Turn passages into structured predictions
Each entry is represented as a prediction with a domain, description, implied mechanism, social role and an initial specificity estimate.
How scores are built
What is measured Specificity, lead time, functional fit, mechanism and societal role.
Compare fiction against candidate technologies
The benchmark combines semantic similarity with rule-based domain and phrase overlap to rank plausible real-world counterparts.
Reward accuracy, detail and anticipation
Final scores blend existence, functional, mechanistic and societal alignment, then adjust for specificity, novelty, lead time, vagueness and redundancy.
Leaderboard
Average adjusted prediction score by author, from 0 to 100. Select an author to open the profile.
Author profile
Domains, subscores and matched technologies
Prediction explorer
Inspect individual prediction-to-technology comparisons. Scores run from 0 to 100.
Select a prediction to view the top candidate match and subscore breakdown.