Hacker News Reader

Comment by hwers | Hacker News Reader

I almost feel as if they have like 100k actual pages that they present (and have looked through manually) and if it's not in that group they just show you the closest one (or say "no results").

View on HN

I have no knowledge or evidence, but this has indeed become my mental model for whatever Google actually does these days. I do wonder what they actually do. All those engineers, for 20 years. Surely they haven't just been scaling up BackRub! But I find it very hard to believe that they're crawling the whole web. I find it very hard to believe that PageRank is still the same PageRank that we understood in 1998. And it looks like they're managing to editorialize quite heavily, even if they're doing it via algorithm somehow. But again, I can't really discern what they're doing anymore.

So for now, I have to disagree with the "garbage in, garbage out" theory. I don't believe Google has the same goal now that they did then.

> I find it very hard to believe that PageRank is still the same PageRank that we understood in 1998.

It's not, because as good as PageRank was, it was vulnerable to being exploited by link farms, which started popping up in its wake. I do remember that by the mid 2000s, about 5% of search results were pages just spamming search keywords and hyperlinks.

Do they say "no results" anymore? It seems like Google ignores parts of your query until you get results, no matter whether they are specifically connected to your search.

Yes, they do.

I frequently research historical topics on India, and I get no results or single digit number of results.

Reply on news.ycombinator.com