logoalt Hacker News

f311a • today at 2:26 PM • 4 replies • view on HN

This is a known problem, when ChatGPT uses web search tool, top 10 results are often LLM generated sources as well.

They now train on their own content.

Eating your own d̶o̶g̶ f̶o̶o̶d̶ poo.


Replies

cyral • today at 3:21 PM

The top 10 results were getting pretty bad even pre-LLM. I remember tons of poorly written articles near the top, just gaming the system with SEO. Everyone was paying some guy on Fiverr $5 an article to churn out hundreds of articles with little regard for accuracy.

➕ show 2 replies
phatfish • today at 3:31 PM

Web search is pointless for certain topics now, if you don't want LLM filtered information. The first few pages of results are obvious slop.

scotty79 • today at 5:57 PM

LLMs are still better because they can routinely go beyond first page of search results.

hyperpape • today at 2:35 PM

Two things:

1. This is a 2022 era picture that LLMs just suck up what's on the Internet and regurgitate it. It hasn't been true for a long time.

2. There's almost no one who has a stronger incentive to distinguish LLM generated text from non-generated LLM text. Training classifiers is pretty manageable.

➕ show 7 replies