Hi there! Search Quality engineer here (though I don’t actually work on the core ranking myself, know very little about the specifics of its evolution over the years, and am nowhere near the news team). Every change we launch to search (and there are a lot) has very stringent requirements to meet. We have to, first and foremost, improve things for our users. That means a radical focus on improving the relevance, quality, and accuracy with which Google responds to your query. We measure this in a lot of ways. We relentlessly evaluate every change based on live traffic experiments (in which our metrics continuously evolve alongside our understanding of how to measure a good experience for our users) and side-by-side comparisons of queries with and without the change by real humans (trained to evaluate things with an eye towards the ideal results page). In every way we can measure, our search engine is the best it’s ever been. We definitely aren’t perfect; edge cases at scale are a never-ceasing job, and core improvements to the experience of searching on google always show that there’s plenty of headroom to improve what we return to users. But I heartily push back against both of these statements:
> Google knows its search engine isn't very good anymore!
And
> That is why they're increasingly trying to pitch their service as a > recommendation engine with special AI sauce, not an index of the web
We provide some content discovery tools (Discover, News), but those are offshoots made possible by our core mission and competencies: organizing the web’s information and making it accessible to users (yes we really say that, though frequently with a wink and a nod). We don’t try to be an index of the web, but instead as the best place to find out anything.
For a high-level read (and some in-depth coverage, like the rater guidelines) on what I’m taking about, consult this page: https://www.google.com/search/howsearchworks/mission/web-use...