The reason for that is because Google's building for a mainstream audience, because the mainstream (by definition) is much bigger than any niche. They increase aggregate happiness (though not
your specific happiness) a lot more by doing so.
It's probably possible to build a search engine for a specific vertical that's better than Google. However, you face a few really big problems that make this not worthwhile:
1) Speaking from experience, it's very difficult to define what "better" means when you don't have exemplars of what queries are likely and what the results should be. The reason search engines are a product is that they let us find things we didn't know existed before; if we don't know they exist, how can we tweak the search engine to return them?
2) People go to a search engine because it has the answers for their question, no matter what their question is. If you had a specific search engine for games, and another for celebrities, and another for flights, and another for hotels, and another for books, and another for power tools, and another for current events, and another for technical documentation, and another for punditry, and another for history, and another to settle arguments on the Internet, then pretty soon you'd need a search engine to find the appropriate search engine. We call this "Google", and as a consumer, it's really convenient if they just give us the answer directly rather than directing us to another search engine where we need to refine our query again.
3) Google makes basically 80% of their revenue from searches for commercial products or services (insurance, lawyers, therapists, SaaS, flowers, etc.) The remainder is split between AdSense, Cloud, Android, Google Play, GFiber, YouTube, DoubleClick, etc. (may be a bit higher now). Many queries don't even run ads at all - when was the last time you saw an ad on a technical programming query, or a navigational query like [facebook login]? All of these are cross-subsidized by the commercial queries, because there's a benefit to Google from it being the one place you go to look for answers. If you build a niche site just to give good answers to programming queries or celebrity searches or current events, there's no business model there.