About

I'm Stephen Byatt, a software engineer. I work in telematics and connected-vehicle systems by day, but this site — Learning Out Loud — ranges a good deal wider than that: it's where I think in public about building software well, using AI with judgement, and staying open to being wrong.

I've been coding since 1980. Yes, I'm that old. I started with a Sinclair ZX80 at just 10 years old, and nowadays I work with a broad development toolkit deploying huge software solutions in the cloud. I have a particular interest in visualising insight from very large real-time datasets. I also code for hobbies - such as the creating the software in my kit car that runs the lights, exhaust valves, cooling, haptics and aero. I also enjoy driving, particularly Mazda MX-5s, on track. I guess that makes sense when my day job is often about helping drivers be safer or faster.

How these articles are written

The ideas here are mine, formed through experience and a fair amount of arguing with myself. My process is deliberately conversational: I talk a topic through with an AI assistant — testing the idea, pushing back on it, and refining the position until it holds together — and then ask it to summarise that discussion as a draft article. I then edit the draft, cut anything I don't mean, and publish only what I am willing to stand behind.

So the AI is a thinking partner and a drafting tool, not the author. The judgement, the claims, and the responsibility for them are mine. The full revision history of every article is public in this site's source repository, so how a piece developed is a matter of record — which is rather the point ofProvenance in the Age of AI Slop.

Get in touch: stephen.byatt@gmail.com.