Rather than asking “what should I build?”, I'm curious how people systematically discover gaps in the industry.
What's your research workflow?
Where do you look for recurring pain points, bad existing solutions, missing tooling, or internal scripts that perhaps should be proper open-source projects?
And once you find a possible problem, how do you verify that:
there isn't already a good tool for it; the pain is common enough; and developers would actually switch to/use a new solution?