#AI #search engines cite incorrect sources at an alarming 60% rate, study says - https://arstechnica.com/ai/2025/03/ai-search-engines-give-incorrect-answers-at-an-alarming-60-rate-study-says/ yes, they really are getting worse...

Whoever decided to replace search engines with ML was either an idiot or a malicious demon #MLsec #ML #AI #search

https://arstechnica.com/ai/2025/03/ai-search-engines-give-incorrect-answers-at-an-alarming-60-rate-study-says/?utm_brand=arstechnica&utm_social-type=owned&utm_source=mastodon&utm_medium=social

Good grief. This is far, far worse than I thought possible. On average, #AI #search assistants were confidently incorrect on > 60% of queries. Especially pleased Grok 3 was incorrect on > 94% of queries! All tip of the iceberg stuff.

AI agents delivered "confident presentations of incorrect information, misleading attributions to syndicated content, and inconsistent information retrieval practices."

AI Search Has A #Citation Problem - Columbia #Journalism Review
https://www.cjr.org/tow_center/we-compared-eight-ai-search-engines-theyre-all-bad-at-citing-news.php

We are committed to surfacing content from the Small Web in our search results, which makes Kagi uniquely different to any other search engine out there.

It is not only part of our mission to humanize the web but we genuinely feel that this improves the quality of search results and discoverability of high quality, relevant content (created by humans!)

How many years I needed to learn how to search my own posts. It's too easy. Type into the #search:

from:@username searchwords

It also works (for yourself) if you have disabled being in the public search.

Switched to #vivaldi from #firefox as part of my #buyEuropean effort.
Has been pleased to see email client there, because it was a big deal for me what to use instead of thunderbird.
Next step will be to pick a new #search engine to move from kagi.
What are EU (or at least UK) options you can suggest? I know about that website for EU alternatives. But wanted to see some personal experiences.

What is everyone using for search that is not google? Tried a couple but not found a good alternative so far #search

it's nice to see reporting that doesn't mince words.

"Google’s featured snippet is pulling in an Amazon AI summary of Adolf Hitler’s Nazi manifesto Mein Kampf that calls it “a true work of art” in the latest AI-related fuckup affecting top search results."

https://www.404media.co/google-amazon-ai-search-mein-kampf-reviews/

Having ditched my subscriptions to NYT (long ago), WaPo and the LA Times, I'm glad to be a paid subscriber to the journalism from places like @404mediaco.

#Algorithmen, die sich selbst auffressen: #Google und #Amazon AI sagen, Hitlers „Mein Kampf“ sei „ein wahres Kunstwerk“. Dabei greift Googles Featured Snippet auf eine Amazon-KI-Zusammenfassung von Adolf Hitlers Nazi-Manifest „Mein Kampf“ zurück. Das ist mega-verf*ckter KI-Mist, der sich leider auch auf die Top-Suchergebnisse auswirkt...

#ki #ai #mist #search #searchengine

https://www.404media.co/google-amazon-ai-search-mein-kampf-reviews/

https://opensearchfoundation.org/en/events-osf/ossym25/ An alternative to Google and Microsoft search engines with the openness in mind is having a conference in Helsinki from 8-10.10 with free online participation.
#Finland #Helsinki #OpenSearch #Search #Degoogle

The new version of SearchTweaks.com is now live!

SearchTweaks is a set of sixteen web-based tools to make your #search experience better, from building topical #Google queries using #Wikipedia data, to easier ways to do date-based searching, to better local #news search. Check it out!

SearchTweaks.com is free, has no ads, and is designed for desktop.

https://searchtweaks.com/index.html

“In this video we showcase the main differences between #Kagi and #Google, and how we are trying to humanize the web by building a premium #search engine without ads, without distractions and without #tracking your data.”

https://www.youtube.com/watch?v=7ZLOBHAtjxk

Looking to get away from Big Tech? Here are some European alternatives you should check out!

@StartpageSearch – Search engine from the Netherlands
@Qwant – Search engine from France
#Ecosia – Search engine from Germany
@protonprivacy – VPN & Cloud storage from Switzerland
#OctoVPN – VPN from Norway
#Cryptee – Cloud storage, based in Estonia
#Jottacloud – Cloud storage from Norway

… And we got you covered with a browser.

Turns out you can circumvent AI search engine answers by using cuss words. Type any query that produces an AI answer. Add a cuss word and submit it again. No AI and mostly the same results.

The more you know.

#Disenshittify, NOW!! We have a chance: #IP!!

"…thanks to internal memos published during last year's #monopoly trial against #Google, we know what they did. They made #search worse. They reduced the system's accuracy it so you had to search twice or more to get to the answer, thus doubling the number of queries, and doubling the number of #ads.

Meanwhile, Google entered into a secret, illegal collusive arrangement with #Facebook, codenamed #JediBlue, to rig the ad market, fixing prices so advertisers paid more and publishers got less." @pluralistic
#adtech
https://pluralistic.net/2025/02/26/ursula-franklin/

Ads harm the web: they slow it down, invade privacy, manipulate behavior, and even spread malware (malvertising).

Removing ads from the web is the way forward.

Kagi will actively down-rank sites with lots of ads and trackers in the results and promote sites with little or no advertising.

Today's winner at the Search Engine Olympics was in the event "Long string exact matching, first online within the last 48h"

and the winner is....

https://www.startpage.com/

...which the only site to return two hits today. Runners up did find one result.

Living in a country where 'gender' is a forbidden word? Lost some of the Web? Make that content findable again and help us develop the #queer instance of our #decentralized #Web #search engine. Head over to https://queer-pears.org/indexer/suggest and suggest content you want to preserve. #Archive links welcome.

PeARS is a #curated search engine. Each instance is dedicated to a particular topic and we encourage human-vetted content. https://queer-pears.org is focused on informing the #LGBTQI+ community.

#Mojeek #Search. Tried a couple of years back and struggled. They deliberately use a basic keyword search rather than second-guessing the user with semantic search. In theory I should like that, but couldn't get on with it. Last year they put in some *basic* semantics - typo correction, etc - and it feels much more usable. Trying it as a daily driver and probably falling back to another search engine 30% of the time (for slightly niche work stuff). https://www.mojeek.com
#Privacy #uk