This expose is a remarkable rebuke of AI by an authentic genius. This post should be made required reading by high school students before they enter the mind-altering world of college and university learning. My tests of AI have been rudimentary and simplistic, but even they quickly demonstrated how even a very simple question about temperature conversions stumped the "program." Dembski has gone much, much farther in exposing AI fallacies. Unfortunately, AI is going to become well established long before its limitations become widely known. (This comment was not subjected to AI.)
This isnt exactly surprising given that we find that AI largely attempts to "cheat" when possible. We will likely see a rush on AI for superstimuli, via porn, etc rather than quality and effort.
At the very least, it seems AGI is going to require some radically new ideas. A realized AGI should be able to give us a James Joyce. As my little exercise showed, LLMs are not even close.
If you want a good laugh, check this one out:
https://x.com/keatonpatti/status/1161284670601990146?s=46&t=vIKLRIDXXDlDBrsDAwVP0Q
Someone had a bot watch 1,000 hours of Batman movies and then write its own script.
This expose is a remarkable rebuke of AI by an authentic genius. This post should be made required reading by high school students before they enter the mind-altering world of college and university learning. My tests of AI have been rudimentary and simplistic, but even they quickly demonstrated how even a very simple question about temperature conversions stumped the "program." Dembski has gone much, much farther in exposing AI fallacies. Unfortunately, AI is going to become well established long before its limitations become widely known. (This comment was not subjected to AI.)
Thanks Forrest. The genius here is James Joyce. Everything else is clever and not-so-clever manipulation.
This isnt exactly surprising given that we find that AI largely attempts to "cheat" when possible. We will likely see a rush on AI for superstimuli, via porn, etc rather than quality and effort.
https://futurism.com/sophisticated-ai-likely-lie
At the very least, it seems AGI is going to require some radically new ideas. A realized AGI should be able to give us a James Joyce. As my little exercise showed, LLMs are not even close.