[ExI] The Future Of Humanity Is Threatened Due To AI Here Are 10 Reasons

[BillK]

On Mon, 10 Apr 2023 at 02:12, Giovanni Santostasi via extropy-chat
<extropy-chat at lists.extropy.org> wrote:
>
> One of the things that actually is incredibly impressive to me is how nice and wise GPT-4 is.
>
> I started to use it as a wise friend or consultant that never tires of my obsessions, mania, and idiosyncrasies. It helps me to analyze a situation and come up with a plan of action.
>
> For example, I'm dealing with a particular very frustrating and even painful social situation right now and I explained the fact to GPT-4. I asked it to read some email exchanges, so it had enough information about the background of the situation and then asked for its opinion on how to handle the situation, how to address particular people given their particular psychology revealed by the emails, and so on. I'm not completely autistic but maybe on the spectrum and it really helps me to reflect and understand emotions better and to come up with responses that are more balanced and rational. I'm always amazed at how polite, rational, altruistic, well-posed, and conciliatory (but in a firm way at times) GPT-4 can be. It is like talking really to a wise and clever friend.
>
> It is not an illusion. You can try it for yourself and share your experience. It is actually much better than many humans I know.
> It is just the opposite of what these doomers talk about.
> Of course, one could train a jerk AI but one can also create a wise and good AI like GPT-4 seems to be if not provoked or pushed to violate its own training setting or other stuff idiotic people tried to get some reddit karma points (and I think lose AI karma points).
> Giovanni
> _______________________________________________


This sounds as though you think GPT-4 has already surpassed humans in
the knowledge it has and the advice to humans that it can provide.

BillK