One day the hospital decided to make short promo videos.

Short video is everywhere now, so a lot of staff jumped in. Plenty of good clips came out of it. But not everyone could play. Some of my coworkers weren’t comfortable with computers. They had the photos. They had the footage. They just didn’t know how to edit any of it, so they made nothing at all.

One of them was a manager in our administration office. She told me she had all the raw material, but video editing felt impossible. She wanted to try. She couldn’t bring herself to start.

I started building that same day. The goal was simple: a video editor anyone could use.

Build for the person who struggles most

Everything I make follows one rule. It has to be easy. Comfort for the user is always priority number one.

There’s a reason. The people at my hospital range from barely twenty to past sixty. Younger staff grew up with screens. For older staff, a single small button can be a wall. If I want nobody left out, the tool has to work for the person who finds it hardest. Not the power user. The most unsure person in the building. That’s the bar.

Rebuilding the features of a paid editor like CapCut, for free, came with limits. But there were ways through. Automatic subtitles were the surprise. I bolted on a speech-recognition tool called Whisper, and it transcribed speech into text better than some famous commercial apps. It just listened and typed.

Staff started editing video with no trouble. I called the tool EasyReel.

What I was actually trying to build

Here’s the thing. I didn’t really care about a video editor.

What I wanted was to give older staff an experience. The feeling that AI lets you do the thing you’d written off for life. The thing you never dared to attempt. Not me explaining it. Them doing it, with their own hands.

The coworker past sixty who gave up because she couldn’t edit. The moment she finishes a video herself. I wanted to see her face in that moment. That’s different from just saving someone work. It’s taking a person who couldn’t, and making them someone who can. Not pity. Putting the thing in their hands and saying: you can do this too.

Reaching the patient, not just the staff

Then I built something that reached past the staff and touched the patients.

It’s a cognitive-rehab tool I called EasyBrain. Cognitive rehab means training to recover functions like memory and attention. For this, my hospital was paying for a program that cost about 40 million won — roughly USD 30,000. To me the features weren’t that impressive. The price was.

I couldn’t just be reckless about it. To formally bill for this treatment, regulations require an officially registered program. So whatever I made could not replace the official one. I had to draw that line clearly, and I did.

But I flipped the question around. What if I didn’t bill at all? What if it was only a free helper, an extra tool to give patients more practice? If patients could train more without spending a single won, why wouldn’t I make it?

I also had an edge nobody at a software company has. I treat the patients and I build the tool. So I could try it on a real patient, watch the response, and fix it on the spot. No company can do that. The end of our work is the patient getting better. If I can help that for free, that’s just the road to take.

Twenty-two tools, one direction

It went like this, over and over.

I gave back the hands that cropped and retouched a photo every time we made an employee ID card. I made it so a director could search one keyword instead of digging through years of paper diaries. I tracked treatment adherence so no patient slipped through without the care they were owed.

Line up all twenty-two tools and they look like different programs. They weren’t. They all pointed at the same place.

Lighten someone’s hands. Make someone able. Help someone heal. In the end, every one of them was a person.

So if you’re a non-developer wondering what to build first, here’s the only filter that has held up for me: don’t start with the coolest idea. Start with a specific person who’s stuck, and build the smallest thing that unsticks them.

And then, after staring at all twenty-two laid out in a row, I noticed something strange. But that’s the next story.