DeGoogle Yourself

@Ephera@lemmy.ml @dating1999@lemmy.ca

site:domain.tld does work, as an and x constraint. I often use it.
It seems to me like the OP's specific and (not x) usage of site:domain.tld is the reason why it isn't working. While the negation prefix (-) does work for tokens/words (e.g. mercury -freddie), it's probably transforming site into a token not to be included in the results (i.e. "any results that don't contain the word "site") which, disconnected from the rest of the sentence (:quora.com), turns the latter into part of what the results should include, so the query ends up being something like:

Filter all the indexed Web results where its contents don't include the word "site", possibly do include "quora.com", possibly do include "Molten", possibly do include "boron", possibly do include "oxide", possibly do include "attacks", possibly do include "silicates"

The negation prefix has a similar effect to that of positive (+) prefix (e.g. "mercury +periodic +table") as it turns the word into a required condition (must be present for "+", must be absent for "-") rather than an optional condition (i.e a search for "mercury periodic table", without quotes, will contain pages with all three words in any order, pages with just two of the three words (such as "mercury periodic" in any order) and pages with only one of the three words (such as "mercury" which would include pages talking about the singer, and pages talking about the planet and pages talking about the Roman deity), ranked by "relevance").

As Quora pages do include "quora.com" somewhere within the page body, the first results will be from Quora because it's part of the parsed condition (which is to optionally include "quora.com" as part of the result while discarding results containing the verbatim word "site").