People never cared about delivering Alices; they were an implementation detail. I think the article argues that they're still an important one, but one that isn't produced automatically anymore

The article is talking about science research in the context of astrophysics, not coding sweatshops.

I was also talking about producing researchers for academia.