Hacker News new | ask | show | jobs
by croes 1049 days ago
There is a difference between a statistical analysis of a text to categorize by certain words or word groups and training an AI model to generate texts on the data used for training.

The later creates massive competition to human writers, the former is just an information for potential readers.

1 comments

Both the former and latter are information for writers. Neither create massive competition for writers (not that there is any law against creating competition), just FUD and better tools for writers.
Things like GPT already create competition for authors even using their names.

https://news.ycombinator.com/item?id=37042561

Pure text statistics won't do the same.

Wrt. your link, the same thing could have happen to this author if these spam books contained complete gibberish (and someone listed it on Amazon/goodreads using the authors name). This isn't legitimate competition (i.e. books written by LLMs that rival the quality and style of the actual author). This is a failure of the selling platform to QC the books they are selling.
LLMs make the scams better just like they will make spam better.

If it's gibberish you know you got scammed, LLM texts look convincing so you don't know for sure.

I agree LLMs can make better spam. But good spam isn't real competition. It's not like anyone is debating whether they should buy the latest book from their favorite author or the latest book from their favorite author's clone LLM (which is known to have written some solid books). Again, this is an issue that needs to be solved by vendors (it seems like all they need is a system where authors get a copyright to their name, and can curate the list of titles published under their name).
That's partly true for established authors,but how many readers will get fooled and think those books are real. See these 11 Harry Potter books from China

https://11points.com/11-amazing-fake-harry-potter-books-writ...

It's worse for new authors, they disappear between all the AI authors.

Publishers and readers will have to search a bigger haystack to find the needle.