SEC EDGAR is an amazing data source. But this wasn’t always the case. Before 1984, companies filed paper reports with the SEC. If you were researching stocks, getting filings could’ve meant physically visiting the SEC. Then we went electronic: •1984: EDGAR pilot begins •1993: mandatory electronic filing begins •1995: EDGAR filings go online •2009: XBRL filing format introduced It took 25 years to go from paper filings to the electronic data you pull from EDGAR today. And the current system works well. But it’s designed for humans, not agents. There are 100+ filing types, 10K+ XBRL types, custom concepts, and millions of documents. Yes, your agent can query SEC data directly and analyze it. But your agent will burn time and tokens finding filings, reading them, and getting lost. The corpus is still too lumpy and unstructured for meaningful, repeatable deep research. We need a translation layer that is structured, cited, searchable, and connected at all times. Enter @findatasets