I have literally never looked at github stars as a measure of quality or had it affect my decision. I have looked at git logs, websites, issues, etc. But I would be genuinely worried if someone used github stars as an indication. So many honestly stupid projects have a lot of stars, and stellar ones have next to none.

https://github.com/EvanLi/Github-Ranking/blob/master/Top100/...

proof here. The top are taken by chinese educational repos. Elastic Search and Spring Boot are the only projects actually used by anyone in the top 10. But why would I trust the stars for spring boot over the fact its used in every java shop on the planet?

I don't rely on stars as the main signal of quality, but very low stars could stop me from looking into the things that I do use as a signal:

   - number of contributors
   - open issues
   - merged and unmerged PRs
   - commit history
   - the code
   - project governance

Some of these are also tied into GitHub rather than the git repo itself

If I am sitting in a review session and the Engineer presenting brings up their options and why they are choosing to bring in technology A over B, and I ask them what their reasoning is. Being unsatisfied by "There are a lot of Github stars" Seems like an absolutely reasonable position to me. This is the equivalent of saying that something seems more true because it has a lot of facebook likes.

GitHub Stars are not a selection mechanism. They are filtering mechanism for most people.

Imagine someone is choosing between three FOSS projects - one has 90K stars, one has 30K and the thirdone has 1K.

In your meeting, the engineer will show you how he decided between the first and the second without even mentioning the third.

Good for you champ.

Another anecdote chiming in here. I've literally never paid attention to GitHub stars for anything important. Except if a repo of a big project has few stars I double check because I'm probably looking at a fork instead of the main repo.

I came to a similar conclusion - that GitHub benefits from a network effect similar to social media. I would really like to leave GitHub, but it's where stuff is happening. Any company seriously looking to replace GitHub should pay some attention the social network aspect of it.

> Any company seriously looking to replace GitHub should pay some attention the social network aspect of it.

Indeed. And I would say it is not just social signals but even non-social authority signals.

E.g. how many other projects depend on this projects and how many downloads happen for its artifacts.

You can see some of these on package registries like npm and pypi where their authority signals help people choose between the right libraries.