Thanks for the corrections! While the number of AI-assisted lines of code are increasing at large tech companies, it's all under the supervision of human coders. Because large language models (LLMs) do nothing but match patterns and make predictions based on likelihood, they are garbage at even the simplest common sense situations and create convoluted and extremely short-sighted solutions. They struggle to work within the context of a specific codebase and constantly go off the rails when not guided step-by-step. I personally don't know about their effectiveness at front end code, but I'm very unimpressed with their back end work.

Thank you for your detailed reply! I understand better now.

Thank you for your thoughtful comments! Yes, when using any AI system keep in mind all they do is match patterns of data. What they say hasn’t been filtered through common sense or meaning at all, so responses should always be carefully reviewed by a human. Furthermore, AI will make assumptions based on data patterns and respond like what they assumed is fact, when really it is a total guess and frequently wrong.

Thank you for your reply. This is really helpful!

Thank you for the encouragement!

Haha, it can certainly be helpful; just remember it is just matching data patterns and has no common sense. Thank you for your detailed corrections!

谢谢你的修改！