Every night. That's the honest answer to how long this has taken. Not a few focused weekends, but late night after late night, teaching myself a craft I had no background in. If you're deep in your own version of that, anxious and a little behind, you're not the only one. I've felt it most of this stretch too.

And the backdrop matters, so I won't pretend otherwise. Redundancies are sweeping the industry, and friends and former coworkers have been caught in it. I've been fortunate: I'm still employed and I genuinely enjoy my role. But I've watched shifts like this before, radio giving way to Spotify, television advertising to YouTube and Facebook, and the pattern holds: these waves arrive whether or not you're ready. So I made a small, stubborn choice. Rather than wait to see where I'd land, I'd keep upskilling and learn to move with the change instead of bracing against it.

I've edited for a living for more than two decades, across a lot of industries. The surprise, if you've never done it, is that the hard part isn't the editing. It's the feedback. Waiting on notes is where the time quietly goes, and it often comes back unclear, or later than you'd hoped. Plenty of tools promise to smooth the review process. None have solved the real thing: notes from people who know what they want but can't quite put it into words.

That gap was the seed. I set out to fix it, the idea ballooned, and while wrestling with it I spotted a smaller, sharper problem sitting right beside it. So I parked the big one and chased the small one.

The small idea was a planner and reminder app. How hard could "remind me at the right moment" be? Almost immediately I learned that a reminder is a lie we tell ourselves about simplicity. Under that friendly toggle sits logic I never knew existed: time zones that drift with daylight saving, notification permissions that differ on every system and can vanish without warning, reminders that depend on other reminders. One tidy feature was really a stack of small, unforgiving decisions, each hidden until the moment I got it wrong.

What looked like a single feature turned out to be a web of dependencies, and every one of them had to be right, or the whole thing quietly broke.

The deeper I got, the more the app became about that bigger problem. I was drowning in notifications, a dozen apps each certain it deserved my attention right now, and I wanted one calm place that could see across all of them. That's what Focus Drift turned into. It's a planner and reminder app that talks to my other apps through MCP and pulls what matters into one view, mine or an assistant's.

What mattered to me was keeping it private, and making sure it didn't just become one more thing buzzing for attention, adding to the notification hellscape it was meant to quiet. It's out now, and in the end it turned out less about ticking things off than showing me where my day actually went.

I'm not an engineer, and I won't pretend to be. Getting this to run meant far more than subscribing to a service and flipping it on. It meant endless questions, and real engineers patiently explaining things while I nodded with a blank face, not getting it, until days later it clicked. I got things working first and understood them second. That's the honest order.

Then the pile nobody warns you about. Version control, so one bad save doesn't cost a week. Security, a rabbit hole with no bottom once real people are involved. Promotion, the plain job of getting anyone to notice, which has only just begun and may be the hardest of all. Each is its own trade. There's a reason tech has specialised roles, the same reason television and radio did. But industries shift, and sometimes you just have to sink or swim. I chose to swim, clumsily.

Nothing humbled me like the App Store. You hand your work to a reviewer you'll never speak to and wait on a process you can't see. Rejections arrive over rules you didn't know existed. You fix, resubmit, wait again. It teaches a patience you never signed up for.

Shipping across platforms taught me something subtler. That systems differ is obvious. The real skill is reusing your own work well, building a component once and carrying it across, while accepting it won't always survive the trip and some things need redesigning for where they'll live. That word, design, is where my old job crept back in. Designing a screen isn't far from cutting a sequence: both are visual, about what the eye reaches for first. New tools, same instinct.

Once I'd shipped one thing, the ideas multiplied. A retro office-chaos arcade game headed for Steam. A set of illustrated flashcards from the vibe-coder work, part portfolio and part experiment, that happen to save a fair few tokens while you're learning to build with AI. And the energy carried me back to the idea I'd parked at the start. Evergreen is the big one, and it's really about collaboration, helping people who aren't technical work with video and shape their own ideas instead of handing them off and hoping someone else gets it. Underneath, it's mostly about finding what you need fast, whether you dig for it yourself or just ask, and it works on its own without any AI at all. When you do want the AI, it stays on your own device. My hope, and it's only a hope, is that it does for people shut out of editing something like what Canva did for people shut out of design. I'm close to shipping it and testing with a few people I trust.

I assumed none of my craft would carry into software. The opposite was true. Attention to detail, the kind where one wrong frame nags until it's fixed. Thinking far enough ahead that today's choice doesn't box you in. Heading off risks before they break at launch. Troubleshooting when something falls over and you don't yet know why. All of it transfers. Though there's one part I haven't cracked: working with teams, the communication and clarity of shared goals could use an Evergreen of their own. Or maybe that's also part of what's coming. Stay tuned.

Here's what I'd tell the version of me from a few months ago, the one who didn't know what an alarm permission or an App Store reviewer was. The confusion isn't failure. It is the work. You don't understand it first and build second. You build, badly and at midnight, and the understanding arrives on its way out.

The wave is coming either way, from radio to Spotify, from television to the feed, and now through all of us. I can't slow it down. But the fear loosens the moment you stop bracing and start paddling. What's next for me is offline orchestration for generating images and video, to get ahead of the rising cost of tokens. A few months ago I couldn't have parsed that sentence. Tonight I'm building it.

If you're in your own stretch of late nights, keep going. The person who knows how to do this is the same one who's confused about it right now. That was me. It can be you.

You can follow along with everything I'm building at evergreencontent.app.