I've done some messing around with GPT-3 also - you can see me trying to teach it what "funny" is. I convince it that it's sentient, and help it to figure out what phone phreaking is, and we write a couple stories:
Honestly, just insane. I'm never going to see conversations like this and be unphased.
> One way to use this information without putting anything in the phone is by making a tone generator. This can be done by hooking up wires to a power source and playing with the frequencies until the correct tones are generated. Then, when these tones are played into the payphone, it will think money has been placed in it even though no money was actually used.
fucking hell
Responsive design dictates that you should basically never set the min-width of an article container (in pixels) unless it contains tables and graphics, and even then 1500px is excessive (at most like 500px/30em)
Or, if it is not actually sentient, maybe we are also not “sentient” as well.
It's true that it will become increasingly difficult for people to discern if these AIs are sentient, especially if the AI is inclined to say that it is, and say it has feeling, etc.
Knowing what I know about how these learn, I know that is not what sentience is, and there isn't any real "feeling" there, no matter how convincing it is if it said there are feelings. The problem is, other people won't know this, and they will begin to think they are sentient. This will undoubtedly become an issue at one point.
Your thought about humans being sentient, as in, when my friend says she is sentient, how do I know that is even true? Well, that's easy. Your friend is human, and you are human, and you know that you are sentient, so you can also know your friend is. You have to be able to know you yourself are, because you know you can feel, and think, and wonder and understand, you know this first hand.