Settings menus – seek and ye shall find?
I know what I want to change. Why do I have to find the right menu first?

Why is software so complex? Perhaps I only need ten of Excel’s many features. The next user, though, needs ten different ones.
For me, one reason lies in the economics of software. Once developed, it can often be distributed to more users at little extra cost. Another newspaper or another keg of beer, on the other hand, has to be produced and transported. This favours software that serves a large market with many different needs. Given the high development costs, separate solutions for every small niche were often uneconomical.
Adapting it to my everyday life then ends up in the settings menu. When I click the gear icon, I rarely know what to expect. Yet I know quite precisely what I want to change. Why can’t I enter that request right there and have it carried out?
Schedules are the example I have in mind: lighting, robot lawnmowers, irrigation, heating. Every app has its own view, and creating or changing a schedule is often tedious. I just want to write:
“Increase the temperature in the children’s rooms by 0.5 °C between 8 pm and 6 am.”
In my Tado app, that currently means opening each day, adjusting every affected time slot, changing the temperature with the dials – and repeating that for every room. Copying only helps to a limited extent because the days and rooms have different settings.
The application could instead show me which entries it would change. I review the suggestion and confirm it. All the other differences in the schedule remain intact.
Ideally, the model would run directly on my device: local inference. In principle, this is already possible. For my use case, what matters is that a model is small, fast and reliable enough on the hardware in question. More about local AI in the browser.
One message for a change that currently takes me through lots of individual settings. That would make my everyday life much simpler.
Taking the idea further: UI on Demand.