1002003004005006007008009001000110012001300140015001600170018001900200021002200230024002500260027002800290030003100320033003400350036003700380039004000

-1000-900-800-700-600-500-400-300-200-1000100200300400500600700800900100011001200130014001500160017001800190020002100220023002400250026002700280029003000

/00 Insights

Notes from the work.

Essays on applied AI, 30-day MVPs, and senior engineering teams, written by the people doing the work.

/01 Latest

//001

Essay8 Jul 2026

Why most AI features never reach production

The demo is the easy part. Most AI features die in the gap between a convincing prototype and a system your business can actually run on.

By: Qualsix Team
Filed: AI

Read the essay

/02 Archive

More essays

00212 jun 2026
The 30-day MVP, nine years in
We still run the 30-day MVP because it forces every hard conversation to happen early.Delivery
00320 may 2026
Evaluation is the product: LLM features you can trust
If you cannot measure an LLM feature, you cannot improve it, and you definitely cannot promise your users it works. Evaluation is where trust comes from.AI
00415 apr 2026
Hiring is not the only way to add senior engineers
A senior hire takes months to land and longer to ramp. Embedding a vetted senior engineer takes a week, and you can prove the fit before you pay.Teams

Why most AI features never reach production

The 30-day MVP, nine years in

Evaluation is the product: LLM features you can trust

Hiring is not the only way to add senior engineers