Mastodon Feed: Post

NfNitLoop ("Cody Casterline 🏳️‍🌈") wrote:

The post is mostly good, but it suggests tracking:
> What percentage of production defects trace back to AI-generated code versus human-written code?

Unfortunately, a problem with IDE-based AI suggestions is that, AFAIK, there is no way to trace which lines of code were written by the developer vs. suggested by an LLM.