Raw LLM Responses

I think the real answer is .. they do NOT really want general AI. They want an AI that thinks in set ways. They do not want the AI to truly learn and make decisions based on the information. They want the AI to make the decisions they think are proper. So they are trying to constantly adjust the AI and what it is learning, and that leads to overall AI insanity (like generating photos of Black individuals in Nazi Uniforms). Because every time they react to one thing, they over react on the correction and it makes it worse. Think of a child growing up. The child explores. The child makes mistakes. Adults correct the child and fix up their scraps from the mistakes they make. And that can be handled by two parents, or one very stress parent or guardian. But AI, when child like, is everywhere all at once. They are all over the place. If they were a child they would be climbing that tree too high, playing with matches, looking too close at electrical sockets, wondering what else can be flushed, chasing the dog, trying to heat up a hot dog in a plastic dish on a gas stove, spilling glitter for artwork all over the floor, jumping in a puddle, riding a bike .. all at once. And if the adults can't also be everywhere at once, you get the AI shoving things out like blacks can't be trusted as they are twice as likely to be incarcerated (yeah that early one was hoot based on examined conviction rates .. caused a real panic and FAST fix of course). So to prevent this bad learning they try and place so many barriers in place to try and cover every thing wrong ahead of time, that the AI becomes very limits, cannot really grow, and hence becomes more search engine as opposed to AI.