Hacker News new | ask | show | jobs
by hattmall 906 days ago
I don't disagree with your assessment in full, but I don't exactly consider wikipedia and Amazon good results. Like they are big enough that if that's the result I want I can go to them directly. So like they aren't bad or wrong, but I can see the case for excluding them. Should something like Webster's dictionary be a top result?
2 comments

I think for single word queries like that Wikipedia covers more ground than a dictionary. Personal preference, perhaps. If I need a definition I search for "define dildo" (Kagi shows Merriam-Webster, Oxford, etc dictionary entries).
Marginalia supports the old Google syntax, e.g. "define:dildo"
Thanks! If you are that "single person" who built Marginalia... hope you are not taking my criticism personally. I am more annoyed by this blog post that uses a few handpicked queries to present generalized long winded conclusions that are completely disproven when using a different set of queries.
Yeah, its me, and to be fair I made a comment to a similar effect myself. Assessing search result quality is very hard, and this is definitely a pretty flattering selection of queries.
On the plus side - in addition to Marginalia's own success, you can take partial credit for how good Kagi search results are (IIRC Marginalia's index is one of the sources for Kagi search results). So... thank you for that!
Marginalia Search isn't trying to be a universal knowledge engine, it's just a website finder.

That's bad if you're looking for a simple answer or basic fact, and good if you're looking for a few hours of reading.