... their ass from a hole in the ground when it comes to LLMs: the inputs for the models are huge, at least I imagine so. I further assume they are too huge to be catalogued and curated by hand. So, how do they keep from hoovering up the stolen data floating around on the "dark web"?
Point is, at some point, should we expect to be able to ask ChatGPT for someone's SSN, DOB, favorite porn, last sex toy purchase, etc.? I can easily see a hacker deliberately dropping some of this info onto a site they know one of these models is "sampling", just for the shits-and-giggles of it.
Or, am I just somkin' crack?