Stuck on a DSA problem for 30 minutes: brute force first, or keep thinking?

Some people say write the brute force right away so you have something working. Others say that anchors you on the wrong idea. Where do you draw the line between productive struggle and spinning your wheels?