Thursday, 12th November, 2026

Used Claude to make a mechanics of maths question generator site for my wife (maths teacher). As always it looked pretty good out of the box. She’s not interrogated it yet but I said anything can be changed in any way. This used the most tokens out of everything I’ve done so far. Almost hit the 5 hour limit after about one hour. I see there’s some promo to use Fable but Opus does everything I’ve wanted so far, so I don’t know why I’d change. Perhaps I could switch to Fable and ask it to review the whole code base of projects and suggest improvements or something? Claude models peer reviewing each other? I’m sure there’s a “best” way to do these things. Also feel like I’m not making the most out of my subscription.

I’m still debating if my hill planner app should do more for me or not. I like researching routes and planning them out and if it did them for me what would I do? If you can’t go hill walking then it’s fun to read about hill walking. I did wonder if there’s some half way point where it does an estimate of everything and then I refine it. Maybe useful for screening options, eg I could go out next week, here are 6 hills and routes I downloaded, make a plan for them. Then I can pick one or two that maybe suit what I’m looking for and go from there. I wonder if it’s just my imagination that’s limiting the vibe coding experience!

Having done some projects, it makes me less critical of other people’s vibe coding. Although I’m wary that projects are fleeting, as it’s so easy to make something then move on. Of course, as long as they publish the code, then I could pick it up and continue if I so desired.

I’m off work tomorrow but if I wasn’t I might not be getting much done: https://status.cloud.microsoft