ghariksforge@lemmy.world to Technology@lemmy.worldEnglish · 1 year agoAdvertising revenue in Twitter crashes by 50%lemmy.worldimagemessage-square203fedilinkarrow-up1279arrow-down115file-text
arrow-up1264arrow-down1imageAdvertising revenue in Twitter crashes by 50%lemmy.worldghariksforge@lemmy.world to Technology@lemmy.worldEnglish · 1 year agomessage-square203fedilinkfile-text
minus-squaretraveler01@lemmy.worldlinkfedilinkEnglisharrow-up0arrow-down2·1 year agoI think these limits were increased to a point where they are not really bad or they were removed. The point of them was to prevent scraping to train AIs
minus-squareTechmaster@lemmy.worldlinkfedilinkEnglisharrow-up7·1 year agoImagine using Twitter to train AI. “Why is my AI such an asshole!?”
minus-squareUniquitous@lemmy.onelinkfedilinkEnglisharrow-up2·1 year agoI’d argue it’s a poisoned dataset. You can’t validly train an AI based on content that contains a non-trivial percentage of bot-generated content.
minus-squaretraveler01@lemmy.worldlinkfedilinkEnglisharrow-up0arrow-down2·1 year agoCould be meant more like to influence elections…
I think these limits were increased to a point where they are not really bad or they were removed. The point of them was to prevent scraping to train AIs
Imagine using Twitter to train AI. “Why is my AI such an asshole!?”
I’d argue it’s a poisoned dataset. You can’t validly train an AI based on content that contains a non-trivial percentage of bot-generated content.
Could be meant more like to influence elections…