The three-hour problem

There is a folder in my head labelled "watch later" and it never empties. A three-hour podcast everyone is quoting. A conference talk with the one idea I actually need. A long video that buries ten useful minutes inside a hundred and seventy. I wanted the substance. I did not have three hours.

When you spend your days building apps, the hours you would have spent learning are the first thing to disappear. So I did the thing I always do when something annoys me enough: I built my way out of it.

Built to buy back time

Whisperer AI started as a tool for one listener: me. Hand it a recording, a video, a podcast, and it hands back a clean transcript, a summary, and the parts that actually matter. Not a vague list of bullet points, but enough that I could skip the three hours and still hold a real conversation about what was said. If I had a question about a recording, I could just ask it, and get an answer drawn straight from the source.

It worked. I used it constantly. The watch-later folder finally started to move, and the time it gave back went straight into building the next thing.

From a tool into a library

Once it was part of my day, the gaps became obvious. A summary is useful in the moment, but I wanted to keep things. So I added PDF export, so any transcript or summary could leave the app as a clean document. Then folders, so all of it had somewhere to live.

That small change turned Whisperer from a one-off tool into a library. People started keeping separate shelves: one for meetings, one for podcasts, one for YouTube, one for university lectures. The app I built to get through a single podcast became the place people kept a whole semester of classes.

The reviews I did not want to read

Here is the harder part to write. Whisperer worked for me, which meant I slowly stopped seeing it clearly. The bugs that mattered were not on my device or in the way I used it, so I never hit them. My users did.

The way I found most of them was the way no developer enjoys: a one or two star review. It stings, it pulls the rating down, and there is no reply that fixes the number.

But every one of those reviews was a map. Each frustrated note pointed at something real I could not see from the inside, and once I could see it, I could fix it. The rating is slow to recover. The app is better because of the people who took the time to complain.

What it taught me

Building for yourself is the fastest way to start and the easiest way to go blind. You stop being a fair test the moment the thing works for you. Whisperer reminded me that shipping is not the end of listening. It is the start of it.

The watch-later folder still fills up. The difference now is that I get to the bottom of it, and so do a lot of other people.