Waiting for search
No experiments yet
Run a topic search, then open this layer.
This panel will stay empty until we can propose actionable experiments from the current signal.
Prediction accuracy
0.000
estimated from signal quality
Training data --
Model --
Layer
Click a layer for single view. Press and hold until green to combine layers before search.
Cause relation
Effect = lighter shade of cause
Parallel event
Cycle recurrence
Line count/thickness = repeated evidence
Node size = predicted attention
Dated newspaper layers
Parallel events at one present moment
Dates are coordinates for comparison, not a claim that reality moves in a straight line.
Machine learning layer
Automated model tournament
Train on old matched observations, test against later old observations, then allow a prediction only when the model passes.
Search a topic to test models.
Simple ML pipeline
Historical train/test gate
Uses old search data twice: earlier windows train the model, later windows test whether it predicts reliably enough.
Collect old data
Train models
Test holdout
Gate prediction
Waiting for ML search
Select ML, then search a topic.
Theory layer
Counterfactual experiments and next-test ideas
Builds experimental hypotheses by combining historical evidence with Law and dictionary context.
Search a topic to generate experiments.
Theory workflow
Loose-End Discovery Lab
Theories are experiments. They look for where current models can break and what to test next.
Law layer
Cross-domain expert review
Collects medical, economic, social, security, and scientific confidence checks before accepting a forecast.
Search a topic to wake the doctors.
Review method
Ranked Domain Review
The strongest domain rises first; weak domains stay on standby.
Master Dr conclusion
Waiting for search
Select this layer after a topic search to run domain doctor reviews.
Requires a stable evidence pass before a definitive master conclusion appears.No law ranking yet
Open Law after a search to see doctor cards and predictions.
0%
fit
Archive focus
Loading dated layers…
Cause relation
Effect = lighter shade of cause
Parallel event
Cycle recurrence
ML prediction trend
Line count/thickness = repeated evidence
Node size = predicted attention
Semantic scan layer
Dictionary of observed relations
Choose a letter, then a word. Smaller words show its strongest description-level relationships.
Load the corpus to begin.
A
Generic concepts share equal root size
Selected word
Choose a word
Dictionary data has not loaded yet.