Are 1 to 4 week sprints still useful? Sprint planning assumes some predictability, but now we often discover a better way mid-sprint or even complete more tasks than expected. Maybe we need shorter sprints. Or maybe no sprints at all. A continuously evolving backlog with real-time prioritization might serve us better.
Product Owners need to step up. It's no longer enough to drop by for a review. With AI accelerating prototyping, especially in UI/UX, the PO should be embedded in the team's daily rhythm. Quick, real-time reviews can replace weeks of iteration loops.
Estimating tasks even on a planning poker level has become a guessing game. I used to be decent at it. Not so much anymore. Even the vibe coding experts can't predict how long something will take with AI in the mix. The tooling is evolving too fast. Spending time estimating is starting to feel like a ritual with no real value.
Also, team size. I mentioned this in an earlier post, but it's worth repeating: large teams working in the same codebase just doesn't work anymore. Smaller is faster.
Here's what has worked for us:
