ɗ𐐩ʃƕρʋMany people think that a software project is code. They would say things like "forking a project is easy, you just make a copy of the repository", or "this clean-room re-implementation is a new project". But a project, whether commercial or open source, consists of people. The code is only a side effect, often necessary to let people from outside of the project make use of it, but not strictly required. There are software projects that produced no code, and ones that produced multiple codebases.
If you think about the project as the people, then it becomes rather obvious that drive-by automatically generated patches don't actually contribute to the project. They are just so much noise. The people who generated and submitted them are not part of the project and usually don't intend to become part of the project. They didn't consult the other project members about their patches and they won't answer questions about them. They are like ephemeral ghosts, an illusion of contribution.
Kale