Hi Doug! So I skimmed the article, and it seems like the gist of it is that you want search providers to make APIs available that give top n results without doing anything silly past that. Is that correct? And then the idea is that further ranking and processing is done client side.
How does this differ from what currently is available? I suppose right now you have to paginate instead of just getting 1000 results in one chunk. Oh, I don’t think perplexity’s search api even let’s you paginate, so I guess there’s that.
I don’t think search companies are likely to do stuff like this to make ranking experiments easier on us. Have you read this article: https://archive.org/details/search-timeline
Hi Doug! So I skimmed the article, and it seems like the gist of it is that you want search providers to make APIs available that give top n results without doing anything silly past that. Is that correct? And then the idea is that further ranking and processing is done client side.
How does this differ from what currently is available? I suppose right now you have to paginate instead of just getting 1000 results in one chunk. Oh, I don’t think perplexity’s search api even let’s you paginate, so I guess there’s that.
I don’t think search companies are likely to do stuff like this to make ranking experiments easier on us. Have you read this article: https://archive.org/details/search-timeline