[00:13:50] They push to see. You don't get feedback for a couple of hours by time, which time you've kind of lost the context and it comes back. But that's still, even, that is still quicker than it's running in production and a customer complains about it a week later and says, why did you ship this terrible software?
[00:14:08] You should never have shipped that. It's clearly, clearly broken. And then you have to go and fix that, and that's even more expensive. But it's also, you know, shift left is, has is also been a kind of excuse to put more burden on developers.
[00:14:21] and developers have to be supported in these activities. Um, and they have to have things, prioritized. They have to try and understand what the things they need to actually fix are because.
[00:14:31] if they were obvious, they would've fixed them anyway. So
[00:14:34] they oversee, you know, they kind of need tooling to, to help them understand that
[00:14:38] Dan Blumberg: with, with that in mind and with the additional responsibilities that developers are, are being given, how do you prioritize Docker's roadmap?
[00:14:46] Justin Cormack: the number one thing is listening to customers. we have a lot of customers. and particularly we're a very diverse set of customers. We have, you know, from the largest enterprises to some, individual developers. So
[00:14:58] listening to them is really important because there's so many of them, and they're so, different in their use cases. And, dock is such a broad tool there's so many different ways. people use it in different, um, difficulties they run into.
[00:15:11] So that's a, a lot of the work is really just spending time with customers, understanding what they're finding difficult, where they are organizationally, and how their organization is thinking about, how this fits in. talking to the teams that look after developers, I mean, platform teams have been really. Becoming a, a real thing for many, most organizations we talked to for in the last few years. And those are really, you know, those teams are really thinking about how they help developers in a very focused way now. So spending time talking to them and understanding their problems helps us understand developer problems.
[00:15:50] Dan Blumberg: you mentioned there's so many. Different types of customers. There are 20 million developers that use Docker in one way or another. And, and if the number I found on the web is to be believed, there's 27 million developers worldwide. So you have incredible market penetration among developers.
[00:16:04] Um, it also means, I'm, I'm guessing that you're thinking a lot about ways you can deepen existing relationships with developers. can you share more on how you approach, that challenge?
[00:16:14] Justin Cormack: there's a lot going on for developers in, in many ways it's a, you know, as we said, more is being put on them because more, more things are becoming software. and so it's a, it's an, it is a really interesting time from that point of view. you know, the saying is that the future's unevenly distributed is, it's very true that,
[00:16:34] there are organizations that arefully containerized, but you still go out and talk to people who are starting on that journey and, there's also this very wide spectrum of, of users at different parts of the journey.
[00:16:46] You know, I've been a dogger for nine years now, and one of the things that's been fascinating to watch is our, audience grows, the kind of questions they ask and the kind of, um, I issues they have, have changed so much because the early adopters were, quite sophisticated users
[00:17:02] Now it's, everyone is using darker and they're using it for a very broad set of use cases.
[00:17:08] for a long time it was a, most of our users were on Linux and then, Mac and Linux. And now most of our users are on Windows. And the Windows users, uh, come from a very different backgrounds and they have different view of the world. They, they tend to be more familiar with Gooeys than command lines,but again, there's a big spectrum within them as well.
[00:17:29] And then, um, you know, I think that obviously things like AI in the last few years have really affected the developer as well.
[00:17:37] Dan Blumberg: Yeah, I, I wanted to ask, what, what are the conversations that are most sort of timely and relevant with Docker's customers? I imagine AI is very much one of those things
[00:17:45] Justin Cormack: almost all of our customers are implementing some form of ai and often, um, for the first time, um, so.
[00:17:54] Last year we released the, um, docker gen AI stack, which is a, you know, kind of example rag stack. So RAG is one of the most common patterns people use for building AI applications. And we've talked to customers who are planning to build hundreds of rag applications. So RAG is retrieval. Augmented generation is basically where you take, some set of data that you have, often, for example, documents or, but it might be images and or video and things as well.
[00:18:22] and you want to use a generic LLM model to actually, you know, answer questions, but it doesn't know anything about your
[00:18:29] specific data. So you a rag app, you build a, um, retrieval engine in front that adds the context in so that it can answer with relevant information.
[00:18:38] And this is, this has become a very successful pattern. It's very widely used. it's not the simplest thing to actually build and deploy in production. And we're talking to a lot of customers about how to make that production journey better.
[00:18:51] there's lots of low level pieces of the stack that just AI makes more important, like working with GPUs
[00:18:58] but we're generally interested in like, how. AI tooling is going to affect all of development and, and the kinds of ways we can help developers in a more broad way.
[00:19:08] We have a, um, doer copilot plugin, that helps you with Docker related, questions but we're also interested in like, there's a lot of tooling that developers use, um, and things they do on a day-to-Day basis where AI could help them that are not just specifically about coding.
