Building in Public
For me, a medium-heavy user of AI tools, the current value and future potential of AI is obvious. But I’m also aware that my context — working in tech-heavy companies, spending frankly too much time on my phone and computer — is not typical.
Being inside the AI hype bubble on LinkedIn makes it easy to forget that the excitement and concern I feel about AI is simply not something most people share. Many are sceptical, pointing to real risks: copyright violations, deepfakes, cybercrime, job losses, environmental costs. Others don’t care at all — AI feels like a tech thing, relevant only to programmers in large enterprises or startups. The technology-heavy discourse bores them.
I find myself thinking a lot about how to reach these people. Because my concern is this: the public conversation about AI — and by extension, policy — is being driven almost entirely by people already inside the bubble. The enthusiasts believe the potential is almost unimaginable and could materialise within years. The sceptics have a much more conservative view of the benefits, and a much longer timeline — if they consider transformative AI a realistic possibility at all.
That imbalance is a problem. If AI does evolve into something with larger societal impact than the industrial revolution — and if that happens in 5–10 years rather than over a century — we need as many informed, diverse voices as possible shaping the norms, the ethics, and the governance. That requires people outside the bubble to understand what’s actually happening, and to have grounded opinions about it.
So rather than talking in abstract terms about risks and benefits, I want to show rather than tell. Coding agents like Claude Code have enabled me — a fairly technical person, but by no means a programmer — to build small applications that solve real, tangible problems I actually face. I’m going to keep building these, and share them freely as open source, packaged in a way that’s useful whether you’re highly technical or not.
I’ll also be transparent about how I build them. Coding with Claude Code is a fundamentally different experience from traditional programming, and I think it’s far more accessible than most people assume. I’ll be honest about where AI helps, where human judgment is still essential, and how to think about concerns like security, privacy, and safety even when you’re not a software developer.
The first project is already here. If you’ve ever had something mysteriously steal keyboard focus on your Mac — your typing vanishing mid-sentence, no idea what caused it — I built a tool that catches it. Read about it in the next post.