• 0 Posts
  • 6 Comments
Joined 3 years ago
cake
Cake day: July 1st, 2023

help-circle


  • But AI has a hallucination problem similar in the way every single human being in existence has an imperfect recollection hallucinatory problem. We have to distinguish accuracy from precision and take that with a grain of salt. I do wonder when the last time people here have used some of these later GPTs, because they absolutely have value and vast consensus of scientists agree.

    How one regulates it, is of course another story that I agree with you on. I’ve already used AI successfully to combat far-right trolls though.


  • These datasets, along with the rules governing them, are a bit more complicated than weighing 4chan and wiki equally. They absolutely weigh peer-review journals and white-listed encyclopedia sources higher. Once they establish a baseline of credibility and identifying a certain grade-level of communication, along with identifying critical-thinking patterns (fallacies, formal logic, etc.)., then you can quite easily filter out the noise.

    Every single time I’ve used a GPT to analyze a conversation with a Trump supporter and myself, it has identified their fallacies I had already identified, and rightfully pointing out fallacies of my own at times – but almost always concludes their argument is far, far worse.