[00:19:23] It's like, one of the questions we've been looking at is how do we help developers use particular tools? Like, um, Dogger hub is, has, has thousands of tools or millions of tools in containers, but it's not always obvious. How do I use, if I have a problem, how do I solve it?
[00:19:38] You spend a lot of time working at how to do ad hoc things, and AI is actually quite good at helping you do these things.
[00:19:46] People are realizing that developers are, expensive and if they, if you can give them some more productivity, it's
[00:19:52] actually really valuable. And they spend a lot, they waste a lot of time. doing things that are not productive and so why not invest the same as you do in, in other parts of your organization on productivity tools?
[00:20:07] I think that this is starting to change and people are understanding that,these tools are actually, are really valuable for developers. And
[00:20:12] the amount of investment that's going in is, is increasing the amount of value that you're getting out of them every year.
[00:20:18] And, your developers are asking for this, not because they, uh, are trying to waste your money. They're actually asking for this 'cause it really helps them do their job.
[00:20:26] Dan Blumberg: what is the cutting edge of, of docker r and d that is gonna drive the next, you know, big productivity gain the next five years?
[00:20:35] Justin Cormack: Well, we talked about the AI pieces and I think, AI is really interesting just from the point of view that, developers have adopted it really fast compared to a lot of other areas. And it's actually, you know, I think there's a lot of things that are really, um,good about the ecosystem that we built in, particularly in sort of cloud native development where it actually makes it easier to deploy ai.
[00:20:57] 'cause like, for example, like we have tests so we can see if the ai, ai generated code is terrible, like more easily. And, um, we have this structure of, how you build stuff that you know is reliable and you have checks and balances before things go to production. although, the infrastructure we have in order to deploy AI assisted tooling is better than many other industries because we are, we, we have that formalization and we have that,
[00:21:27] um, understanding.
[00:21:29] So I think, there's a lot we can do with ai.
[00:21:31] there's still a lot of complexity but you can encapsulate that and, and modularize it in a more effective way so that you don't have to understand it all at once.
[00:21:42] Or it can be divided up, you know, amongst teams. So, developers understand the context in which their application runs. The ops people understand the context in which production cluster is configured. Finding a way that they,can work together more effectively, um, so that they don't have to both understand each other's context, I think is helpful.
[00:22:05] And I think we're still in this kind of weird state where we, we haven't quite, you know, helped developers build with the right. Low complexity stuff and, and build applications that are really simple to, kind of understand and build.
[00:22:22] we've left them all the options open because containers can run anything, but we've may maybe given them too many things to be concerned about. And I think that we have to, we have to do a bit more about giving, we get asked a lot about, guidance and recommendations, and I think we need to actually, Simplify some of the stuff that's come out.
[00:22:42] And so trying to understand which, what the happy path, the golden paths, the right ways to build maintainable understandable applications are, and
[00:22:53] I think is really important right
[00:22:54] Dan Blumberg: Yeah. You, you haven't used this word yet. I don't wanna put words in your mouth, but it sounds like what you're saying is you need to be a bit more opinionated.
[00:23:01] Justin Cormack: I, I think some more opinions. Yeah. I think
[00:23:05] this
[00:23:05] Dan Blumberg: that's a tricky thing with open source, you know, but I, I, I'm curious
[00:23:08] how you, how you approach this sort of platform versus, you just said like the Golden Pass or the, you know, the, there are certain better
[00:23:14] Justin Cormack: well, 'cause I, I mean, yeah, the Golden Path term is used a lot by, the platform teams who are trying to, make their developers more productive, I think there's a few different focuses for platform teams.
[00:23:24] Some, some platform teams are in the business of like deciding what the platform is. But a lot of the, the really strong platform teams are in the business of looking to see what developers are are doing and trying to guide them into a smaller number of paths that can be
[00:23:40] managed. Better and give them higher productivity and more commonality in the organization.
[00:23:46] you know, this is how you containerize things. This is how you build things. This is how you operate them. This is the kinds of things you that are gonna be costly and you don't want to do. These are the things that are gonna be, that are simple and cheap and you do want to do.
[00:23:57] And I think working on those paths and simplifying is really important.
[00:24:03] Dan Blumberg: who's the typical buyer for Docker products? Is it, is it a CTO? Is it individual developers?
[00:24:09] Justin Cormack: it's usually a platform team or the VP of engineering,
[00:24:12] sometimes it's a strategic thing, especially where it's about containerization in the cloud journey as a whole organization, and then then it's more likely the CTO.
[00:24:21] Dan Blumberg: Yeah. Well the, the reason that I asked the question is, is that it oftentimes with big companies that the CFO or the CEO has to sign off eventually. And
[00:24:29] I'm curious if there's one thing that you wish those who you know, are not steeped in DevOps, would better understand, uh, about the complexity you're talking about, about the ways that you know developers' lives can be more efficient about the ROI, you know, thereof.
