Inspired with Alexa von Tobel
00:15:04 · Alexa von Tobel, Dan Lorenc
[00:15:04] S1: That's amazing. Dan, over the last couple of years, you guys have expanded beyond containers into libraries, OS packages, and more recently into workflows and agent related products just announced this year in 2026. How have you thought about your own product expansion? And then let's get... Let's touch on agentic workflows and Yeah. What that's gonna look like. [00:15:25] S0: We're we're fourteen minutes in, and we're finally talking about AI. Yeah. That's kinda correct. [00:15:29] S1: Hey. By the way, you're welcome. Yeah. Yeah.
00:15:33 · Dan Lorenc
[00:15:33] S0: I mean, it's probably the most exciting and scariest time to be in software engineering in, you know, the history. It's... You know, software engineers mostly been the same for the last thirty or forty years. Different programming languages, different tools. But at the end of the day, people, you know, still been writing this stuff by hand. The tools have gotten a little bit better, but, you know, there's been no massive step function increase in productivity since the first programming languages came out. And now all of sudden, that is getting flipped on its head, and the entire way we build software and products and ship features and road maps and where is your mode and differentiation and all of that is just kind of thrown up in the air, and everyone's trying to figure it out all at the same time. But, yeah, we've been able to leverage a lot of this stuff over the past couple... The past, say, eighteen months, really, to accelerate how we're shipping code, letting us do more things safely and securely. So, yeah, the business we started with container images, and we've, you know, we spent a whole year trying to use AI automated, and nothing worked. And then all of sudden, like, overnight, the models and the harnesses and everything got better. And we've been able to, you know, use AI to ship... To add more things to the catalog even faster and maintain that software in the catalog even faster. So, yeah, we've expanded. We're trying to to secure all open source out there and containers where we started are just one small portion of that. It's one of these things where the... We would not have been able to build these products or this business the way we were trying to if these tools hadn't gotten there. And, thankfully, they kind of did it the exact right time for us.
00:16:55 · Alexa von Tobel, Dan Lorenc
[00:16:55] S1: I wanna quickly double click on what you just said, which is like we truly are in a step change function where code was written by a human for the last thirty to forty plus years. Every improvement was sort of a micro improvement, and now we're actually seeing a true step function change of, I would say, maybe even two to three happening. What do you think this looks like in three to five years? What can you see? It would just play it out for us, as somebody with the front row seat. [00:17:34] S0: Yeah. Three to five years. Wow. That... That's a long time frame, especially with the the curve and rate of growth going this fast. But I think the next year or two, I think we're we're gonna completely shift the way software is written. The analogy I like to use to explain, you know, the the old way, what's happening now, and then where we're going is woodworking. I've used this this story a couple times, but, you know, we've been building software by hand. It's like you're doing hand woodworking. You had saws. You had manual drills, all of that kind of thing. And if you've ever done that, it's fun. A lot of people still do hand woodworking. Well [00:18:05] S1: I have. It's great. [00:18:07] S0: It's slow. You get to take your time. It's harder to make mistakes. [00:18:10] S1: Wanna Fairy. It's it's it's... We got... It's very therapeutic.
00:18:13 · Dan Lorenc
[00:18:13] S0: Yeah. But it also takes forever. Yeah. And, you know, the the kind of agentic tools that came out over the past year, like the cursors and the quad codes and all of these kind of things, it's like we gave everyone power tools all at the same time. Like, now every engineer has a circular saw instead of a hacksaw. It's lot of fun. You're moving a lot faster. You're gonna cut a finger off if you haven't been trained on how to do these things, safely. It's a lot more dangerous at the same time, but it's a lot more fun, and we're cutting a lot more wood a lot quicker. But that's also not how, you know, more... Like, modern furniture and, all of this stuff actually gets built out of wood. And that's sort of where I think, you know, in the next year, we're gonna end up with. Automated assembly lines and factories and systems that do all of this. People are still gonna have those tools around, and they're gonna do, you know, one offs here and there and use those to build these automated systems. But that's not how software is gonna be created. It's gonna go straight from idea to production, in these automated systems that ensure you have quality, that ensure you have security, that ensure it meets the requirements, that ensure the system isn't gonna go down in any one of these stages. And it... The role of a software engineer is gonna change. You're gonna be operating in that factory rather than, you know, writing the code or being involved and being in the loop on all of this. It's the the first time ever where ideas are finally the bottleneck again. You know? Everyone always used to say ideas are worthless. The only thing that matters is execution. That kind of middle part of building a product, all of that hard execution people had to do in building and shipping, that's getting compressed now. That work is getting a lot cheaper. It's getting a lot faster. And you can build anything you want in a weekend now, and this is still the early days for all of this software.
00:19:48 · Alexa von Tobel, Dan Lorenc
[00:19:48] S1: I... By the way, it is the early days, and I, had fun and literally built an entire platform last weekend just in a few hours having some fun banging around. And to your point, ideas will become the bottleneck because the build is not the bottleneck. What is the role of a software engineer in the future in five years? [00:20:09] S0: Like That kind of factory operator role. Right? You know, the the... One of the problems a lot of organizations have now, and, you know, we talk to CTOs and people all over. Their engineers are using Claude Code and Cursor, and they're spending tons of money on tokens, but the road map is still kinda going at the same speed. We're not kind of turning all of that increased code being written into increased product velocity yet. [00:20:31] S1: Why is it moving at the same speed, Dan? I I I wanted to ask you that.
00:20:34 · Dan Lorenc
[00:20:34] S0: Yeah. Yeah. And I think I think it's, like, there's other bottlenecks here. Right? If you write code in your local ID and press save, that doesn't go right to production. That doesn't go right into users' hands. There's dozens of other steps that happen in between in between that. And it's just like any other assembly line or optimization problem. We've... It completely eliminated one of the bottlenecks, but it just pushes the bottleneck downstream to somewhere else. And that's the kind of thing that organizations have to figure out. You have to get to that spot where, like, when an engineer presses save and when they merge a commit, it goes straight to production, confidently. And until that happens, yet the bottleneck has just kinda moved, you've got so much code you can't ship at all. Turning that that shipping speed and iteration speed and being able to deliver features to your customers quickly is, you know, the next part of that bottleneck. It's like the old days of, like, you know, on prem software. Companies used to... You know, my first job was at Microsoft, and I worked on a... An on prem CRM system. System. And they shipped every three years on DVDs. It didn't matter how much stuff you shipped. It wasn't getting into customers' hands until they burned the next set of DVDs. And then everyone figured out SaaS was a lot better for a bunch of reasons, but one of them is you could add new features to a website every single day. Every time someone hits refresh, you can get value out a lot quicker. That's a similar analogy, I think, to what the industry has to figure out coming up soon.