Show Your Working
Useful Before it's Finished
Writing documentation can be a chore, so why are we suddenly keen to write instructions for Claude? Perhaps it helps that we're already in our editor, rough bullet points will do, and the payoff is immediate: our tools work better for us. Meanwhile, useful explanations disappear into chat histories, and half-formed ideas stay private while colleagues work through the same problems. What can writing for machines teach us about sharing our thinking and finding people to think it through with?
Taking a few detours through Alexander Graham Bell's experiments with sound and Delia Derbyshire's working tapes, we'll explore what happens when ideas travel beyond their original purpose - and why the working matters alongside the finished result.
I'll share some observations, tensions and low-effort things I've been trying: making rough notes findable, distinguishing experiments from recommendations, and leaving enough reasoning for someone else to question your approach or build on it. In the spirit of the talk, I'm sharing my working too - not a finished method, but useful breadcrumbs and a starting point for comparing notes. This isn't about documenting everything or writing it twice for humans and agents. It's about writing for yourself first, while making it easier for others to find your thinking and join in.
Andy Normington
Senior Embedded Software Engineer
Focusrite
I work across Focusrite and Novation. Embedded systems are my technical home, but much of my day job happens between teams: improving the tools and workflows that help engineers get products out of the door, shaping our shared engineering platform and helping teams get the most from it, and turning security and architecture challenges into practical ways forward. I enjoy joining the dots, sharing what works, and making the next team’s path a little less bumpy. I came to software engineering through music and audio, via electronics, and I’m still curious about what happens when software meets the real world, which has a habit of ignoring the diagram.