[00:24:43] Justin Cormack: I think that one, one, the one thing they can do is actually like, listen to the developers. Well, like the developers
[00:24:50] will tell them like, please help my CFO understand why we, why we, our organization needs to use dark air.
[00:24:56] Like the, the developers are, are there. And there's a lot of organizations now who are much better at listening to developers. I think that, you know, modern software and the kind of repeatability,
[00:25:07] the security piece that comes from the way that. Containers are just deployed once, and a trackable, that the audit trail you get from containers is still underestimated in like how important it is in terms of understanding what's going on in the organization and forensics and those types of things.
[00:25:25] so Docker SCS is really backed by, this kind of supply chain database of everything that you've built and deployed and what's in it and what's, uh, not in it, and how it was built and where it came from. you know, that those underpinnings are really powerful in it in terms of, uh, like under understanding what's going on in your organization and really having observability over the software process.
[00:25:48] And I think, that's still something that people, uh. Still discovering that they need,
[00:25:54] for a lot of larger organizations, they're just starting to understand those things.
[00:25:58] 'cause they built, really often very, very complex processes around, things like change control that were pre-con container. and so they're kind of having to, there's, they're having to understand how to adapt those processes to actually, well, something that's way more efficient at scale.
[00:26:14] Um, and actually lets them deploy and, um, build software much faster. but it involves organizational change And I think that, that speed thing is obviously really important to organizations.
[00:26:24] 'cause the old processes were very, very slow. My, my favorite change control story is, um, I worked on a project years ago at, um, Heathrow Airport The change control process for everyone, every, or every organization who worked in the entire airport was such that while Heathrow Terminal Five was being brought up, nothing could change at all. so the prices in the duty free shops could not be changed while the airport was being, uh, brought up.
[00:26:56] That was the kind of change control process that, was historically was seen as. Okay. And normal, and necessary to control what happens. Whereas, with a containerized system, you've got a lot of metadata. You can understand the changes. You can, trace them back to source code.
[00:27:13] You can see what the changes were. You can manage change in a, in a much more granular fashion. You don't have to freeze everything for, for years at a time.
[00:27:23] Dan Blumberg: Yeah, you're reminding me there was a, a period when I was at the New York Times where the entire iOS app, which had been built like in secret, like in Cupertino, just had to be rebuilt. no changes could be allowed for like nine months or, you know, advertising might beg, please put this one change in even though we're building the new thing.
[00:27:38] but I love, I love that actual very physical example of the duty free shop prices, like couldn't change because it would all the, all the downstream effects.
[00:27:46] Justin Cormack: MUSIC BUTTON HERE
[00:27:47] Dan Blumberg: You mentioned thatthe laptops that, that people who are using Docker are using has changed dramatically over, over the years Is that more a function of Dockers partnerships with Microsoft and others, or is that actually just a developer trend that, that Windows machines are taking off again?
[00:28:02] Justin Cormack: it's a mixture of different things. I mean, I think that, windows machines are predominantly a function of, the kind of comp, you know, larger organizations tend to be very Windows
[00:28:13] Dan Blumberg: banks
[00:28:14] Justin Cormack: banks for ex
[00:28:14] Yeah.
[00:28:15] Most, although it's, it's less strong than it used to be. There's like, there's more of a mixture
[00:28:20] and, and, there's more diversity in those organizations too.
[00:28:23] Um, I think the, the Mac has been an interesting, I mean, I. Developers love their laptops. Like they're real power users and they like fast, high performance, long battery life, uh, nice screens because they stare at them all day. and they care about things like type typeface rendering in a way that, um, perhaps someone who's just watching videos might not care so
[00:28:46] much. And
[00:28:47] you, you know, if you look and keyboard Oh yeah, keyboard, yes. Keyboards. And so they're very demanding and like, there was a lot of complaints about the late Intel max. Um, and people were moving away from Mac completely. And thenapple Silicon Max came out and they're like, oh, okay, these are really good.
[00:29:04] developers are very, very obsessive about their machines. And I mean, there's all this question about like, a developer's just gonna move everything, all their whole development environments in the cloud. But they've been quite reluctant to do that.
[00:29:18] And there's a bunch of reasons around it. But I think part of it is just like the laptops have got better.
[00:29:23] Dan Blumberg: Absolutely. And I'm interested how, with the move to the cloud and how you've. Dealt with things like, just, just the way things feel or like latency and things like that. I, you know, that which, which is I'm guessing is one of the hesitancies of people to, to move from desktop to cloud.
