Mastodawn

@nixCraft @duckduckgo @google @Bing @brave Search engines should only return results where the source material has a citation attached (and the citation exists).

Change my mind.

@davidr @nixCraft @duckduckgo @google @Bing @brave So if I'm searching for an answer to a computer problem, and someone has solved it and details it in their blog — what would the relevant citation be?

Jun 18, 2023

@fishidwardrobe @nixCraft @duckduckgo @google @Bing @brave The manufacturer's documentation, source code or similar.

@davidr @nixCraft @duckduckgo @google @Bing @brave Software doesn't have a "manufacturer" and sadly often has little documentation. Source code is only intelligible to a few.

For argument's sake, let's say the blog post links to the issue in Github – which has sat unsolved for six months, and is basically an unanswered question, so doesn't prove anything. The blogger writes, "I couldn't find anything about this anywhere, but here's what seemed to fix it for me."

Where's the citation?

Jun 18, 2023

@fishidwardrobe @nixCraft @duckduckgo @google @Bing @brave You are putting the onus on the wrong person. You should be asking the polluters of our hard-won small nugget of true knowledge why they should be allowed in, not allowing them in by default unless I come up with a perfect system.

@davidr @nixCraft @duckduckgo @google @Bing @brave So when you said "Search engines should only return results where the source material has a citation attached (and the citation exists)." was that putting the onus on the LLM folks? Or is that putting the onus on everyone else?

You seem to be asking everyone else to prove they have a right to a search result?

Jun 18, 2023

@fishidwardrobe @nixCraft @duckduckgo @google @Bing @brave No, it's putting the onus on knowledge pollution engines to prove they have a positive contribution to make.