If its the former, people are wasting their own time, money and energy when spending a similar amount of time searching vs prompting might yield better results for their use case. If its the latter, unless there is a killer feature, which I dont see in the github repo, why not point people to use the actively maintained project instead of the one the vibe coder is going to let die on the vine once they get bored. Its fracturing the space, for what gain? (no pun intended)
Also how much of this code is the LLM lifting from Stellarium either indirectly in the training data or actively searching for it to reference the code in its context?