Sourcing & Methods

Anyone can publish a confident voice over stock footage. The thing that separates a deep dive from a hot take is the paper trail. Here's how we work.

Primary documents first

Wherever a primary document exists — a court filing, an official report, peer-reviewed research, government data, declassified archives, unit records — we cite it directly rather than relying on someone else's summary. Every episode that makes factual claims ends with a numbered Sources list linking those documents, so you can check our work.

Fact, inference, and the line between

Some stories live partly in records that were lost, burned, classified, or never digitized. When we connect documented dots into a larger picture, we say so — what the record proves versus what it strongly implies. We'd rather tell you where the certainty ends than pretend it doesn't. A well-sourced inference is still an inference, and we label it.

AI as a research tool, not an author

We use AI to do what it's genuinely good at: sweeping large archives, cross-referencing thousands of pages, and surfacing connections a single researcher would miss. Then we trace every surfaced claim back to a primary source a human can read. AI widens the net; the citations are still the catch.

Corrections

Found something wrong, or have a primary document we missed? That's a gift, not an attack. Email hello@deep-dive-podcast.com and we'll fix the record.