Hacker News Reader

Comment by ivape | Hacker News Reader

The LLM should know when it regurgitates content from a specific site and page, that should be the new “pageview”. Me telling the LLM that I don’t want the first answer it gave me from a common site is the same as me going deep into Google search result page 10 (like the old days).

View on HN

> LLM should know when it regurgitates content from a specific site and page

I dont think that's how LLMs work.

Maybe some reverse attribution to determine how much credit to give different sources in the old top k results. Maybe a llm that is trained on fact finding, and giving proper % of credit. But this would probably require a collective contract in place to benefit site owners, which means its never going to happen.

That's what Google does now for some of its AI search results. There's a link icon next to some of the answers.

No, you have it backwards. What google does is have it's ai read the URLs that search found, summarize them, and tag the URL.

You don’t think it can run one more iteration on its own output to provide attribution? They could even create a synthetic page view for the sources with an exclusive bot and not change anything about how things are done (the bot is told to visit the site registering a page view).

> You don’t think it can run one more iteration on its own output to provide attribution

Please explain to me how an algorithm that takes a sequence of tokens and outputs a URL would work.

Reply on news.ycombinator.com