At some point the tools I had built stopped being side projects. They became the thing 660 people leaned on every day.

And that changed the job. Building something is the easy part. I know that sounds backward. But once people depend on a tool, your real work is no longer making it. Your real work is keeping it from breaking. That part never ends.

The night drives

The first lesson was small and stupid.

One evening I fixed something just before leaving, didn’t check it properly, and drove home. By the time I got there, the tool was misbehaving. So I turned the car around and drove back to the hospital to fix it. That happened more than once.

I assumed this was just how it had to be. I even had a neat excuse for it. For security, I had locked all my tools inside the NAS — a small always-on storage server that sits in the building and runs everything. Reaching into that NAS from outside felt like tearing down the very security I had built. So I never even tried. In my head, the only way to fix a problem was to be physically sitting in front of it.

Then that familiar question came back: is there really no other way? I went looking, half expecting nothing.

There was a way. It is called Tailscale — a tool that builds a private, encrypted tunnel from anywhere straight to your own machine. Earlier I had used Cloudflare to give staff a door into our systems from outside. This was different. This was a door only I used. And because only I walked through it, it was actually safer, not weaker.

After I set it up, I could fix a tool from my couch. A problem at midnight no longer meant turning the car around. From anywhere, in real time.

What stuck with me wasn’t the tunnel. It was the realization underneath it. The road existed the whole time. I just didn’t know it was there. So much of what I had given up on as “impossible” turned out to be a path someone had already paved — safely — years before I needed it.

The blackout

The next lesson was much bigger.

A transformer at the hospital exploded. The whole building lost power. The NAS shut off, and the moment it did, every single tool I had made went dark at once. Everything froze.

That was when it hit me. The things I had built were now things that, when they stopped, the whole hospital felt it. I had started this just to take one small annoyance off one person’s plate. Now it sat in a place where it was not allowed to fail. That felt good and heavy at the same time.

I needed a plan. And oddly, the answer was sitting inside the same day the disaster happened.

Our hospital is three buildings. The power for two of them is wired together, but the third runs on its own line. So when the main side went dark, the third building stayed lit. One side went down and the other kept running.

There was the answer. Don’t keep the NAS in one place. Split it.

Building something that can’t go down

I was going to need bigger storage soon anyway. So I decided to grow and split at the same time.

The plan is simple. Right now the NAS sits in one building. First I put a bigger, better one in the separate building. I move the data over to it. Then I place a matching unit back in the first building. Two identical NAS units, in two buildings on two power lines, each holding a copy of the other’s data. If one building loses power, the other keeps running. A system that stays up 365 days a year.

A few days later, the facilities team added a UPS to the NAS — an uninterruptible power supply, a battery that keeps a machine alive for a few minutes after a blackout so it can shut down safely instead of just dropping dead. That detail mattered to me beyond the hardware. For a long time I had been carrying this alone. Now the hospital was starting to put its hands under it too.

The weight, and the lesson

Looking back, it’s almost funny. A person who didn’t know what code was a year ago was now lying awake thinking about how to keep a system alive through a power outage.

That is what it means when the thing you built gets heavier. And here is the part worth keeping, if you build anything real: the day people start depending on your tool is the day your job quietly changes from building it to protecting it — plan for that day before it arrives, not after the lights go out.

I didn’t plan for it. I learned it by driving back to the office at night and by standing in a dark building. But carrying that weight changed me. Somewhere in the middle of it, without noticing, I had grown.

And something stranger was happening at the same time. Building new things was starting to feel almost easy. That’s where this goes next.