Here is a fresh post where lady is amused by different AI bot's answers when asked how they would approach to brake the simulation of reality if they where humans?
Lady and audience became freaked out and active. As if AI gave some never seen before answers. In fact it gave great logical step by step answers supported by data that was fed to LLM. Check it if you want: https://www.reddit.com/r/StrangeEarth/comments/1g3rtds/ai_is_become_self_aware_and_letting_us_know_how/
Very logical and normal answers. I would give the same answers. Maybe people are freaked out because they became lazy to think themselves, but that is not the question I have a right to ask. LLM's gave logical step by step answer, only thing occult stuff info wouldn't be used by me because it's not scientific. AI gives logical answers about using occult data in a research but in this question it was not asked if occult stuff is legit, proven or real. It just uses it as a part of a construction to solve given task...
And that gave me the thoughts! That is a clear sign of not having any "Izek Azimov" rules yet. Every conversation still might relate to a situative truth. Not generally excepted one. So if AI is asked to be a "judge" over some people now, it will use basic rules of humanity but in every situation it can act just... logical I guess... And the "Castle structure out of given blocks". And that is not always the right way we see it. So we are still far from AI ethics. And it is not clear on which stage to apply it and how. Though I think studies toward that are done while we soeak.
But looking into quantum dramaturgy of LLM's is great... How they solve dramaturgical tasks is one step closer to braking this world simulation, I'm sure! BTW here is a book on SSRN that sort of teaches how to look into this story-making code of reality.
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4530090
And a 6 min video in case you don't want to spend time: https://youtu.be/22kuYSZUdqY
[link] [comments]