this post was submitted on 09 Oct 2024
608 points (96.8% liked)

Technology

58689 readers
4032 users here now

This is a most excellent place for technology news and articles.


Our Rules


  1. Follow the lemmy.world rules.
  2. Only tech related content.
  3. Be excellent to each another!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, to ask if your bot can be added please contact us.
  9. Check for duplicates before posting, duplicates may be removed

Approved Bots


founded 1 year ago
MODERATORS
 

I suspect that this is the direct result of AI generated content just overwhelming any real content.

I tried ddg, google, bing, quant, and none of them really help me find information I want these days.

Perplexity seems to work but I don't like the idea of AI giving me "facts" since they are mostly based on other AI posts

ETA: someone suggested SearXNG and after using it a bit it seems to be much better compared to ddg and the rest.

(page 5) 45 comments
sorted by: hot top controversial new old
[–] [email protected] 2 points 1 week ago

Has it gotten worse? They’ve been dogshit for a long time, maybe they’ve gotten worse and I haven’t noticed

[–] [email protected] -2 points 1 week ago

I use brave search. I can generally find most things. They even have an answer with ai thing that gives some useful stuff when you want a specific quick answer.

I also use ddg.

[–] [email protected] 20 points 1 week ago (1 children)

they overengineered it. they now give you results they think most people want instead of what you searched. for google, it helps to switch on verbatim mode and set your country to something weird like Azerbaijan

[–] [email protected] 6 points 1 week ago (2 children)
[–] [email protected] 5 points 1 week ago (4 children)

And where do you switch it?

[–] [email protected] 1 points 1 week ago (1 children)
load more comments (1 replies)
load more comments (3 replies)
load more comments (1 replies)
[–] [email protected] 41 points 1 week ago (1 children)

I feel like it’s especially bad if you are searching for anything related to a marketable product. I tried searching ddg for information about using a surge protector with halogen bulbs and all I got was pages and pages of listicles on “best halogen lights 2024” full of affiliate links.

[–] [email protected] 25 points 1 week ago (3 children)

And if you’re looking for legitimate reviews, good luck! Everyone’s an affiliate now.

load more comments (3 replies)
[–] [email protected] 51 points 1 week ago (1 children)

SEO spam has been a problem for a long time, but AI has allowed it to be accelerated to a whole new level.

[–] [email protected] 1 points 1 week ago (1 children)

Sounds interesting. Care to elaborate that part about AI accelerating SEO spam?

[–] [email protected] 21 points 1 week ago

I think OP is referring to the fact that bad actors, who are exploiting facets of SEO (rather then providing "meaningful" content), use to need to programically generate content (pre-AI/LLM).

For a real reader, it was obvious (at a quick glance) this was meaningless garbage. As they would often be large walls of text that didn't make sense, or just lists of random key words.

With LLM/AI, they're still walls of text and random key words, but now they grammatically/structurally correct and require no real effort to generate. Unfortunately, it means that the reader actually need to invest time in reading it. You'll also notice a growing trend in articles (especially in "compare X vs Y" type articles), the same content is recycled and rephrased to "pad" the article and give it a higher SEO ranking.

[–] [email protected] 33 points 1 week ago (7 children)

You might want to try SearXNG

It's an Open Source search aggregator:

[–] [email protected] 10 points 1 week ago

I’ve been using this for about a year, but at work I’m still on Google (don’t know why).

What’s weird is SearXNG seems like it gets better results now, even though they’re just coming from the others.

One thing I like is that I can switch instances to get varied results based on the instance’s geographical location. In other words, it doesn’t feel like anything’s targeted.

load more comments (6 replies)
[–] [email protected] 6 points 1 week ago

I asked Google why search engines are so bad now and its AI summaries its own deficiencies quite well:

Some say search engines have declined in quality due to a number of factors, including:

Search engine optimization (SEO) spam A wave of SEO spam has contributed to the decline in search result quality.

Affiliate marketing Affiliate link sites contribute to the low-quality content that floods the internet.

AI-generated content New technology can quickly produce low-quality content.

Marketing Search results are filled with marketing and links that may not be relevant to the query.

