Maybe? LLMs sometimes provide suggestions that are novel or highlight overlooked ideas. Sometimes I'll oneshot a narrow problem and invite coding agents to get creative about how they chart some complex data and am pleasantly surprised by the results.

I'm absolutely not snobby about where the ideas originate or how they are realized, if the end product functions well. I don't don't want to waste time assessing newly-reinvented wheels. This problem predates the advent of vibecoding for the masses.