I've been having trouble falling asleep lately. Not the anxious kind — the opposite. I'll be lying there, ready to call it a night, and my brain just won't stop. Features I want to build. Products I want to prototype. Problems I just realized I know how to solve. I close my eyes and within thirty seconds I'm mentally architecting something.
This keeps happening. Not every night, but enough that I've noticed. My mind starts spinning with possibilities and it doesn't want to power down. A new feature for OnPlane. A new tool I could prototype this weekend. A whole new product I hadn't considered until ten minutes ago.
This isn't a productivity hack post. It's me trying to figure out why my brain has started treating bedtime like a whiteboarding session. I think I have, and it has everything to do with how dramatically the relationship between ideas and execution has changed.
Ideas Used to Be Cheap
For most of my career, ideas were free because execution was expensive. You'd have a great idea in the shower, get excited about it for a day, and then reality would set in — it would take months to build, you'd need a team, the cost would eat your runway. So you'd file it away in the "someday" folder in your brain and move on. The friction was a natural filter. Most ideas died not because they were bad, but because the cost of finding out if they were good was too high.
That filter is gone.
I had a thought the other night — lying in bed, obviously — about building an AI day trader. Not a high-frequency bot with real money. A sandbox: let Claude manage a paper trading portfolio, make decisions based on market data, learn from its wins and losses, all with limited capital to contain the blast radius. A year ago, that's a months-long project. Data pipelines, model integration, trading interface, risk management. Real engineering. Now? I could have a working prototype by Sunday. And that realization — I could actually build this right now — is exactly what kept me up until 3am sketching architecture in a notebook.
Every idea now comes with an immediate follow-up: "I could actually build that." Not theoretically. Not someday. This week. That changes the way your brain processes ideas at a fundamental level. They stop being abstract daydreams and start feeling urgent. They have weight. They demand attention. And they don't stop coming just because it's two in the morning and you have calls at eight.
The Bottleneck Moved
For my entire career, execution was the bottleneck. You could only build as fast as you could code, as fast as you could hire, as fast as you could coordinate. Creativity was never the limiting factor because you always had more ideas than bandwidth. The backlog was always full. The constraint was always time and hands.
That ratio has flipped. With AI agents handling execution, I can now build faster than I can prioritize. The constraint has genuinely moved from "can I build this?" to "should I build this, and if so, when, and what do I not build to make room for it?" That's a completely different problem. And it turns out it's a harder one.
I work a full-time job during the day. Evenings and weekends, I'm building BCK Systems, pushing OnPlane toward beta, and now thinking about this day trader experiment. A year ago, that list would be laughable — one person making meaningful progress on side projects of this scope alongside a day job isn't real. But each of these is now genuinely achievable. The tools are that good. And that's intoxicating and paralyzing at the same time.
The Discipline Nobody Warns You About
Here's what nobody told me: the hardest part of having powerful tools isn't learning to use them. It's learning what not to build. When everything is executable, focus becomes the most valuable skill you have. And focus means killing ideas that are genuinely good. Not bad ideas — those are easy to let go. Good ones. Exciting ones. You have to let them die, or at least put them in a coma, because they're not the most important thing right now.
I'm still learning this. I caught myself at midnight last Tuesday sketching out the day trader's architecture — data sources, decision logic, risk parameters — when I should have been focused on OnPlane's beta. The idea was exciting. It was also a distraction. Knowing the difference in the moment, when your brain is firing on all cylinders, is harder than it sounds.
This is a new kind of discipline. Not the old "manage your time, eat the frog" advice. It's about managing your own creativity when the tools have removed every excuse for not acting on it. When you can build anything, deciding what to build becomes an act of willpower.
What This Actually Feels Like
My evenings look different now. I get home from work, and instead of unwinding I'm already thinking about what to build. I spin up agent teams, and in a couple hours I have working features that would have taken a full sprint before. Real features, tested and integrated. Then a new idea hits and I have to physically stop myself from opening a new project and just starting. The pull is strong. The activation energy is almost zero.
The excitement is real and it's the best kind — the builder's high of making things exist that didn't exist yesterday. I haven't felt this way since I first learned to code and realized I could make a computer do things. That same sense of limitless possibility, except now I have over a decade of experience to aim it at real problems instead of text adventure games.
But there's a cost. My mind doesn't downshift. I'm out with friends and half-present because I just figured out how to solve a problem I've been chewing on for days. Trying to fall asleep and suddenly I'm designing a monitoring system I don't even need yet. I catch myself mid-conversation, eyes unfocused, mentally building something instead of being present. I'm 33 years old and I have more energy for this stuff than I did at 22 — because the tools finally match the ambition.
The Good Problem
I know this is a good problem. Having too many things I want to build is about as good as problems get. The position I'm in — where the limiting factor on what I can create is my own ability to prioritize — is a privileged one, and I'm aware of that.
But I also know I need to get better at the prioritization part. The builders who win in this era won't be the ones with the most ideas or even the best tools — everyone's going to have access to the same tools eventually. They'll be the ones who can focus relentlessly on the one thing that matters most and resist the pull of the new and shiny long enough to finish what they started.
I'm working on it. I've started keeping a list — not a backlog, just a place to put ideas so my brain knows they're captured and can stop holding onto them. It helps a little. Some nights I can get to sleep by midnight. Other nights I'm still staring at the ceiling, designing something I have no business designing at that hour.
Tonight's probably going to be one of those nights. I had two new ideas during the writing of this post.