Open Source
At the end of 2025 I spent 100 hours on a single open source project. I did similar things over the past twenty years but always lost interest. This time was different. The software is actually used at my company. Spoiler: I will happily discontinue it.
My track record with open source is not great. I built a gamification engine that no one used. I wrote Symfony bundles for Magento product catalogs. Who is using Magento still? I created PHP extensions for a microframework that died years ago (bye Silex). I built parallel test execution for CodeceptJS. That one nearly broke me. Debugging parallel processes in someone else’s JavaScript codebase. I’ll never do that again and it’s not part of the original software anymore.
My super power: writing software no one uses.
In fall I built n8n-openai-bridge. It connects chat frontends to n8n workflows. Another niche thing.
What kept me going:
- The software is used at roadsurfer by regular people
- Actively asking for and getting feedback on Reddit and GitHub
- Offering contributor status to users
- Commenting on related posts and pointing people to the project
Then n8n released a built-in feature that does the same thing.
I am happy about it. No more maintenance and no need to support it over the next years. Bugs are someone else’s problem now. The infra team at roadsurfer has one less Helm chart to worry about. I get overly excited about new projects and quickly lose interest. This website will probably suffer the same fate.
Update January 4, 2026: I launched a new open source project called pop.