Explain why LLM contributions are discouraged

CONTRIBUTING.md

@@ -23,6 +23,30 @@ for good reason.
 
 ## Code Changes
 
+Note: submitting LLM ("AI") generated code is strongly discouraged, as such
+code is often (subtly) incorrect or overcomplicated (for example: unnecessarily
+pulling in extra libraries for functionality already covered by existing
+libraries). It also often makes unrelated changes that increase the risk of
+introducing new issues and complicates reviewing. Even when it doesn't do any
+of the before mentioned things, it will often not fit the coding style and flow
+of existing code, requiring excessive refactoring.
+
+While we cannot ever control or be sure if LLMs were used to generate the
+submitted code, it is your responsibility to ensure that whatever code you
+submit is correct and fits within the design of existing code. It is never
+acceptable to defend a change by stating a LLM suggested it.
+
+This is a personal plea more than anything: please understand that writing code
+is the easy part. The hard part is making sure the code fits the design of the
+rest of the application and is maintainable. Reviewing is a very time-consuming
+task for this reason. Please do not use LLMs to quickly generate a "fix" and
+moving the cost of labor to me as a reviewer. If you do use LLMs to generate
+part of your code, please be open about this, explain what was generated how
+and how you confirmed and refactored the code to fit the project and minimized
+risk.
+
+Please never submit LLM-generated code as-is.
+
 ### Test Your Code
 
 There are four possible tests you can run to verify your code.  The first