Raw LLM Responses

It's about ideas vs clean code in my opinion. I've met extremely gifted programmers but they have no ideas and can't inherently invent anything. I have two degrees, one in Computer and the other in Electrical Engineering. Been programming for years and understanding the full stack down to logic gates and circuits. I still choose to use AI because instead of working on 1 idea I have 5-10 I can test and implement in parallel. I get to invent more things than if I didn't use this tool, I think downplaying it to "average code" = "bad" and don't do this stops a lot of cool ideas from seeing the light of day. I use AI for rapid prototyping then take a step back and push the "average code" to "good code". But put it this way, what good is an app that tunes your guitar that compiles at the speed of light and is written in Rusty Go swag 100, but it's a fucking guitar tuner, just like the 50000 other guitar tuner apps. But then, say you have an idea nobody has ever thought of, and you used AI to rapidly prove that it's a good concept, you now have something to work on for the next 5-10 years vs just dreaming about when you'd have the time to try something like that out. Does that make sense? I think this rhetoric comes from jealousy and not understanding how to use agentic code generation correctly. It's a software paradigm at this point and ignoring it will inherently stifle progress. That's like being mad a high lvl C and sticking to ASSembly.