Why are you still using branches?
Notes
- Ball argues for trunk-based development: push small changes straight to main instead of living on branches.
- He treats long-lived branches as a smell; even in teams that used PRs, a branch lasting more than a day or two worried him.
- The idea is continuous integration with the trunk, which he says has worked in large organizations and is how the Ruby project works.
- The common objection is that some work “can’t go on main yet” because of migrations, architecture changes, or a big refactor.
- His reply: split the work into safe increments—add a column and deploy it, then helpers, then migrate pattern A to B one by one.
- He puts true exceptions at about 0.5% of cases, not 50%.
- He admits the old cost of this style was real: feature flags, staged rollouts, two configs at once, and other “meta programming.”
- That busywork sometimes made a fat branch feel easier, even though start-and-stop delivery is riskier than a system that keeps flowing.
- Agents, he says, are good at those chores, so continuous integration onto main is now much easier.
- He says he is not mainly trying to convert people to trunk-based development.
- The point is to use agents to rethink how you develop software, not just to replace typing.
Next question: why are you still using branches? We talked about PRs. I don't think they make sense anymore, if you rethink them from first principles. Yes, if you only use them to trigger CI, then sure, go ahead. That's a fat commit, not a PR. But for mandatory code review?

































![founder content is a system [full playbook]](assets/2074879329536774615-cover-3_2074877807600009216.jpg)











