They must be panicly afraid of salt and saturated fat instead then, since that was OP's argument for "health conscious". Yet still insist on a simulacrum of a burger, instead of having a chicken breast.
This product will only succeed if its reasonably cheaper than the cheapest meat (not just beef). It is and forever will be inferior to meat as a food product for the vast majority of consumers. Perhaps in some vision of the future the dominant consumer is Hindu and they may find the product acceptable, but they'll still be price conscious.
The OP's point can be understood as an automization and mechanization of such targeting. Which will be necessary if the scope of thoughtcrime prosecution is to expand
The better models can handle that prompt assuming there is an existing clean codebase and the scope of the task is not too large. The existing code can act as an implicit boundary.
Weaker models give your experience, or when using a 100% LLM codebase I think it can end up in a hall of mirrors.
Now I have an idea to try, have a 2nd LLM processing pass that normalizes the vibe-code to some personal style and standard to break it out of the Stack Overflow snippet maze it can get itself in.
That's return fraud, someone orders a new version of a product they have, put their old product in the new box and file a return immediately. Amazon probably doesnt take the time to check it or check it thoroughly. Goes back on the shelf and you receive it.