There's a difference between static chatbot-generated code that you commit to your VCS, and dynamic code generated based on user content. I'm not talking about the former case.
> I'm not talking about chatbot generated code that's committed to your VCS.
I'm talking about code that's dynamically generated in response to user input in order to perform the task the user specified. This is happening today in agentic systems.
Should we relax python syntax checking to work around the poorly generated code?
I'm talking about code that's dynamically generated in response to user input in order to perform the task the user specified. This is happening today in agentic systems.
Should we relax python syntax checking to work around the poorly generated code?