They do, the whole tone and the lack of understanding of Docker, kernel threads, and everything else involved make it sound hilarious at first. But then you realize that this is all the human input that led to a working exploit in the end...

Welcome to vibe coding. If you ever lurk around the various AI subreddits, you'll soon realize just how bad the average prompts and communication skills of most users are. Ironically, models are now being trained on these 5th-grade-level prompts and improving their success with them.