In this issue...
Quantum #670 had an open rate of 31% and a click rate of 10%. The most clicked link was (unsurprisingly) this Eclipse Map.
·Over the past few weeks, I've created MapSplit, an interactive online tool that allows you to select statistics across the UK, and play with splitting them geographically North/South or East/West.
It was inspired by the numerous maps that show where the UK is 50/50 on population (most recently by Quantum-featured Alasdair Rae's population centre of the North).
Please share on your social media, and please share your favourite maps and splits 😃
·Don't miss the opening and closing articles about coding being the easy part of implementing software. Opinions vary, but I think one of the articles is missing the point by focusing on "easy" in absolute terms: no one really thinks that coding is easy generally speaking, but anyone who's put software into production will know that the slow part of the process is elsewhere (governance, support, etc).
On a similar note, there's a brilliant little article that suggests that LLMs reward expertise, and I can't disagree with it, at least at this stage of the development of AI. This applies both at a small scale (e.g. if I use Claude Code to develop software, I'm going to be much better with a programming language that I know than with one I don't) and at a larger scale (I can definitely use it better for computer science than for civil engineering). There is a subtle difference in saying that this is about "good prompting": anyone can write good prompts, but writing good prompts with experience of the subject matter makes a difference. However, I can also say that with models like Opus 5 this gap seems to be reducing, at least anecdotally.
Of particular interest: John Gruber's article retraction (about an interesting App Store rejection), and the San Francisco version of an AI-generated isometric map.
'till next week,
Giuseppe @puntofisso.bsky.social
