From what I understand, you will get the content from the closest peer. So, if someone was on your local network, ie node 3 , you would get the content from them.
I’m working on as webapp to be a search engine for hash based content. It doesn’t host the content ( ipfs does that ). It will be a place where users can post hash + metadata. Something like …my content = username+hash+file details(size, type, date). Users can search database by author, hash, file details and also rating. When you find something you like, you can view it in ( if it’s browser content ) or download/pin it. App will keep track of how many people pin a certain hash and how people rate the file.
This will allow you to quickly browse through huge lists of hashes and filter it in a way that suits you. Not like google where search are manipulated by various factors.
This also gives users the chance to see which other user posted content and how others have verified, rated that content.