← Back to archive

How AI Changes Shape Up and the Cool-Down Week

5 August 2026 · 54 views · 0 likes · 0 reposts · 165 blog views

AI automation decision-making product productivity

I read an interesting take on what happens to Shape Up now that AI has changed the bottleneck. Shape Up is 37signals' product method, from Ryan Singer's book - you work in 6-week cycles with a 2-week cool-down between them. Instead of estimating a feature you give it an appetite - a fixed time budget the scope has to fit inside. Scope hammering is how you cut the work down to fit, and if it doesn't ship in six weeks it doesn't get an extension.

The whole method was built around scarce programmer time, and features that once took 4-6 weeks may now take a week. All of them assume building is the expensive part.

"cool-down" is becoming "tune-up". The new name fits the purpose better - not recovering from a sprint, but deliberately improving the product.

Tune-up has a concrete focus: bugs, refactoring, rough edges, inconsistencies, dev/AI tooling, etc.

The work stays self-directed and people choose what to improve based on the problems they personally run into.

Tune-up shrinks from 2 weeks to 1 (or even smaller?!). Two weeks was long enough to lose momentum, especially now that meaningful work can be finished in a few days.

The gap between cycles still matters. It leaves room to finish shipping, announce releases, plan the next cycle, and start it with a clean slate.

Small improvements compound. A tune-up week goes to the product’s paper cuts rather than another big feature.

Judgment becomes even more important: what to build, what to polish, and what actually improves the product.

Source: world.hey.com/bb/tune-up-d6c…

View on X

How AI Changes Shape Up and the Cool-Down Week

Originally posted on X

PS: I'm pretty active on 𝕏 if you'd like to follow along

Related posts

Archive

2026

2025

2024

2018