27 comments

[ 3.3 ms ] story [ 27.7 ms ] thread
It’s still better than duckduckgo requiring a Microsoft account to see TFA!
I fail to see the link between duckduckgo and Microsoft and TFA, what connects them?

https://duckduckgo.com/?t=h_&q=google+news+ranks+ai+generate...

If you tap the NY Post link below the "Recent News" title you'll be taken to a Microsoft powered news page that only lets you read a few paragraphs before demanding you install some malware or app or whatever.
Interesting, for me that just takes me to the page + content. I do run a bunch of ad blocking stuff so maybe that's the difference?
Did you see the plain "Expand Article" button under the blue "Download App" button?
No, just "continue reading in the app".

Edit: I do see the button on other articles that are served "on MSN.com". I dunno.

Reading this article from the exclusive ny post triggered an 8-deep navigation tree into some virus scan scam, without clicking on anything.

Maybe that’s why they are outranked in search…

Did you happen to grab a screenshot? I am curious how it lines up with some malware I saw impacting news sites a few years back.
While this is a real problem, I appreciate the unintentional self-own to complain that AI trash is being ranked above the usual NY Post trash ;)
I would love to see a more "private" version of internet where I can simply get results and such from a specific set of whitelisted providers.

I have a feeling the news industry might actually be doing better now than in past.

Isn't this essentially what Kagi and other search engines that let you create "lenses" and block/allow sites are?

I think this version of the internet exists today, it just isn't built in to Google anymore.

Google used to have a feature where you could filter out whole domains for any reason. I wonder why they don't bring that back and then use that as a (non-determinative) signal for deciding which domains to ban themselves. They index a lot of sites obvs but you'd think a user-curated list of the top 100 or 1000 spammy domains would be useful info to have.
Google has never had that feature. There have been browser extensions that could do it though
LOL it was probably an n% experiment that OP got into!
There is the Programmable Search Engine feature that you could set up to do this kind of filtering.

Here's a quick example that excludes *.gov domains - https://cse.google.com/cse?cx=55d5f36c6916f4526

The visual style of the results is much more reminiscent of an older style search engine results page which I kind of like

Probably a few more zeros on that 1000 to get to the spammy domains, and including those doing seo hijacking
Imagine all the emergency meetings it would trigger at Pinterest, Quora, Business Insider and Forbes
I wonder if the ripoffs were ranked higher because they appeared newer?
(comment deleted)
I might be missing something but this article seems deceiving. When you're "sorting the results by most recent date of publication", of course more recent articles will rank higher regardless of where they're from.