How I use AI

People sometimes ask whether AI writes this site.

It helps me write it. There's a difference.

The thinking is mine

Everything I publish starts with something I actually did.

An experiment on my own site. A probe against an API. Log data from my infrastructure. A feature I tested when it shipped. Something I broke, tested again, and eventually understood.

The experiments are mine. The data is mine. The findings, opinions and mistakes are mine.

AI helps me turn all of that into something readable.

I usually start with notes, data, screenshots and half formed thoughts. AI helps me structure it and produce a draft. Then I edit it, check the claims against the data and publish it.

Nothing goes live unless I've verified it myself.

I also don't scrape or aggregate other people's research and repackage it as my own. If I'm writing about a finding, I've tested it.

Why I work this way

Most of what I research changes ridiculously fast.

AI products get updated constantly. Features get patched. Endpoints disappear. Retrieval behaviour changes. I've published findings that stopped working weeks later because the underlying system changed.

Spending two weeks polishing every article by hand would be pointless if the research is outdated by the time I hit publish.

I'd rather get useful, verified findings out quickly than prove I personally typed every sentence.

This site is free. My time isn't infinite.

There's no paywall. No gated PDF. No course waiting for you at the end.

I write it. You read it. That's pretty much the entire transaction.

I also run two companies and raise three children. This site gets whatever time is left.

AI assistance is often the difference between publishing something useful while it's still relevant and leaving it in my notes until it goes stale.

I'm not writing for rankings

I don't open a keyword tool, find something with search volume and decide to write 2,000 words about it.

I write about whatever my experiments uncover.

Sometimes those posts rank. That's nice, but it's not why they exist.

My test is much simpler:

Can someone read this today and actually do something with it today?

If the answer is no, I'm probably not interested in publishing it.

And yes, Google is fine with AI assisted content

Google has been clear about this since 2023. Using AI doesn't automatically violate its guidelines. The problem is producing content at scale primarily to manipulate search rankings, regardless of whether a human or a machine produced it.

That's not what I'm doing here.

And ultimately, readers decide whether something is useful.

In the two months leading up to mid August 2026, I have over 1M page views. Cool eh?

I'd rather judge the work by whether people find it useful than by which keys were pressed to write the sentences.

There's also a fairly obvious point

I spend a large part of my time researching how AI systems crawl, retrieve, reason about and cite information.

I use these systems every day.

It would be pretty strange if I refused to use them to help me write about them.

If that's still not for you

Some people simply don't want to read anything written with AI assistance. That's completely fair.

Now you know exactly how I use it, and you can decide whether this site is for you.

No hard feelings either way.