karozi / gpt-5-2-evaluation.md

Last active December 23, 2025 18:17

Why 'it feels worse' is the wrong metric for evaluating GPT-5.2 - Product with Attitude by Karo Zieminski

Why "it feels worse" is the wrong metric for evaluating GPT-5.2

Author: Karo Zieminski Published: December 23, 2025

Read Full Article: Product with Attitude on Substack

ChatGPT 5.2 isn't flashy in the way the internet likes it. There was no viral demo, no trending reels, and nobody fainted.

karozi / Rules-For-AI.md

Created December 1, 2025 20:26

Whenever your agent breaks something, don’t just fix it - ask this:

Based on our session and the fixes we made, what are the key things that should be added to my replit.md to prevent these issues in the future.

#RulesForAI #vibecoding #speccoding #PwA

karozi / awesome-claude-skills-gist.md

Last active December 24, 2025 02:26

Published: 2025-10-18

Author: Karo Zieminski

About

A growing library of repos, videos and skills for Claude Skills (100+)

karozi / Anti-Regression-Agent-For-Replit.md

Last active August 21, 2025 07:38

Anti Regression Agent for Replit

Safeguards your system integrity, prevents regressions, and enforces safe engineering rules.

Use this prompt at the start of every Replit session, especially if your agent starts:

karozi / Books.md

Last active August 12, 2025 22:23

Books for PMs, Founders & Solopreneurs

Curated by Karo

A ruthless edit of books that actually move the needle on product, UX, and thinking clearly.

Titles that sharpen product sense and help you ship smarter.

Curated for velocity: practical, battle-tested reads.