TAG

# Claude Code

2026-08-03

AI's Standard Isn't in the Model. It's in the Last Version You Nodded At

Once you compromise a single time, AI lowers the bar, then keeps compromising, and ends up producing garbage — that observation is correct. The reason isn't that the model isn't smart enough. It's that a standard doesn't exist inside the model at all; it exists in the context. The most recent sample in the context is the standard, and every one of your fine, ship it moments is calibrating it. This piece uses my own concrete failures from three days to take the mechanism apart: a cover that cropped $45 billion down to $5 billion, a paragraph added purely to hit a word count, three straight days of believing a success report that was wrong. It also explains why written rules fail the same way — the number of rules is a counter of failures, and it only covers the failure modes you already thought of.

2026-07-08

Hundreds of MCP Servers and Claude Skills, and Barely Any Are Truly Free and Open Source. I Checked Them One by One and Turned It Into a Directory

I wanted to add a few MCP servers to Claude, and the more I searched the more annoyed I got. Out of hundreds, half are only free if you hand over an API key. A whole batch flies the open-source flag but really means source-available, not for commercial use — Sentry's MCP is under the FSL license, and Anthropic's own document skills flatly say all rights reserved. Some repos don't even have a LICENSE file, which legally means all rights reserved by default. The ones that are actually MIT or Apache, install-and-go, no account needed, you can only tell apart by opening every LICENSE one by one. I went through sixty-odd of them and collected the genuinely free and open ones into a bilingual directory: To Be Free. This piece is about how I sorted them, which of the truly free ones are worth installing first (gstack, ruflo, the official MIT servers…), and why this is the next step in my rebuild-free-software line.

2026-06-02

Stop Learning, Start Doing: The Only Thing Standing Between You and AI-Native PM Is Action

In the AI era, product managers don't need to hoard knowledge — you'll never out-know AI. Ask it on the spot instead of studying in advance. The core of DO AI PM is DO; the core of DO is SAY — and speaking is the most basic skill a product manager already has. There's no prerequisite. The only barrier is that you haven't started.

2026-06-01

Vibe coding is already obsolete — and that's great news for product managers

When AI writes the code, what's left is judgment: deciding what to build, for whom, and what 'good' means. That has always been product management. Here's why not knowing how to code can be an advantage — and how to do it on purpose.

2026-05-30

Speak it, AI builds it: I made this website with a single sentence

The first doaipm post. Not knowing how to code is an advantage — this very site was "spoken" into existence with Claude Code.