Then you build a helper tool that analyses the differences between good code and bad code.

Then you set it loose on its own source - and you have an AI that takes over the world. (Or at least github, which is more or less the same thing.)

Perhaps instead of using StackOverflow "correct" answers, someone could create a site that has example of anti-patterns and code smells, which you could then use to analyze the code on Github and link to possible alternatives.

That's basically what linters do.

ESLint catches many

Have the code editor report back to the cloud the amount of times a certain snippet was "selected"/used. And just rank the drop-down box/autocomplete options accordingly.