I would argue Astral's products are in a substantively different category (developer tools) than numpy or scipy (numerical libraries), and the constraints on them are significantly different (e.g. like crypto code, you shouldn't roll your own numerical code, and instead use trustworthy libraries, which is what numpy/scipy wrap). Everything from their choice of language to which libraries to depend on (and which ones to make optional) to their installation system has been about being conservative and ensuring that the bootstrap process is as painless as possible for Python-only devs.
Astral to me seem to be targeting a small subset of the community, and their ignorance of the needs outside of it make me very wary of adopting their tools.
Astral to me seem to be targeting a small subset of the community, and their ignorance of the needs outside of it make me very wary of adopting their tools.