Growing amount of unmaintained duplicate code and digital waste

What are we going to do about this?
Any x y search on GitHub, Android: Permissions for example will yield hundred of results.
We currently have limited means of sorting this information. by star, commit history etc. If we do not implement an objective ranking system for repositories that utilizes both community engagement and static analysis then future developers are going to be left as we are trying to find gems in a massive digital waste storage site.
“In other words, 70% of the code on GitHub consists of clones of previously created files”.
-DéjàVu: A Map of Code Duplicates on GitHub

GitHub Topics is the way to go:

usually forked repositories don’t add topics, so the chances of finding the “source”/upstream repository using topics are higher.

Unfortunately, you can only navigate one topic at the time, but if more users start using topics smartly, it will be easier to pin-point projects according to their categories.

Also, you can always filter results by programming language, which drastically reduces results. There also some other filters that can be added to the results URLs, which might not always be accessible via the WebUI, but which are documented on GitHub Help.

In any case, you can’t have freedom without chaos — the former entails the latter. So there’s no pragmatic solution to this age-old problem.