[00:29:38] Justin Cormack: Yeah, I mean, our thesis at the moment is people don't want to actually move, for example, the editing environment to the cloud because that's the most latency sensitive piece. Um, but we've done a lot of work in making things like this startup time for remote build be, as fast as possible.
[00:29:55] and often the performance can be higher because you've got more, you've got more bandwidth in the cloud to pull things down, and you've got potentially more resources. But you know, we've done a lot of performance work, both on Docker desktop locally and, uh, on the cloud interfaces because it's, you know, like it is really important.
[00:30:12] Like, if it feels slow, people won't use it. But yeah, we, that's why we,don't think moving the editors to the cloud is where developers want to go right now. 'cause that's the bit that the latency matters
[00:30:24] a
[00:30:25] lot. Um, a lot. of the other things you can, with careful caching and careful pre-work, doing things earlier, you can, you can kind of work around it so it becomes less visible.
[00:30:35] Dan Blumberg: Yeah, I, one of my favorite examples of this in the very early days was Spotify, Daniel Eck, wanted, if you tapped on a piece of music for it to stream, you know, from the cloud instantly, and I think, think, I think it was something like, it had to be less than 300 milliseconds because his hypothesis with it, if it didn't, didn't feel like it was on your phone, you weren't gonna engage.
[00:30:53]
[00:30:53] Justin Cormack: Yeah. And, and developers are very sensitive to it,
[00:30:56] So, you know, for us, performance engineering is really important and we've, we've done a huge amount of performance engineering work on Docker desktop.
[00:31:02] 'cause it wasn't, it wasn't good enough. And over the last year, it's become enormously faster. We, the startup time is now pretty much instantaneous. Again, like it was when we first launched it, it kind of slowed down and,the VM performance is like enormously faster than it used to be.
[00:31:17] Like, these are things that, we focus on because we care, our users care, you know, and,it wasn't really sufficient
[00:31:25] Dan Blumberg: Yeah. I, I want, I wanna ask you one last question I asked you, uh, about what's one thing CFOs and, and others, should know, and you said they should just talk to their developers and listen to their developers. I'm curious if there, I'm gonna flip the question. Is there a way that developers, engineering managers, others can speak in ways that those, you know, holding the purse strings will understand better?
[00:31:48] Do you have any one tip on like, communication for, for developers or development managers?
[00:31:54] Justin Cormack: Yeah, I, I think there's been a very, um, there's been these attempts to like, have these conversations about developer productivity and how to measure it and do our own space metrics, and like trying to use those as a way of kind of communicating in the organization. I, I don't know that they've been that great at having, you know, making that dialogue.
[00:32:19] Make sense, to the other half because I think they often seem to land sort of in between like the, 'cause the developers on their own don't actually care about those metrics per se, because those metrics are metrics that are correlated with the developer having, being productive and being, having good experience, but they're not the way that the developer feels that, and those metrics also don't necessarily resonate with the, CFOs because they seem a bit like, why do I care how many times a day these people ship Like that doesn't seem like so, so I think that, and there was a lot of complaints about the McKinsey study.
[00:32:59] I.
[00:32:59] Um,
[00:33:00] Dan Blumberg: saw
[00:33:00] Justin Cormack: kind of, so I think the conversation is not landing very well right now. I still think that we haven't found the right way of having those conversations.
[00:33:11] the best way for developers to communicate is to get better at communicating how they're working with customers and how they're focusing on business outcomes. Um, because that's, that's the language of the business and I think
[00:33:26] that, product focused and, business focused engineers are, becoming more common.
[00:33:32] like, what are, what are we bringing to the business? how are we helping the customers? what's getting better from the customer's point of view? I think that's, that's, that's a really, that's the kind of the way that communication needs to work.
[00:33:44] Dan Blumberg: Justin, thank you so much.That's Justin Cormack, Docker's CTO.
[00:33:48] Speaker: I'm Dan Blumberg and this is CRAFTED. CRAFTED. is brought to you in partnership with Docker, which helps developers build, share, run, and verify applications anywhere without environment confirmation or management. More than 20 million developers worldwide use Docker suite of development, tools, services, and automations to accelerate the delivery of secure applications. Learn more at Docker.com.
Special thanks to Artium where I launched CRAFTED. Artium is a next generation software development consultancy that combines elite human craftsmanship and artificial intelligence. See how Artium can help you build your future at Artium.ai
And CRAFTED. is produced by Modern Product Minds, where my team and I can help you take a new product from zero to one and beyond. We specialize in early stage product discovery, growth, and experimentation. Learn more and sign up for the CRAFTED. Newsletter at modernproductminds.com.
[00:34:44] OK, until next week... in the meantime, please share CRAFTED with a friend...
[00:34:49] Justin Cormack: I kind of hope this works, but I doubt it will.