Mastodawn

I just read "No, AI is not Consicous" by Ted Chiang. It struck me what I find uncomfortable about using LLM-generated code.

LLMs generate tech debt, but they also generate moral debt.

LLMs have no moral agency. They can't take responsibility. It's like working with interns. You can't expect them to know better.

How do you know good decisions were made? Who will take responsibility when tings break?

Taken to its logical conclusion, this line of thinking is absurd—and damning.

The Atlantic