Gregory's Server(I shared this on Twitter and it's interesting how some people there are very resistant to the idea that the deeply personal insights ChatGPT gave them about themselves might be bogus junk)
|
7 comments
@simon >deeply personal insights ChatGPT gave them about themselves might be bogus junk And then - govs use it to define "value of human" by asking AI to predict behavior or people base one their previous actions/social group they represent. https://youtu.be/FkKPsLxgpuY?t=1112 Veritasium video - 18:33 time @simon It's probably even better than Cold Reading, and that's what most future telling boils down to after all  @Profpatsch yeah, it's becoming apparent that a lot of people share a LOT of personal questions with ChatGPT, which means it's more like red-hot reading for them - so YMMV but it can be a lot more "insightful" than I first suspected @simon For all the amazing power of the underlying semantic pattern engines underneath LLMs, it feels like people are absolutely determined to treat them as something much, much deeper. The end result is, like, doing your taxes with a combination of Brian Eno’s Oblique Strategies cards and help from a cold-reading psychic @simon I’ve been thinking a lot about written fiction lately and have come to the conclusion that it’s only possible because the human brain is so hard wired for in person communication that it automatically assumes that intelligible words == talking to a human. I think this is the source of most of the misunderstanding of llms too. |
@simon Has anyone spun up a daily personalised horoscope service based on an LLM yet? That thing would print money.