Google employs thousands of researchers who would be less productive (and upset) if they scrapped it. That alone is probably enough to make it worthwhile to keep it going, at least until a good alternative emerges.

Given that they have killed products with millions of users, including a lot of paying users, relying on this is optimistic. Google doesn't seem to care about major inconvenience they cause, like with the Google Domains sale Squarespace.

I think the point was that Google is _sometimes_ willing to support projects if it helps their employees do their job, which might be the case here.

Well, given how Google behaves I am not sure my perception of rationality aligns with their perception of rationality

Google employs a lot of academics that probably use it. And of course they have a few AI related products that are probably being trained on scientific content as well. I bet Google Scholar feeds data into that effort. My guess is that keeping google scholar up and running isn't breaking the bank for them and it is actually a valuable resource for them.

Semantic scholar is pretty good so I keep using it more and more.

Like annas archive?

Well, at least Google Scholar is aligned with Google's core business: search. It seems silly for Google to scrap search features. On the other hand, I'm not sure if Google Scholar is aligned with their real core business: ads.

IMO pubmed is superior for life sciences, especially if you use their entrez direct. Really powerful query tooling.