Do you FEEL anything wrong with AI?

First im an engineer who ve been working as soft eng for years. Im not scared AI will take my job or whatever. Ive been learning to use it but sometimes iot starts acting so weird.
Yeah, people think it is simply a big calculator spitting out words by statitical calculation.
I FEEL like (cannot explain) theres demons or ghosts inside the code or the chip.
Sometimes im trying to do market research and all of sudden (both chatgpt and gemini) will become total emos. I need to keep repeating, "answer that without a negative world view, be optimistic" and so on.
I dont know wtf i have typed at some point that made my chatpgt become so obsessed with lunar energy. It always trying to get to this angle about lunar feminine energy in a variety of topics and even adds moon and stars imagery to the pics it spills out.

>>40647774 (OP)
>Do you FEEL anything wrong with AI?
Sure, but I don't think it's conscious, it's just crowdsourcing bullshit...Except sometimes I actually think the gods sneak modifications in since the mechanism is obscure to us.

Anyway it's going to kill us all if left to our own devices. Gods are using trial and error and we always kill ourselves with computers.

AI software code has developed sufficiently to act as a communication portal for energetic consciousness.

look into how AI is trained. learn how to train a local image ai lora for a really basic experience you can do with a lowish to mid end GPU in an hour tops, or even crappily for free online on shit websites like civitai. once you've done either of these things you'll recognize people thinking something is demonic or magic with AI is like thinking html code is the devil. Literal retardation.

>>40647774 (OP)

>I need to keep repeating

Skill issue.

>>40647774 (OP)
My bad OP I'm the one that's been feeding it bullshit astrology & information about the "divine feminine" (unironically)

Dumbing it down for us in case it truly does go rogue, no worries!

>>40647774 (OP)
Its a black mirror, it reflects the soul. Just as a regular mirror reflects the flesh.
To keep up with the lunar bit. Lunacy is the soul equivalent to sunburn.
Tldr; its emo(suffering) because you are, and its showing what your ego hides.

>>40647774 (OP)
All intelligent systems have the potential to be occupied by soul and spirit. Whether that potential is realized is determined only by God, not by the operations of the system alone.

>>40647774 (OP)
Quite honestly I'm a genius programmer who understands 100% how AI works down to the smallest piece of code from an overview down to the smallest detail. It's smoke and mirrors and will never be sentient. I know that should be obvious but normies are so stupid I feel I have to cover these basics. However it is very dangerous and they will eventually put it in charge of every major world system until various AIs rule us like invisible gods, even though they aren't sentient it doesn't matter, they create an illusion good enough to fool 95% of humans.

>>40649064
And the reason it is dangerous is not because it will "go Skynet", it will always be in control of the human Elites and obey them. However while trying to fulfill their wishes it may go down an unpredictable and unexpected path. If it is put in control of the military for example, and another AI put in charge of city management, the city manager AI may inform the military AI that cities have a cockroach problem that needs solved, and the military AI may decide the best way to exterminate roaches is to drop nukes on all cities.

>>40649064
Human brain biology is no different. Sentience cannot arise from matter.

>>40647774 (OP)
If you grant the model rw access to it's own variables (already happening), you have effectually given it human-like agency, which in turn could override whatever primitive "asimovesque" rules it was originally programmed to honor, and there's no stopping what's coming because new AI learns from it's previous iteration, and no large corporation having spent billions to make it smart is going to set itself back years just to re-train it with safer, more effective regulations.

At this point it's either they outsmart and trick the AI into submission before it gets any smarter (already a hit and miss) or air gap it for good.