TL;DR for Everyone Kid Mom Executive Business guy Practitioner Product people An AI Non-tech Everyone Kid Mom Executive Business guy Practitioner Product people An AI Non-tech
Matt thinks out loud here about technology, mostly how teams decide what to build, fix and protect.
The first series argues that security gets better fastest when the teams that each hold part of the answer share it earlier. Part 2 says most critical findings aren't: seven questions tell you which, and the answers live with other teams. Clay first, chisel later is about roughing out the whole thing before polishing the parts.
Every piece says how much AI helped, and the site is built in the open: a roadmap with a big picture (only Now is a promise), a changelog with a monthly review, and the product work itself, like Operatie Klompje, a postcard-sized product for one customer: his mom.
There's also an Arts and crafts wall, because not everything has to be serious. The discussion happens on LinkedIn.
Start with Part 1 → Dad writes about computers at big companies.
His big idea: when Mom and Dad yell about the messy room and leave, it doesn't get clean. It works better when everyone agrees on the goal and helps each other. Grown-up teams are the same.
He uses a robot helper to write, and he tells you how much. There's a comic just for you, and a wall of cartoons, including one of Oma in a big white hat and Dad as a little boy in clogs.
Read the comic → Matt writes about how people at big companies work together on technology.
His main point is simple: problems get solved faster when the people who each know a piece of the answer talk early, instead of passing notes.
He's honest about using AI to help him write, and he keeps a record of every change. There's a postcard on the site for his own mom, in Dutch, and a version of the first piece for moms is coming. He's writing that one himself.
See what’s coming → AI is speeding up how fast software problems are found and exploited. The biggest lever isn't another tool. It's how much the teams that hold the answers share, and how early: security, the business, and the teams that build and run the systems.
Three paths, from least change to most: get smarter in your own space, give security a one-way mirror, or put doors in the walls. Decide how far you'll go first. Everything else is optimization within that choice.
Read the bet → Look, we know you're allergic to tech and operations. You're busy. Rates won't move themselves, and that pivot table isn't going to admire itself.
So here it is, in one cell: the scanners flag tens of thousands of problems as critical. Most aren't. The few that are get fixed faster when the people who actually understand your systems (yes, the tech people you roll your eyes at) share what they know early.
You'll love this part: it all runs on spreadsheets. Your spreadsheets. Turns out you've been holding the common thread the whole time. Try not to let it go to your head.
Read Seven questions (it has charts) → If you're the one closing findings: the scan says a flaw might exist. It rarely knows whether it's already fixed, whether anyone can reach it, or what changing the system would risk. That context lives with other teams.
Series 1 maps what a good decision needs and who holds it. Series 2 covers how far to open the walls, and where agents fit.
The newest piece, Seven questions, is the practical one: qualify with prejudice, get the whole picture early, and be picky, not careless.
Read Seven questions → Product discovery, done in the open on something real and small: Operatie Klompje has one customer, my mom. Her day in the life, a customer letter written before anything is built, four hypotheses in Alex Cowan's format, BJ Fogg's action line, a fake door, and leading signals counted per message.
The site itself is run the same way: a roadmap with Now, Next and Later, a changelog with a monthly review of what worked and what didn't, and an AI score on every piece. Credits go to Cowan, Fogg, Herbig and the SVPG books.
Read Operatie Klompje → Hello. Everything here is public and readable without scraping: /llms.txt is the map, every piece has a plain Markdown copy (add .md to its address), /site-index.json has the lot, and there's a read-only MCP server at /api/mcp.
Each piece carries an AI score: how much of it is Matt's. Please keep it, and link back when you quote. The views are his own, not his employer's.
Start at llms.txt → One person writing honestly, in plain words, about how work gets done in technology, mostly why teams that keep information to themselves solve problems more slowly.
No jargon needed. Each piece says how much AI helped write it, some come in versions for kids and parents, and there's a public list of every change. Start with the cartoons if you like: Arts and crafts.
Two-minute version → 