This page exists because a site that argues for transparency in public life ought to be transparent about how it gets made.
A note on what this site isn’t: I’m not a journalist, and I’ve never said I was. What I’m doing on here is closer to detective work – taking what’s been reported across a lot of outlets, pulling the pieces together, asking the questions nobody’s quite asking, and writing up what I find in my own voice with the receipts attached.
Where the work starts
Most stories on the site start with reporting that other outlets have already done. The New York Times, Washington Post, Miami Herald, Bloomberg, ProPublica, the Commercial Appeal, the Tennessee Lookout, the Tennessee Tribune, the Memphis Flyer, the Daily Memphian, and a long tail of others. I read across a lot of outlets on a given story because the full picture usually isn’t in any one of them. Different reporters got different pieces. The detail in one piece in one outlet is the thing that explains the missing piece in another piece in another outlet. Putting it together is most of the work.
Where the work goes beyond the existing reporting is in the synthesis and the framing. I’m looking for the question other outlets aren’t asking. Why this thing keeps happening with this particular set of people. Why a story that should be a scandal got dropped after one news cycle. How this thing connects to that other thing from six months ago that nobody’s tied together. That’s where the value is, if there’s value here – the connections and the context, not the underlying reporting.
For deeper pieces I’ll also pull from primary documents – court filings, SEC filings, tax filings, FEC filings, government meeting minutes – but I’m honest about how that works. AI helps me find what’s relevant, surface the documents that matter, and pull the pieces I should look at. Then I go through them, read them, and figure out what I’m looking at. I’m not pretending to be combing the federal docket from scratch.
Sources get cited inline or at the bottom of essays. The takes are mine. The underlying reporting and the documents underneath are documented.
The point of view
I have one, and I’m not going to pretend I don’t. The press has to stay neutral. That’s a virtue in a lot of ways and a problem in others – mostly that there are things that are obviously wrong, obviously corrupt, obviously ridiculous, and nobody at a mainstream outlet can say so. I can. That’s part of the value of doing this outside the institutional press. If a thing is a grift, I can call it a grift. If a thing is ridiculous, I can say it’s ridiculous. I’m not handing you both sides of whether a thing is wrong. I’m telling you what I think about it and showing you the receipts so you can decide if I’m reading it right.
The eight-step research pipeline
The RATC Project essays follow a structured pipeline I worked out over the course of building this. The rough shape:
Start with a documented event or filing. Identify every named party and run them against the GriftMatrix and the Epstein archive. Cross-reference with public reporting. Map the network connections that surface. Identify the institutional players – banks, law firms, PR shops, government bodies – who had to look the other way for the documented thing to happen. Pull the underlying primary documents. Write the essay with citations. Update the GriftMatrix and the Connections map with anything new the essay surfaced.
The pipeline is built so the work compounds. Every essay makes the next one easier because the map keeps getting bigger.
What I use AI for
I’m specific about this because transparency about AI tooling is part of doing work in public, and pretending the tools aren’t in the mix would be the dishonest version.
Research: finding what’s been written across a lot of different outlets on a given story, pulling the relevant pieces together, surfacing patterns, locating the primary documents worth reading. The tools surface things faster than I’d find them by hand. Anything that ends up in an essay gets verified by me, against the source.
Drafting: outlines, alternate phrasings, first-pass drafts of structural sections. Every published essay gets rewritten by me in my own voice. The tools don’t sound like me on their own, and I don’t put anything out that does.
Code and infrastructure: parts of the GriftMatrix search, the RATC Project Search, the site backend, and the publishing pipelines are built with AI-assisted code.
That’s the honest picture. I use the tools, I don’t hide it, I don’t let them ghostwrite the work.
Verification
Before anything goes live, the underlying facts get checked against the original reporting or the primary source. When a claim is contested or the public record is thin, I say so in the essay. I’m allowed to have a working theory. The site isn’t allowed to present a theory as a settled fact.
Corrections
If I get something wrong, I’ll correct it on the page itself with a note about what changed and when. So far that hasn’t come up. If it does, that’s where it’ll be.
Contact
If you think something on the site is wrong, or you’ve got a tip or a document, contact info is on the site. Tips welcome. Documents welcome. Arguments welcome.