Using generative AI and natural language processing on multiple UFO sighting databases, we finally have the tools to answer questions like Hynek first posed in 1972
175 strangers filed independent reports about the same thing. 96% agreed on the color. 100% agreed on the sound. I ran the data on the Tinley Park Lights.
My first approach to analyzing 152,000 UFO sightings had a 34% accuracy rate. I almost published it anyway. Here's what went wrong with semantic embeddings, why I pivoted to LLM extraction, and the methodology that actually worked.
I analyzed entity encounters in the NUFORC sightings database. Shadowy beings: 68% of witnesses report fear. Greys: 38% involve abduction. Light beings: more awe than fear.
I analyzed 152,000 UFO sighting reports. Some patterns: triangles almost never appear during the day. Disks are 5x more likely to involve entity encounters. And the majority are silent.
