AI Disruption

AI Disruption

Superpowers vs gstack: AI Coding's 3 Weak Spots, Half-Fixed

Superpowers vs gstack compared. Learn how these AI coding frameworks fix different failure modes: wrong requirements, poor quality, and broken delivery chains.

Meng Li's avatar
Meng Li
Aug 04, 2026
∙ Paid

“AI Disruption” Publication 10,400 Subscriptions 20% Discount Offer Link.


After installing both Superpowers and gstack, Claude Code refused me for the first time.

It said: “Are you sure you want to do this? I think the problem you’re trying to solve might not actually be this one.” Then it genuinely stopped and waited for me to think it through—without writing a single line of code.

That was gstack’s /office-hours skill at work. And before that, Superpowers had already trained me into a habit: never declare something done until the tests have all passed.

Together the two frameworks give you 37 skills. It took me two weeks of tinkering to realize one simple thing: they treat completely different diseases. Jesse Vincent’s Superpowers manages “how to write good code.” Garry Tan’s gstack manages “what to build and how to ship it.” One is engineering discipline; the other is product process. Installing only one isn’t enough; installing both feels too heavy.

Total skills across the two frameworks: Superpowers 14 + gstack 23.

User's avatar

Continue reading this post for free, courtesy of Meng Li.

Or purchase a paid subscription.
© 2026 Meng Li · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for great culture