PostgreSQL's MVCC is bad. So is everyone else'sboringsql.com 17 pointsradimm1 month ago4 commentsSaveHideCopy link On HNComments−twoodfin1moAn LLM can’t write a technical article with a strong, clear narrative to save its weights.This is a great example. Why was this article written? What is it supposed to teach me? What is its unique point of view?Of course it doesn’t have answers to any of these questions because they weren’t in the prompt.−radimmOP1moOP here - thanks for feedback. They weren't in prompt because there wasn't any (for polish yeah, not hiding it).−Nican1moI think the article is great as-is, showing how the initial architecture choices can have deep consequences.−eschaton1moIf someone wants me to spend my time reading something, they need to be willing to spend their time writing it.
Comments
An LLM can’t write a technical article with a strong, clear narrative to save its weights.
This is a great example. Why was this article written? What is it supposed to teach me? What is its unique point of view?
Of course it doesn’t have answers to any of these questions because they weren’t in the prompt.
OP here - thanks for feedback. They weren't in prompt because there wasn't any (for polish yeah, not hiding it).
I think the article is great as-is, showing how the initial architecture choices can have deep consequences.
If someone wants me to spend my time reading something, they need to be willing to spend their time writing it.