In this video I tell the story of the infamous “144 point Jira story” and what it looks like when agile goes wrong. You know the vibe: everyone is focused on their own ticket, their own microservice, their own PR… and nobody is watching the end-to-end behavior. I use a YouTube-algorithm-style example to show how a feature can become half-implemented, super buggy, overly engineered, and so coupled that changing one thing breaks five other things. At that point, it’s not just annoying—it’s risky, hard to test, and basically impossible to reason about. So what did we do? We created one giant story: redesign and implement the recommendation algorithm. Yes, 144 points. The point wasn’t “lol big number,” it was acknowledging reality: most of the work was design and requirements thinking, then building a foundation, then finishing implementation, and finally tests (which should be happening throughout). I also talk about why junior engineers get burned when they skip design and jump straight into coding, and why tests are the insurance policy that make refactors possible. It was stressful, I probably wouldn’t do it again, but it paid off—once the tech debt was handled, we could ship enhancements week after week.