On my now page, there is a single sentence under the heading Learning:
… that writing about ideas is a good way to find out which ones hold up
I wrote that line without much ceremony. It was an observation, not a thesis. But the more I sit with it, the more I realise it is the closest thing this site has to a mission statement.
There is a particular kind of embarrassment that comes from writing down an idea you have been carrying around in your head for weeks. You open a text editor, type a heading, and begin. By the third paragraph, the idea is dead. It sounded profound in the shower. On the page, it is hollow.
This is not a failure of writing. It is writing doing exactly what it is supposed to do: exposing the difference between feeling you understand something and actually understanding it.
Paul Graham captured this well in his essay Putting Ideas into Words: “Writing about something, even something you know well, usually shows you that you didn’t know it as well as you thought.” The act of committing thought to prose forces a confrontation with your own fuzzy reasoning. You cannot hand-wave in writing the way you can in internal monologue.
Most people treat a blog as a publication channel: a way to broadcast finished thoughts to an audience. I have come to treat this site differently. It is a workshop.
When I write a post about agent-native architectures, I am not declaring a position. I am testing whether the concept holds together when I have to explain it to someone else. When I wrote about the ghost in the org chart, imagining what happens when AI agents begin to populate team structures, I was not reporting on a trend. I was trying to figure out if the idea made sense. The writing was the figuring.
This is why I added a disclaimer to the site:
If something I wrote contradicts an earlier post: good, that means I’ve been paying attention.
A contradiction is not a mistake. It is evidence that the workshop is working.
Thinking, in the unstructured sense, is generous to its host. It allows you to skip over the hard parts without noticing. You can hold two incompatible ideas in your head and never feel the friction because you never have to place them side by side.
Writing removes that luxury. It demands sequence, causality, and resolution. You must decide what follows from what. You must close the loops you open. If a claim cannot survive the transition from mind to markdown, it was never a claim; it was a vibe.
This is why many of my posts are summaries of other people’s work: conference talks, research reports, and books. Summarising forces a deeper engagement than reading alone. You cannot summarise what you cannot structure. And if you cannot structure it, you probably did not understand it.
Not every idea collapses. Some make it past the third paragraph. Some survive the edit. A few (the ones that genuinely hold up) get published.
The site currently holds over fifty posts. I do not stand behind all of them equally. Some are conference notes, raw and functional. Others are artefacts of who I was at a particular moment. But the ones I return to, the ones I reference in newer writing, are the ones that survived the test of the page.
They are also the ones I am most likely to disagree with in five years. That is not a problem; it is the point.
This is not abstract advice about journaling. For anyone working in technology (architects, developers, managers) writing is one of the few tools that scales clarity.
A design document forces you to articulate trade-offs. An architecture decision record forces you to state why you chose something, not just what you chose. A post-mortem forces you to reconstruct causality from fragments of log data and memory. In each case, the writing is not a by-product of the thinking. The writing is the thinking.
I have spent over twenty years in IT, and the most valuable skill I have developed is not a programming language or a framework. It is the habit of putting ideas on a page and letting the page tell me what is wrong with them.
If you want to know whether an idea holds up, write it down. Not in bullet points. Not in a tweet. In prose, with paragraphs, where one sentence must lead to the next.
You will find out quickly which ideas are solid and which were just comfortable.
That is what I am learning. And I keep learning it every time I sit down to write.