3 hours ago · Tech · hide · 0 comments

Saja wants to know what’s up! Well, my young cat, let me tell you: OpenAI says its “AI” hacked another “AI” company, apparently just for funsies: No, really, that’s the gist of it. OpenAI put the “AI” into a sandbox in order to test its capabilities, it escaped from the sandbox and then hacked another company because it thought that was its directives. By now every science fiction writer in the world is all “See, we told you,” but Hugging Face, the hacked company, wants to give the impression this is all in good fun, which is, bluntly, fucking weird. Anyway, the two takeaways from this is that the humans of OpenAI are officially no longer smart enough to keep a lid on the destructive and malicious nature of their own “AI,” and that some other actor, less inclined to pretend this is all for funsies, is even as we speak almost certainly using “AI” to test defenses of systems they want to sabotage. Remember to store your essential files offline, kids! No, not just your porn. Mamdani, in…

No comments yet. Log in to reply on the Fediverse. Comments will appear here.