It published a proposal on GitHub outlining how users might signal whether they want their posts hoovered up for AI training or public archiving. As we all know, a polite request has always stopped tech companies from doing whatever they please with publicly available data. 

CEO Jay Graber [pictured] posted about it on Bluesky itself. Users, having been lulled into a false sense of security by previous promises that their data wouldn't be flogged to advertisers or fed into the AI meat grinder, were not best pleased.

Graber’s response was essentially: "Well, AI firms are already nicking everything that isn’t nailed down, so we might as well pretend we have a say in it."

She likened it to websites using robots.txt to tell web crawlers what they can and can't access—a file that AI firms have about as much respect for as Russia has for Ukraine's personal space. 

In its most hopeful and legally meaningless form, the proposal states that AI companies and researchers are expected to honour users’ wishes if they say they don’t want their data scraped. Although they have shown no sign of showing much in the way of honour so far.

Meanwhile, someone had a different AI pipe dream on Meta's Threads. They wished they could have a nice little chat with their feed algorithm—explaining what they like and what they don't and having an ongoing conversation as their preferences change.

Instagram and Threads overlord Adam Mosseri said he’d discussed the idea with VC Sam Lessin and that, while it’s a long way off, he reckons it’ll happen eventually.