The HN API is not efficient - it says so itself in its readme. Which means neither is this. Making 1000+ HTTP requests to dupe check feels kind of insane when all you are doing is URL comparisons.
Why not instead make a single call to: https://news.ycombinator.com/from?site=whateverdomain.com, and parse that single result?