A Brief Detour: Two Writing Challenges and What Came Out of Them

Regular Clusterflick series readers: I got distracted. Twice 😅 In the last week I entered a couple of dev.to writing challenges, and both turned out to be good excuses to write about things that were already on the series roadmap — just earlier and in a slightly different shape than I'd originally planned. The first was the 1️⃣ DEV Weekend Challenge: Community , which I used to write about the film club discovery and "near me" features, which I finally took the time to build. The second was the 2️⃣ Built with Google Gemini: Writing Challenge , which pulled forward what was going to be a later post about using LLMs in the data pipeline. Both are standalone submissions, but they're very much part of this project: Making London's hidden film clubs discoverable Three Things I Learned Using LLMs in a Data Pipeline The LLM post in particular covers things I'd have gotten to eventually in this series — the matching pipeline, the reason key trick, defensive JSON parsing. Worth a read if you've