True lol, i feel like you are the only one getting my point in this thread xD
I still think it won't be able to make as quality ones as humans can
I guess it depends on what you mean by quality. If you mean detail, then I'm sure it's just a matter of time. If you mean creativity-wise, then no, these systems can't and won't.
You'd need completely different technologies that either matches/exceeds the capabilities of the human brain or at least does so in terms of artistic creativity, which isn't a standalone thing. What we call creativity or inspiration isn't a self-contained system, it's intertwined with the rest of our brain and pulls from other aspects of how it functions, from our abilities to see, hear, physically feel, emotionally feel, smell, create an internal world model and extrapolate predictions from it, make novel connections between information gathered through experiences which are subjective and unique to the individual.
Generative ML systems can only look at a bunch of examples of something, break it down into chunks of data, how they relate, statistics on how they relate, then use that along with some weights and some pseudo-randomness to help the illusion that something more is going on, then spit out some predictable content.
That content may not seem predictable with a very large model if you haven't seen enough examples of its output but the system is pseudo-deterministic. Even if you use truly random number generation, the output is still going to have follow rules that ultimately make it deterministic, otherwise you'd end up with too much garbage output.
This is why with these models, you have a randomness/temperature setting you can change.
