Description

Hi! I've been building an OSINT research tool/framework called Res Ipsa. Res Ipsa is designed for investigations where provenance, source dependency, contradictions, missing evidence, competing explanations, and interesting patterns become difficult to keep separate as the corpus grows.

Though Res Ipsa cut its teeth on another public corpus, the UAP corpus became one of its first serious stress tests due to the significant tranche of files that have come out in the past several months.

I gave it the frozen five-tranche UAP documentary corpus found on war.gov (http://war.gov) and spent the last month testing whether it could investigate that material without collapsing toward either confirmation or dismissal. It also felt appropriate given UAP/UFO research has always been a subject I've enjoyed doing deep dives on.

This is the resulting public report: The Shape of the Record.
https://www.resipsaresearch.org/publications/the-shape-of-the-record-… (https://www.resipsaresearch.org/publications/the-shape-of-the-record-…)

What interests me most isn't whether everyone agrees with its conclusions. The idea of the framework itself is neither to nullify the believers, nor is it to pander to the skeptics. Instead it seeks to clearly map the edges of the known. I want to know where the research is strongest, and also where the weak points are from the perspectives of people who have lived in these files as avid hobbyists and field experts.

If you know these archives and see a source that has been misidentified, an apparently independent account that's actually derivative, a conventional explanation we've underweighted, a connection we've flattened, or one of our “missing” records sitting publicly in an archive somewhere, please point to it.

Res Ipsa mapped 43 formal Null Spaces in this corpus. Some are ordinary documentary gaps. Others sit exactly where a missing radar record, technical analysis, enclosure, dataset, or correspondence file could materially change what can be concluded.

So consider this both a report and an invitation to audit the research. The entire idea of Res Ipsa moving forward is handing serious OSINT researchers a tool for research that is not simply expansive, but defensible and capable of being audited.

This report was produced by me working inside the Res Ipsa research framework using AI as a collaborative scope/tool, rather than inference generator. Within it, the human serves as the soul, the AI serves as the scope. The framework/tool is currently still in the alpha testing mode, so not yet released for others, but will be opening for beta soon. This outreach is more to gather input and thoughts on the reception of the output offered by other enthusiasts and experts.

Tell me what it got right. More importantly, tell me where it breaks.

One other thing I'd genuinely encourage: if you use ChatGPT, Claude, Gemini, Grok, or another AI system, give it the report and ask it to critique the work. Don't prime it with whether you think the report is right or wrong. Ask about the methodology and sophistication. Ask how it mitigates the issue of hallucination. Finally, inquire about where reasoning is strongest, where it is weakest, what appears underweighted or overclaimed, whether the conclusions follow from the evidence presented, and what it thinks Res Ipsa may be missing.

I'm actually interested in seeing those responses too. One of the largest insights I've seen is that different models often notice different failure modes, and part of this public insight expedition is finding out whether the research product remains defensible when it leaves the environment that produced it.

Thanks!