I don’t think (completely wild guess here) AI content crawlers should have any more impact than the dozens and dozens of spiders that make up must of my own site’s traffic.
The impact was magnified for Twitter because it generates so much new content every second. That wasn’t an issue when Twitter had a nice, properly cached API and it shouldn’t be an issue for fediverse instances either because we have RSS and caching and we’re not so stupid as to turn those off. Like, what kind of moron would do that?
The issue comes when those AI bots start commenting and posting here. From what I understand, bots are a large reason why Beehaw keeps defederating from instances with open registration: bots are difficult to moderate without good moderation tools.
to be fair, that latter argument about the magnified effect on twitter operates under the assumption that elon wasn’t just lying to cover up that he didn’t pay his google cloud bill. the amount of users who view and create that content still create a much higher load on the servers than AI scrapers that want to read it once and save it somewhere for training
I don’t think (completely wild guess here) AI content crawlers should have any more impact than the dozens and dozens of spiders that make up must of my own site’s traffic.
The impact was magnified for Twitter because it generates so much new content every second. That wasn’t an issue when Twitter had a nice, properly cached API and it shouldn’t be an issue for fediverse instances either because we have RSS and caching and we’re not so stupid as to turn those off. Like, what kind of moron would do that?
The issue comes when those AI bots start commenting and posting here. From what I understand, bots are a large reason why Beehaw keeps defederating from instances with open registration: bots are difficult to moderate without good moderation tools.
to be fair, that latter argument about the magnified effect on twitter operates under the assumption that elon wasn’t just lying to cover up that he didn’t pay his google cloud bill. the amount of users who view and create that content still create a much higher load on the servers than AI scrapers that want to read it once and save it somewhere for training