A crawler wanted to know who talks to whom on our server (www.martinmuc.de)

submitted 5 days ago by JensSpahnpasta@feddit.org to c/fediverse@lemmy.world

13 comments fedilink hide all child comments

you are viewing a single comment's thread
view the rest of the comments

[-] tiptoes@sh.itjust.works 4 points 5 days ago

Fascinating data presented in a clearly AI written or at least heavily edited format that makes it really hard to read without massive skepticism.

[-] M1k3y@discuss.tchncs.de 9 points 5 days ago* (last edited 5 days ago)

Probably AI translated. troet.social is a german instance, there's a link to the original text at the top.

Edit: The German text also has some AI-like features.

[-] mountain@lemmy.zip 4 points 5 days ago

+1, I have actual trouble reading stuff like this, even though the content is interesting. My eyes just glaze over. I would rather read someones unfiltered low effort garbled mess of a post than a polished article sounding like ai

[-] JensSpahnpasta@feddit.org 1 points 5 days ago

Please provide some evidence that this is AI written.

[-] tiptoes@sh.itjust.works 9 points 5 days ago

The phrasing and pacing sound repeatedly “but wait there’s more” style that AI seems to favor. Always trying to re-emphasize a point already made to make the reader interpret it a specific way.

Not to mention the em-dashes.

Sometimes good writing sounds like that too, but good writers also know to not use and repeat that same style over every paragraph.

[-] backpedal@fedinsfw.app 3 points 5 days ago

Not posts. Relationships.

One paragraph in and the DiGiorno pattern (it's not x, it's y) is an immediate red flag.

I'm not saying humans don't do this, but it's overwhelmingly an AI thing now.

[-] CovertOperative@piefed.zip 2 points 4 days ago

Notably, AI does it all the time when there isn't any good reason to highlight a contrast, like here.

[-] XLE@piefed.social 1 points 4 days ago

https://www.pangram.com/history/f5763e48-bee9-4e17-b86c-6b65f8acd175?ucc=uqYDBGYZEF1

[-] JensSpahnpasta@feddit.org 1 points 3 days ago

Dude, AI detectors do not work and those idiots running them are increasing the harm done by AI, because now people and students have to defend themselves against AI-usage allegations. Your tool even says "confidence limited" and "We believe that this text is 100% AI", which is really idiotic.