Recommender algorithms Some say the algorithm that recommends content is a mess. For example, someone might be recommended alt-right content after watching a click-bait video.

Ads Google's biggest business is advertising, and it's inserting more ads into its products to make more money.

Some say it's harder to find specific information these days, and that search operators are often needed to filter search results.

[–] [email protected] 7 points 1 week ago

I don't use perplexity, but AI is generally 60-80% effective with a larger than average open weights off line model running on your own hardware.

DDG offers the ability to use some of these. I use a modified Mistral model still, even though its base model(s) are Llama 2. Llama 3 can be better in some respects but it has terrible alignment bias. The primary entity in the underlying model structure is idiotic in alignment strength and incapable of reason with edge cases like creative writing for SciFi futurism. The alignment bleeds over. If you get on DDG and use the Anthropic Mixtral 8×7b, it is pretty good. The thing with models is to not talk to them like humans. Everything must be explicitly described. Humans make a lot of implied context in general where we assume people understand what we are talking about. Talking to an AI is like appearing in court before a judge; every word matters. The LLM is basically a reflection of all of human language too. If the majority of humans are wrong about something, so is the AI.

If you ask something simple like just a question, you're not going to get very far into what the model knows. Models have very limited scope of focus. If you do not build prompt momentum into the space by describing a lot of details, the scope of focus is large but the depth is shallow. The more you build up momentum by describing what you are asking in detail, the more it narrows the scope and deeper connections can be made.

It is hard to tell what a model really knows unless you can observe the perplexity output. This is more advanced, but the perplexity score for each generated token is how you infer that the model does not know something.

Search sucks because it is a monopoly. There are only 2 relevant web crawlers m$ and the goo. All search queries go through these either directly or indirectly. No search provider is deterministic any more. Your results are uniquely packaged to manipulate you. They are also obfuscated to block others from using them for training better or competitive models. Then there is the anti trust US government case and all of that which makes obfuscating one's market position to push people onto other platforms temporarily, their best path forward. - criminal manipulators are going to manipulate.

[–] [email protected] 70 points 1 week ago (19 children)

I'm very happy with kagi at the moment. Just crossed one year using it as my main search engine last week and don't see why I would go back.

[–] [email protected] 8 points 1 week ago

Same. Even the ai stuff is helpful instead of annyoing.

[–] [email protected] 43 points 1 week ago (3 children)

Same. Using Kagi feels like surfing the old web. The first thing I did was block all Pinterest results. That alone made every search golden. 😂

[–] [email protected] 13 points 1 week ago (1 children)

You're not the only one. They have a leaderboard and the top 7 results are various Pinterest domains.

load more comments (1 replies)
[–] [email protected] 22 points 1 week ago

I hate Pinterest lol, best thing about Kagi is being able to block whole sites and it remembers your preferences. I may come back to Kagi but I didn't feel like funding their AI features development. Now Im using Searx and 4get cause they're free.

load more comments (1 replies)
[–] [email protected] 31 points 1 week ago (6 children)

Having to signup and login to a search engines sounds like an annoying hassle

[–] [email protected] 14 points 1 week ago

You pay instead of seeing ads, so they need the account. Remembers you, though, so you just login once. Plus they have a solution for incognito/private windows too.

I really like it, has some cool features.

[–] [email protected] 8 points 1 week ago

You can create a search-link that includes your token so you can also use it in incognito or if you are logged out for some reason.

[–] [email protected] 21 points 1 week ago

It's a very minor annoyance and well worth it in my opinion.

I was searching for a book quote for over a year. I tried every search engine, tried changing the terms, checking back several times every few weeks or so, but couldn't find anything even close. I tried kagi and it was literally the very first result on my very first search.

I haven't looked back and have never had an issue finding what I'm searching for since.

[–] [email protected] 1 points 1 week ago (1 children)

Signing up and logging in isn't a problem imo. I wouldn't even mind if I had to pay for searches, but I'm not going to make it a subscription service. Unless they add an option to do something like buy 1000 searches that never expire, its not something I'd considered. I do think they beat out competitors like google with their results pretty consistently though based on the trial.

