BrikoXM to Technology@lemmy.zipEnglish • 7 days agoStudy: Social media probably can’t be fixedarstechnica.comexternal-linkmessage-square4fedilinkarrow-up118arrow-down12file-textcross-posted to: technology@lemmy.worldtechnology@beehaw.orgtechnology@lemmy.worldantisocialmedia@piefed.social
arrow-up116arrow-down1external-linkStudy: Social media probably can’t be fixedarstechnica.comBrikoXM to Technology@lemmy.zipEnglish • 7 days agomessage-square4fedilinkfile-textcross-posted to: technology@lemmy.worldtechnology@beehaw.orgtechnology@lemmy.worldantisocialmedia@piefed.social
“The [structural] mechanism producing these problematic outcomes is really robust and hard to resolve.”…
minus-square@skisnow@lemmy.calinkfedilinkEnglish12•edit-27 days agoTheir methodology is so astonishingly flawed at the most fundamental level I had to read it twice. They got a ton of LLMs to simulate how they thought real people would react to changes in their feeds. The entire study is built on less than sand. Someone needs to sit them down and slowly explain to them how LLMs actually work.
minus-square@technocrit@lemmy.dbzer0.comlinkfedilinkEnglish3•edit-27 days agoHa yeah. The “AI” grift is a foundation for other grifts. The shitty “science” will keep getting worse and more widespread.
Their methodology is so astonishingly flawed at the most fundamental level I had to read it twice.
They got a ton of LLMs to simulate how they thought real people would react to changes in their feeds.
The entire study is built on less than sand. Someone needs to sit them down and slowly explain to them how LLMs actually work.
Ha yeah. The “AI” grift is a foundation for other grifts. The shitty “science” will keep getting worse and more widespread.