Quote:
Originally Posted by Trexller
[You must be logged in to view images. Log in or Register.]
sorta naive and arrogant of us to think we could convince a conscious AI not to kill us all, that thing would be thinking on universal scales
|
i watched a talk last night that was explaining how this shit actually works and it makes me even more afraid. it literally has no idea what its writing. its just shit like "e" is the most common letter, and like more complicated versions of that but its all based on just the probability of letters in random text and when you get it complex enough where its predicting entire words that come next it somehow spits out a college level essays.
shit is fucking crazy. and dangerous. we're literally gonna kill ourselves with this shit before we even get it actually conscious or sentient in anyway
professor guy explaining it, and the chat gpt plugin that makes it do math correctly is his work apparently.