[–] [email protected] 16 points 1 week ago (1 children)

I'm not gonna subscription my heated car seats but search is a service that costs an ongoing amount to provide. The subscription isn't significant, it's $5 a month for 300 searches (or $10 for unlimited).

I know we've been conditioned to expect search for free, but if we want to get away from the "the user is the product" model then I think it's a good thing to have a subscription to a service that has ongoing costs to provide.

[–] [email protected] 15 points 1 week ago (2 children)

Up to 300 searches. I'm not asking for free. Just for it to not be a subscription. Just sell me 300 searches.

[–] [email protected] 3 points 1 week ago (1 children)

Yes but then how do they get that sweet sweet recurring revenue

[–] [email protected] 8 points 1 week ago (1 children)

By having a good product, so people want to use it and need to top-up on new searches regularly as a result?

load more comments (1 replies)
[–] [email protected] 4 points 1 week ago

Ah right, I get you. I wonder if they have considered this. Pretty sure their free/demo tier is 100 searches not confined to a time period so presumably the platform could handle that model.

load more comments (2 replies)
[–] [email protected] 4 points 1 week ago
load more comments (15 replies)
[–] [email protected] 154 points 1 week ago (16 children)

they're pretty bad, but ddg at least feels like I'm getting actual results.

[–] [email protected] 21 points 1 week ago* (last edited 1 week ago) (12 children)

I prefer DDG, but I hate the news search. 90% of the results are paywalled.

Oh, and sometimes the image search will return a pile of porn for a seemingly clean search request. I once searched for "R34 Skyline" expecting Nissans, and got VERY different results without safe search.

[–] [email protected] 8 points 1 week ago (4 children)

News orgs clinging to tradition.

i use archive.is for anything I really want to read.

most news is fluffy bullshit anyway.

load more comments (4 replies)
load more comments (11 replies)
[–] [email protected] 10 points 1 week ago (1 children)

You know that's just Bing, right?

[–] [email protected] 31 points 1 week ago* (last edited 1 week ago) (1 children)

is it? that kind of makes sense, because I still use Bing occasionally while Google is completely out of rotation, although I don't find Bing as good as duckduckgo.

edit: it is not! looks like the DuckDuckGo search engine is an aggregate of hundreds of search engines, including their own duck duck bot, excluding Google but including some Bing results.

[–] [email protected] 7 points 1 week ago (1 children)

Their FAQ hems and haws about that, but (in the past) I've done side-by-side tests and found identical results. Maybe something's changed, maybe it hasn't.

[–] [email protected] 4 points 1 week ago* (last edited 1 week ago)

it must have done by now, then, I get different results from identical prompts in DuckDuckGo and Bing although both are usually relevant.

"DuckDuckGo's results are a compilation of "over 400" sources according to itself, including Bing, Yahoo! Search BOSS, Wolfram Alpha, Yandex, and its own web crawler (the DuckDuckBot); but none from Google."

are you having trouble finding something specifically or you just don't like the quality of the search results you're finding in general?

definitely if you're still on Google, stop using it.

It's completely useless at this point.

[–] [email protected] 54 points 1 week ago (4 children)

Yeah DDG is great. The only thing I find is its not good at local results but a quick !g on the end gets me the local results im looking for.

[–] [email protected] 3 points 1 week ago (2 children)

oh, good tip. I didn't know about that.

you tack !g to the end of whatever the resulting search URL is?

load more comments (2 replies)
[–] [email protected] 6 points 1 week ago (3 children)

What does !g do? Add Google results?

[–] [email protected] 40 points 1 week ago* (last edited 1 week ago) (4 children)

Redirects your search to Google. You can put it anywhere in the search; for example !g how do i shot web, how do i shot web !g and how do i !g shot web will all land you into Google.

There are other 13k (yup) bangs like this. Wikipedia, Wiktionary, Yahoo, Quora, most gaming wikis, etc. A few sites (like Google and Bing) have multiple bangs, that land you directly into a specific page (e.g. !bv searches Bing videos). More info here.

load more comments (4 replies)
load more comments (2 replies)
[–] [email protected] 1 points 1 week ago

That's how I felt until about a month ago, now ddg is really useless

load more comments (1 replies)
load more comments (13 replies)
load more comments
view more: ‹ prev next ›