The age of sprints in tech is dead. We are in the age of continuous building

I think the age of sprints in tech is over. We're moving into a phase of continuous building.

For a long time, most teams I've worked with ran in two-week sprints. Planning, grooming, estimation, standups, retros. It was just how things were done.

And honestly, it made sense. Building software used to be expensive. Shipping took time. Changing direction was costly.

But AI has changed the dynamics quite a lot.

These days, execution isn't really the bottleneck anymore. The harder part is deciding what's actually worth building next.

With tools like Cursor, Claude, GPT, and agents, I can spin up features very quickly, refactor large parts of a codebase, and prototype ideas in hours instead of weeks.

So planning two weeks ahead feels less useful than it used to. A lot of assumptions don't hold for very long.

The feedback loop is much tighter now. You build something, ship it, see how users respond, and adjust almost immediately.

There's less need to wait for sprint boundaries or batch work into fixed cycles.

Sprints optimise for predictability. The way I work now optimises for responsiveness, and those feel like very different mindsets.

My current workflow is closer to a continuous loop. A moving backlog, daily prioritisation, letting agents handle most of the execution, me reviewing and steering, shipping often, and learning from real usage.

No big ceremonies. No velocity charts. Just a tight feedback loop between ideas and reality.

Sprints were designed for a world where building was expensive. AI created a world where deciding what to build is the hard part.

And in that world, waiting two weeks to ship something doesn't really feel necessary anymore.

It just feels slow.