Talking to AI vs. letting it figure it out

I’ve noticed something interesting after 10+ hours intensively using Mocha. I was struggling with trying to make a sticky side navigation work. I tried a bunch of different ways to explain how it should function but it kept on bugging. Then I simply said “audit the navigation and fix whatever you think needs to be fixed”. Granted, I had to do it twice, but it finally worked.

Sticky side navigation

So there seems to be a spectrum: knowing how to “talk” to AI is important to get good results, but sometimes, just letting it figure it out works best.

When you think about it, it makes sense for that specific use case: I was trying to explain in my own words how a sticky side navigation should work, but I dare you to try — it’s much more difficult than it sounds! When actually, this behavior is spread out across the internet so I did not need to reinvent the wheel. Just telling the AI to apply best practices worked better.