Sidenote: While writing this, the VSCode AI-driven auto-complete suggestions have been interesting to observe because there probably isn't that much public code out there that does this, but it clearly has it in the training data because it keeps suggesting extremely specific things that are usually a little bit off. It shows how problematic these tools can be when the suggestions aren't just some large scale average of implementations.
(I escaped all of them because I am writing this my way.)