I'm personally frustrated with Google as much as anyone else, but I know they are metrics driven and I wonder if, according to a certain metric, Google is nailing it just as much as they ever have been, or perhaps are even more so.
Or they've shifted strategies to optimize for things that give them a high hit rate, perhaps things that leverage their knowledge in maps, or people's desire for extremely recent news. All of the unconscious times that we use !g because we don't trust DuckDuckGo enough to get the right answer.
I don't know what the right answer is here, I think the question "does Google suck" breaks down into several sub questions that get at the essence of what we mean, and certain variations of the question the answer is yes, such as niche discovery or deep dive research, credible long-term commitments to any kind of specific algorithm whatsoever that can help you grow an extinctual relationship with the product, and for certain other questions the answer is no.
So it's not even that I disagree, I just think anytime I see this feeling expressed, it's always reiterating that very vague first pass subjective unprovable version, and it's not breaking down into sub questions.
- https://news.ycombinator.com/item?id=29772136 / Google no longer producing high quality search results in significant categories, 1290 comments, Jan 2022
- https://news.ycombinator.com/item?id=29392702 / Ask HN: Has Google search become quantitatively worse?, 502 comments, Nov 2021
Anyone competent enough to make an information based website worth ranking on google front page will be better off making a youtube video on the subject instead and that's what they're doing.
Maybe LLM’s auto transcribing and pinning timelines to them (like loom does) is the solve.
- Exact token searches (quoting, + prefix) are completely gone.
- Doesn't work properly until JS loads, leading to false starts.
- Screws with browser history so that often I cannot hit back to go to the results again, I land on the previous page before the search, and going forward goes to the result I clicked.
- 80% of the time I can't copy or save images on mobile. (This one is especially irksome).
- Shows nearly a full page grid of shopping results I don't care about if I search for any product. If I wanted shopping results I would use the shopping tab!
- Adds :~:#text= to URLs I click into, making sharing links I search up obnoxious.
It just goes on and on. These days I use Kagi, despite the friction/flakiness of the plugin on iOS.
Edit: looks like the newer 2.0 plugin just got updated a couple days ago to supposedly fix the lack of redirects. Seems to work so far! And now it redirects on any search engine site you allow the extension on, and doesn't on ones you don't, which is a fair bit nicer if a bit confusing to configure.
I have also had the problem with my browser history and always blamed my device.
Maybe I was giving Google too much benefit-of-the-doubt.
I have many times searched for an exact phrase I know is in a result and could not find it on the first two pages of results.
I suspect it's a combination of Google prioritizing more and more revenue-maximizing changes (like inserting a distracting, space-hogging YouTube video section into many results pages) plus so much content now being SEO optimized (by both human authors and AI generators) to the point of being nearly useless.
I now have a Userscript that runs on Google Search pages that automatically makes over a dozen conditional changes to both my queries and the URL parameters, like adding "-site:youtube.com" to searches not on the Videos tab. I also have a couple of "Fix Google Search"-focused Firefox add-ons installed as well as at least half a dozen uBlock Origin Google Search-specific rules to hide various annoyances. Combined all this makes GSearch still sort of useful but I can tell it's a losing battle and I now often use either DuckDuckGo or Bing .
Saw this the other day https://bsky.app/profile/drewtoothpaste.bsky.social/post/3kk...
These kinds of posts are all just hearsay and based on a billion variables of people's search history etc that Google is trying to adapt to.
This isn't like an SQL query where a weird result 1 in 1 million times would be an indication of a bug.
Reasons mentioned: SEO is winning, being an advertising company is contradictory to being a good search engine, people and content moving into big silo sites some of which are deep web, greater ease of just producing a giant amount of autogenerated sludge and sites and Google trying to be too smart.
https://twitter.com/lillybilly299/status/1756714094596358351
Try kagi. I was skeptical at first about the idea of a paid search engine, but my experience has been they are mostly simply better than google for all searches related to software development, and at worst, equal.
This breaks or at least alters how google ranks sites. It becomes unusable.
Someone needs to look at search from first principles. Everything from scratch.
I'm not sure I agree with that conclusion, but I do believe it represents a forward step from the status quo, where machine produced content dominates but is merely in the C to C+ range for quality.
Seriously I’m rennovating a house and the top of the google page usually tells me whether Home Depot or Lowes sells what I need in seconds.
Millions of people are using google every day to search for other things and getting results. I use it constantly without issue for a number of daily topic searches.
For example try searching "best coffee grinder" and "best coffee grinder Reddit", the quality of results (and arguments) is night and day. This is true for any "best x Reddit" query.
We don’t specifically want Reddit results, we want results that are not obvious SEO/AI spam. Ideally I’d like the ability to add ":humanblog" to my queries to get only blog results, but it’s not possible, so "Reddit" is the best option left.
The deal is publishers (including stackoverflow, reddit, etc etc) let Google scrape their pages because they send traffic. That's the trade.
If the trade, instead, is google scrapes your pages, uses that as AI training data to give "searchers" answers, and then and never sends you traffic, site owners will block Google's scrapers. And just like that, all the training data disappears.
Google could afford to pay publishers, but that also opens the doors to governments the world over demanding payment for scraping and links, so... yeah. And also, then, what is their advantage vs OpenAI, and how does Google remain ungodly profitable if their tac jumps thousands of times, etc? Tough spot.
There are some queries that have useless results by nature of being useless questions. "Best shovel" etc.
That's when Google started to go downhill quick.
On a different note: yes, google results have degraded very noticeably in the past 2-3 years, to the point where you are better off using Yandex (except for politics and contemporary history - for these look for some other source...)