← Back to Tools

History Predictor

Enter any birth moment and find the most similar historical events (post-1500 AD, n=7,132). A neural net maps 10 astronomical positions into a 768-dimension semantic vector and compares them against a corpus of historical event descriptions. See which moments in history resonate with yours.

Note: the historical source is largely Eurocentric — a limitation worth keeping in mind when interpreting results.

Neural Net Version (recommended)

Uses a trained neural net to amplify positional vectors into a richer semantic space — lower loss, better matches.

Classic Version

The original version — uses direct distance comparison on the 10-element positional vectors.