Raw LLM Responses

Hey, I really appreciate you being open to this conversation. I definitely don’t have all the answers, but I’d much rather figure things out alongside other artists than leave it all up to tech bros optimizing purely for profit. I'm more involved in performance and composition, and I’ve been especially curious about using AI to create new kinds of instruments. Projects like Google NSynth have inspired me to think about new ways to design expressive VSTs. That said, I completely agree with you that AI is currently better suited for mundane tasks like EQ or mixing. As for the AI-to-art pipeline feeling generic or error-prone, I think that comes from the nature of how these models work. They generate the most statistically probable result based on a given prompt, which often means popular or predictable outcomes. The way casual users approach music is very different from how you or I might engage with it. If they hear a mistake, and that’s assuming they even notice it, it often doesn't matter to them. For them, it’s just something to have on in the background, not something to experience in full depth or dynamic range. That disconnect is at the heart of a lot of misunderstanding. From what I’ve seen, two key things contribute to the gap. First, art is about beauty, and beauty takes time, care, and experience to develop. Second, many people using text-to-creative tools don’t care enough to reach for that kind of beauty. They settle for whatever is "good enough" because they’re not used to putting in the effort to refine their work. A lot of folks haven’t ever done a creative project at all. It's a bit like how some people say they're bad at math and give up. Our culture often discourages people from being multidimensional, especially when we’re trained to fit into rigid roles from a young age. Back during the Renaissance, people aspired to be skilled in both art and science. That spirit feels missing today. Also, as someone who works in tech, I totally get what you're saying. The mindset around AI is often focused on speed and results. That works fine in software, but when applied to art, it results in outputs that feel hollow or even broken. Those rough edges are what frustrate creatives. When the general public embraces those flawed outputs, it can feel disheartening, even insulting. I'm not happy that some artists may lose their jobs, whether it's because AI replaces them or because other artists become more efficient by using it. But I do see a possible future where AI lets artists work faster and more affordably, which could actually increase the demand for custom work. If artists can deliver commissions at five times the speed and with greater consistency, more people might be willing to pay for personal, high-quality art. That outcome is far from guaranteed, but it gives me a reason to stay hopeful. About that quirky AI music, I wish I had a link. I heard it while eating at a restaurant and laughed a little with my tablemates. They had no idea it was AI-generated until I asked the owner, who confirmed it. It wasn’t what I would call good pop music, but it had a lot of charm and was clearly tailored to the restaurant’s theme. If I can get a copy or a public link from the owner, I’ll definitely send it to you in a DM. Thanks again for taking the time to talk this through. This kind of thoughtful back-and-forth is rare in conversations about AI and art, and I really respect that you're willing to engage in thoughtful dialogue over hot takes.