From question to primary source

Each lesson opens one real public archive and walks the same arc: a real question, a query that translates it, and a spreadsheet of primary sources you can read, judge, and cite. You'll pull every presidential document about Mexico from 2018, the week of a State of the Union with the full 39,000-character speech — then make the machine read it for you and hand back only the sentences that matter. You'll compare how English-speaking and Spanish-speaking Reddit talked about the same collapse, and trace TV coverage of the Haiti earthquake down to the actual, watchable, captioned broadcasts. Everything runs on Google's free tier — Colab and Gemini. No credit card; an anonymous account works if you prefer privacy; a US .edu address unlocks the advanced tier free; and when your research outgrows the free tier, the same code is one step from scaling. Investigar is the same programmatic skill as Crear and Motor, pointed at evidence — and it's where Entender's 'verify instead of defer' gets its tools. No prior coding required: Gemini writes every line, and you keep every script that works. The library keeps growing — new archives are in production now.