Raw LLM Responses

Hi! Computer Science student and hobbiest here! I love hearing about all of this from the mouths of the artists. I hate how misunderstood "AI" (a term I hate immensely by the way) is in the public eye. Machine Learning is a wonderful thing we've developed, the ability to make algorithms far more nuanced and precise, absurdly sensitive to patterns and details, than anything hand made is so useful! See the medical fields, and various scientific frontlines, and the power and use of machine learning algorithms is apparent. *However,* we are abusing the strengths of machine learning algorithms to make party tricks-- fun to see, but wholely useless. "Generative AI," as in machine learning algorithms trained to generate pieces of creative works (mostly of the visual variety, but not exclusively so), do not solve any problems, but rather is a neat trick the math can do if you bully it hard enough. Machine Learning is best used for finding patterns and sequences where the human mind would struggle to see, either due to scope or sublty. *However,* (2: Electric Boogaloo) even then not everything needs the powerful pattern perusing program. No need for an "AI" to be trained to tell me if a number is even or odd. Just modulo 2 and see if the remainder is 0 or not. That logic will always work, not by fine tuning a rube-goldburg series of mathematical operations to find some "hidden, subtle pattern," but my simply using good ol' logic. Extrapolate this out and my point is that there's a point, more of a boundary, at which using AI is not viable, not useful, or just a plain misuse of the tool. Last thing I want to , I do see some genuine benefit for "AI" in the creative field. As a *tool*. But not in this generative sense. I mean machine learning algorithms that can try to aid an artist in some way or another that helps them express their creative vision by their own hand. Not try to replace them or make "art" for cheap or whatever. Tl;dr: Machine learning is cool. Use it for what its good for (pattern and sequence recognition and prediction), and not these toxic party tricks that do nothing good for anyone. Progress cancer research or something-- a far better use of CPU cycles.