Hmm. Reverse-engineering this page with BigQuery is surfacing a lot more results than the page itself (for 2019 atleast).

https://docs.google.com/spreadsheets/d/1he6ca0BBYbj2ZEOEpu8L...

    #standardSQL
    SELECT id, title, url, score
    FROM `bigquery-public-data.hacker_news.full`
    WHERE timestamp > '2019-01-01'
    AND REGEXP_CONTAINS(url, '.pdf|arxiv.org')
    ORDER BY score desc