Indie Diary #9 — What do we do now?

I’ve been looking forward to writing this new diary entry for a while, and it only took me about two years to come back to writing it. Why so long? Plenty of reasons. And they have been piling up one after another, delaying it even more.

But here I am, finally hitting the keyboard writing one of my stories, to tell you more.

It’s been so long!

Was it? Indeed. For about 2 years, I’ve been much less active on social media. And even less on my blog and my projects.
The only place where I’ve stayed somewhat active was conference speaking:

• Security practices for mobile at Mobilis in Mobile 2024,
• Kickstarting backend development with Swift at ServerSide.swift 2024,
• SwiftUI techniques to craft better APIs at Swift Connection 2025, Do.iOS 2025, ARCtic Conference 2026
• CLI Driven Development at SymfonyDay Montréal 2026.

I’m really grateful for the recognition I get from being selected in various calls for papers, and meeting all those great people around the world. I really am.

Speaking is always an incredible experience, in so many ways. And I won’t stop there.

Aside from speaking, I was pretty much avoiding the rest. I didn’t feel like building or sharing stuff. Mostly because I wasn’t feeling the craving of building and being active. That craving had simply been replaced. But also because of plenty of unexpected life events, the kind of surprises that only life has the secret for. Thankfully, nothing too bad or unrecoverable from.

What could replace the craving of crafting?

In August 2024, my partner got pregnant. It was something we had planned for, and from then until his birth in April 2025, it became our main focus. Since then, our new life as parents took over.

Aside from removing almost any free time I had for building things, it also replaced the need to do so.

Learning to be a father is probably one of those adventures of a lifetime, and I’m grateful to experience it as much as I can, working from home, with my partner, and my son.

Have I really stopped?

Not really. There were attempts, a few projects started, but they never reached any kind of maturity.

After releasing the new watchOS app for Padlok I stayed active on my apps a bit, for groundwork and maintenance. Nothing fancy.

On the open-source side, ButtonKit, my most successful library, got a bit of love, including the symbolEffect release I’d been looking forward to for almost a year.

But aside from that, it was a very quiet time outside of my daily job.

The revolution

Then came the fall of 2025, and for me the beginning of the revolution that will shake our entire industry. Call it Claude Code, Codex, or whatever. I really do think we’re not on the edge of something yet to come. We’re already inside a profound transformation of our role as developers.

At first, I wasn’t sure what I was looking at. I started experimenting last summer. Tried to implement a few things. Sometimes I was impressed. Sometimes I was not. But the more I used it, the more disbelief I started to feel about this technology. Not only was it relevant when used wisely, but it felt like everything was about to change.

I started using it at work, and I used it to start crafting things again in my very light free time available. It allowed me to land a few improvements on ButtonKit, yet to be tested and released. Along with a few prototype ideas for apps I might never release, but that started being impressive in terms of technical difficulties that were solved in a blink.

And it gave me room to experiment on server-side Swift, where I’d always wanted to invest more time and energy to grow this community. Claude and Codex built various things on my instructions in record time:

• The server implementation of engine.io and socket.io based on Hummingbird and modern Swift,
• A Google Cloud Platform toolkit to improve Hummingbird or Vapor integration when deploying a swift binary on that cloud provider,
• An abstraction of AI Agent and Conversation for both client and server, allowing to switch providers (OpenAI, Anthropic, Gemini, AppleIntelligence) without having to refactor all the logic built on top of a specific API. This last one was very interesting, but it is already outdated and mostly useless now that Apple promised to open-source FoundationModels.

I did all of those things on a very tight schedule with very little free time to spare on those projects, while keeping a full time job and taking care of a toddler. That’s the part that really shakes me: it showed me what is now possible with so little time available.

Where are we going?

Around me, I find a lot of different behaviors and beliefs regarding what we are currently looking at. Some think that AI will never replace developers. Others that it will replace them all, or almost all of the non-senior ones.

I believe that like it happened for manufacturing and factories over the last century, it will, for sure, mean that it’ll be possible to produce the same, or even more, with far fewer people; and therefore that large companies will lay off a lot of their workforce. I also believe that there will be a period where being hired as a developer will be harder than ever.

But I also do believe that like for manufacturing and factories, an equilibrium will be found at some point. Way more software will be produced, by far more startups and indies than ever before. Just as robots reduced the entry cost of production, AI agents do the same for the tertiary sector.

But it’s only speculation.

What do we do now?

I suppose the answer is simple, and straightforward: we adapt. We, developers, have always been keen on learning new technologies, new patterns, and new things. This is one more.

I’m not sure I understand people who absolutely don’t want to try AI to develop anything. For me, it’s missing out on one of the biggest evolutions of the industry that will, whatever you think of it, change it all. Even if it turns out not to be, it’s a big risk to refuse to embrace it on the assumption that it’s just a trend. If you’re wrong, here goes your career.

We’re living in both exciting and frightening times. Even for my indie projects, I feel that I will now have time again to build new apps, and maintain the old ones without sacrificing my family time. Therefore, the indie scene and the App Store will be more competitive than ever. Building a business might become even more difficult than before.

The final question is the price. The $20/month plan doesn’t let you do that much. But it’s burning way more value in tokens than it’s worth. There will be a day when providers will start asking us to pay the price. And I wonder what will happen if no one is willing to pay the true price of AI.

Maybe the future is a future of local/self-hosted AI, on powerful machines, with open-weight models that will be just as good as the latest ones from Anthropic or OpenAI. After all, they are just about six months behind, as long as we’re able to produce enough GPUs and RAM for everyone, making it possible to afford one without selling your car and a kidney.

I also wonder if I’ll take as much pleasure crafting with AI as I did in the past, crafting by myself. I have a lot of projects on the shelf. I guess only time will tell.