Hacker News Reader

> If you're capped at browsing the first 10 pages or whatever, that's entirely reasonable.

Not to me. I’ve tried a bunch of times to look for obscure shit I’ve seen before (and know exists) but have bumped up against the limit. This is especially annoying when the thing I’m looking for vaguely sounds like a more popular topic, and so the first 10 pages are just about the more popular version.

If you can't find it after 10 pages (100 items) you likely couldn't find it even after 100 (1,000 items). It's diminishing returns, so more pages isn't the solution.

The solution to avoid the more popular version as much as possible is to exclude keywords associated with it, and/or to add required keywords associated only with the thing you're looking for. Exact string matches ("go programming language" rather than "go") help too.

Any kind of news event can be repeated by dozens of articles. It’s a really absurd system and it also means you can’t easily find alternative perspectives on events.

Searching by date range doesn’t work either since it brings up current articles as well.

For many search terms the first few pages are nothing but advertisement and sites gaming Google search. I'd love to go beyond page 10. Not with Google, even if says "100k results found". It just doesn't bring them money to link to those pages

Pre-Google, it was pretty common to have to dig several pages into the results from early search engines to find what you were looking for. Seems like we’re just returning to the bad old days.

Even earlier google had lots of useful results on the 5th page. Google was better at finding me what I was looking for when they were a search company breaking into the ad space. Now they are an ad company with a vestigial search segment who's only modern purpose is to be a page to serve ads on.

> I can't even tell what this blog post is complaining about

They’re not complaining, they’re advertising their own (paid) service (which serves Google results programatically) while at the same times being able to point to customers why they’re getting fewer results than a regular Google search (they aren’t, because Google’s number doesn’t reflect what you can look at).

I agree that it's very poorly written. The heading is especially confusing, it took me a long time to understand that "SerpAPI" was the search term, not a Google service. It is also the name of the company he's promoting, which makes it even more confusing.

But I don't agree that it's "obvious" that "you're not going to be able to browse to page 187,398,384 to get those results. I would argue that a search result is only a search result if you can actually view it. If not, it's just a marketing, or statistics.

Feels like SEO blogspam trying to capitalize on Google's currently negative public profile after the Stadia cancellation, and general negative google sentiment.

Really easy to get clicks (and upvotes on HN apparently) by complaining about literally anything Google related.

Maybe but it's from 2021.

Personally I'd prefer something like "Returned top N of 166,000 results" and easy access to those N results (first/best match, last/worst match, paginate left/right)

why do you think this? nothing in the article indicates this, and Google's questionable search results (compared to what they once were) has been a nearly constant topic of discussion for years now

> Google's questionable search results

This is exactly my point. This article spends a thousand words to say "google doesn't actually let you see every result from the billions it claims to have."

Okay cool, I'm not gonna read a billion search results anyways, that's why I ask google about a topic.

The writing is poor, the "findings" mundane, it's a marketing fluff piece that tries to convince you there's a problem and then plug their service as a means to solve it.

I’m not conspiratorial about this, but I’m confident that an article like “Google doesn’t actually serve you all the results” drives more clicks, engagement, and outrage than “literally no sufficiently large search tool gives you all the results, including Bing, DuckDuckGo, Wikipedia, GitHub, the Library of Congress, etc.”

I don't think that's obvious at all. This first came up a while ago here I think. Even after having the time to think about it, it's still not obvious.

The search page for coffee says "Page 1 of about 3,600,000,000 results"

How are we to know that we can't load page 360,000,000? Maybe it's obvious if you're familiar with search engine internals/algorithms, which the vast majority of people won't be.

> I can't even tell what this blog post is complaining about, it seems so badly organized and written.

That's because it's an ad for the service, not a real blog post.

Most people never go past page 1 of the results.