I fully agree with you, and especially on the 'temporarily' part. The good thing is, they can one-shot as much as they want, but almost any fully LLM-generated project that pops up still has inferior code quality compared to those projects from people who actually know what they are doing, so people who prefer code quality will ignore most of the AI spam anyway, just like we do. People deploying one-shot projects out there also have a very notable lack of common sense, skill, and knowledge on topics like security and/or code optimization (obviously, because they don't code), which is mostly obvious both in their way of argumentation and their published artifacts.

They are happy to have their ads-maximized agreed-to-all Cloudflare/Anthropic/OpenAI/Google/Apple/Microsoft accounts so they can one-click deploy that "worker" with all the UI gears spinning automatically, because who needs to know how to configure a web server anyway? :^) And wow, have you heard that Cloudflare offers an operating system now??