I’ve started relying more on AI-powered tools like Perplexity for many of my search use-cases for this very fact - all results basically warrant a pre-filtering to be useful.
Unfortunately the spam arms race has destroyed any chance of search going back to the good ole days. SEO and AI content farms means we’ll need a whole new system to categorize webpages, as well as filter out human sounding but low effort spam.
Point being, it’s no longer enough to find a page that’s relevant to the topic, it has to be relevant and actually deliver information, which currently the only feasible tech that can differentiate those is LLMs.
It would be interesting tho to use a LLM to spot AI/SEO crap and add whole domains to a search blacklist. In that case we wouldn’t need AI to do the actual search, and this could easily just be a database for end users by the SE’s side (kinda like explicit content filters).
I’d call that option “Bullspam filter” and leave it on “moderate” by default.
This is one solution to the issue, and it seems silly you are being downvoted for it.
Google became what it became, and years of seo optimisation cat & mouse play has reached new heights. Those obviously target Google instead of their competitors for now.
Would that we could have perfect search results, it would be beneficial to google as well.
Skill issue. Old version docs tend to offer you a redirect to more recent docs, and even then something sintactic like an “IN” operator is unlikely to change in form or structure between versions of a database engine.
Old version docs tend to offer you a redirect to more recent docs
Sadly, the docs, I’ve worked with (openstack and ansible) frequently, don’t do this. They have a button to go to the latest version of the docs, but not to the equivalent page on the latest version. This means I have to find the equivalent page again, from the integrated search usually.
And yes, a lot can change between versions. New features can get added that solve your problems or older stuff can get removed.
Been there. Done that. FML on searching for programming help some days. Versioning is a nightmare as the way you “used” to do things is no longer relevant and the rest of the results are some asshole saying it is a duplicate question that was answered 10 years ago…that is no longer fucking relevant!
Sorry. Yesterday sucked. I hope today is less frustration and more things working like they are supposed to.
As someone who is trying to teach themselves a few new things this year by diving to projects using them… I seriously, seriously feel you. It honestly makes me question whether I should just abandon each project I start, both professional and personal.
All the relevant hits are from years and/or 2+ versions of whatever ago or forum posts with dead links to an alleged solution.
I feel like in the past I could just dive into something and search my way through it. Now I feel like that era is over and I question whether it’s me, my niche project idea, the disappearing community, or just the search engines.
Oh, that stuff happens all the time. The one that really pissed me off was Microsoft 404-ing basically their entire KB system.
That thing was standing for so long you could still find Windows 9x stuff on it, and it was glorious.
Around the time they stopped supporting windows 7, they bricked the entire thing up and started a new system. Overnight, all the Microsoft help article links went dead. Find a good forum post about an issue that you’re having and someone replied with a link to the MS KB saying little more than “this should work” followed by a sea of commenters saying thanks, that worked, but when you follow the link, it goes nowhere.
And right before they did that, they started removing footnotes from KB articles that only dealt with older OSes, so if you ever needed to go back and find something, it just wasn’t there anymore. For example certain RGB packing formats were only supported on newer OSes and the footnote used to tell you that, but then it disappeared. I have been directly affected by that multiple times.
I wonder if the internet archive has a full copy of the KBs from before Microsoft dumped them. I’d love to set those up in a web server so I can reference them as needed.
Some of this is just because some of these frameworks and technologies have been around for a while and they iterate frequently. I see a ton of Azure content that is obsolete after only a few years.
It doesn’t matter. List all the crap you want, but show me the most up to date official documentation for the postgres “IN” operator in the very first result! It can’t be that hard.
Doesn’t mean the statement is less true, the enshitification of google is a symptom, the disease is the internet as a whole. Google and LLMs screwing the web, M$ screwing windows, Apple’s existence by itself, Meta monopolizing and screwing social media, and don’t get me started with streaming platforms and other media industries are all symtoms.
Considering all of that, yes, the internet enshitification is very real.
But anyway, the cool thing about the internet is that you can find your nice cozy niche and stay there.
That’s how the 90s internet was. If the megacorps want to be in here, fine. I’ll just stay in Lemmy. And when Lemmy starts sucking, I’ll move to somewhere else.
This is why I’ve really grown attached to Kagi (paid search engine).
It’s made the internet usable again. I’m honestly surprised how much of difference there is. I’d really recommend people give it a shot. (there’s a free trial for it)
I generally agree with that, but as an aggregation service it would need to justify not providing any actual content/information with its price structure. The same argument against AI models trained with user data.
For now, ignore my recommendation, as I don’t yet fully know my stance on this, with the information provided.
However, I can say that I’ve been super happy with the search results. I don’t use their email service. Just the search and the access to all of the LLMs that are out there.
I don’t know what shady shit you’re referring to. They do AI, but I don’t use any of that. IMO their core strength is the search engine and how it works for you rather than against.
Though if I would use postgresql documentation very often I could just use the Kagi feature that rewrites URLs with a regex, so I can replace it always with the latest version.
Interesting, my Kagi results gave W3Schools, geeks for geeks, and postgresqltutorial.com before the official docs, but hey still way better than OP’s results!
Kagi has search personalization where you can lower/raise/pin specific domains (one of kagis main selling points) and I blocked geeks for geeks and w3schools, as these are irrelevant for me and I don’t want them in my results
I don’t think that’s possible with searxng (but I’m not 100% sure, but I can’t seem to find that feature)
I know there are browser extensions which can filter out domains in search results for different search engines like google and duckduckgo.
But the pinning/lowering/raising is a bit trickier to implement as an extension, because what kagi does is basically:
Load 3 pages of search results in the backend
Show a result as the first entry if it matches a rule for pinning
Influence the search ranking algorithm with the lower/raise rules of the user
Filter out blocked domains
It would be possible but not as “streamlined” as Kagi does.
Don’t get me wrong, Kagi definitely has its rough edges and the search ranking algorithm is sometimes very unpredictable, but it provides good enough results for me to be worth the 10$ per month for unlimited searches.
if Kagi were open source sure, but it’s $10 a month and the CEO it is kind of an asshole. And a generative-AI-bro (please don’t make me call them GAI-bros)
sopuli.xyz
Active