Like he says, the "comment" describes what he wants to do. That's not what humans are interested in. The human already knows "what he wants to do" when they read the code. It's the things like "why did he want to do this in the first place?" that is lacking in the code, and what information is available to add in a comment for the sake of humans.
Remember, LLMs are just compilers for programming languages that just so happen to have a lot of similarities with natural language. The code is not the comment. You still need to comment your code for humans.
> Like he says, the "comment" describes what he wants to do. That's not what humans are interested in.
When I'm maintaining other people's code, or my own after enough time has gone by, I'm very interested in that sort of comment. It gives me a chance to see if the code as written does what the comment says it was intended to do. It's not valuable for most of the code in a project, but is incredibly valuable for certain key parts.
You're right that comments about why things were done the way they were are the most valuable ones, but this kind of comment is in second place in my book.
Remember, LLMs are just compilers for programming languages that just so happen to have a lot of similarities with natural language. The code is not the comment. You still need to comment your code for humans.