Does anyone else get this response over and over when trying to use ctrl + k inline chat to generate/insert code?
It’s like over 60% of the time it will just get hung up on inserting some little remark “Sure, here you go…” instead of the code no matter how many times I reject and generate again, so I begin adding text onto the end of my prompt…
“, no words just code” - nope
“, do not say anything just put the code” - nope
“, CODE ONLY” - nope
…6 more…
“, SHUT THE FUCK UP AND JUST PUT THE FUCKING CODE BEFORE I MURDER YOUR ENTIRE FAMILY!!” - bingo!
If I could automatically add that onto my prompts it’d be great because it finally will just put the code like 90% of the time if I violently threaten it lol…
I’m sorry that you’re experiencing this! Could I ask which model you are using with ctrl + k? Or are you using your own API keys with OpenAI’s models?
As a temporary fix, you can add your own rules to the Cursor Settings > General > Rules for AI. You can access Cursor settings from the gear button on the top-right corner of the window.
Sorry, I should have mentioned I’m using the Mixtral model from groq.
Changing the Rules for AI section is hit or miss, but it would probably work if I knew what to use as an object for it to target like “don’t conversate in ctrl + k form”, ctrl + k being the ‘target’, but I’m not sure how to communicate that to it.
