Notes from the build.

Legacy systems don't need more modules

Four problems every monolithic system has—data entry, locked-up access, offline reach, and oversight—each solved in days with a thin AI layer over the data the system already holds.

Adrian Chmiel

May 29, 2026

How far does Claude Design get you?

Anthropic's Claude Design, tested on a multi-view resource management dashboard: how it holds a design system across screens, how fast prototypes arrive, and what the Claude Code handoff is worth.

Jerzy Czarnecki

May 22, 2026

AI for QA: A First Look

Can AI handle QA? We tested Claude Cowork and OpenClaw on acceptance criteria, test plans, and UI automation—which one succeeded at repetitive testing tasks, and where free models still fall short.

Dawid Leszczyński

December 3, 2025

How to do Vibe Coding as a Developer

Some developers hand half their coding to AI. Others won't ship a line of it. A one-week vibe coding experiment—full workflow, tool configs, and the verdict.

Wiktor Podobiński

July 31, 2025

Human Prompting: A Reversal by Design

We've spent the last few years learning how to prompt AI. But lately, the AI has started prompting us. And it's proving better at it than we expected.

Yevhenii Hordiienko

July 16, 2025

Task-Doing Agents in Real Time

How AI voice agents handle phone calls, talk like real people, understand intent, and book meetings before the caller hangs up.