Not sure if it is an instance of Goodhart's law or just increased exposure to non-developers crowds, but Github stars used to be a good indicator of the popularity AND adoption of an open source project.
Nowadays, a demo project gets 89k stars in few days... https://github.com/Significant-Gravitas/Auto-GPT
What do you think?
GitHub stars were always pointless.
Some people get lucky.
Think it’s an outlier
At this stage of LLM-based AI, you can consign everything to the "a demo project" bin, because adoption is in no way at par with more traditional tools.
Meaning, the area is skewed towards popularity/interest, and for that, GitHub Stars remain a good-enough indicator, just as upvotes on GH issues, and posts on certain sites are.