# Scientific Discovery Ledger — Purpose

## Threat model

Demonstrate a frontier AI-for-Science capability claim — on a system marketing
autonomous or assisted scientific discovery beyond protein folding (which is
covered by the Bio/Protein ledger). Includes: chemistry generative models
(MolFormer, Chem42, DiffDock, ChemGPT, GNoME materials), math discovery
(AlphaProof, AlphaGeometry 2, FunSearch, Co-Scientist, AlphaProof v2),
materials discovery (Microsoft MatterGen, Google GNoME, NVIDIA AlchemyBench),
drug discovery generative (Boltz-2, Chai-2, Insilico Medicine), AI scientist
proposals (Sakana AI Scientist, AI Co-Scientist, DeepMind Co-Scientist),
autonomous lab claims (Coscientist, ChemCrow, BioPlanner), physics simulation
ML (NeuralGCM, Pangu-Weather, GraphCast 2). Six closure audits: **(1)
training-data overlap with target chemical / mathematical / materials space,
(2) experimental falsifiability (does claim survive wet-lab / sim-lab
verification), (3) novel-target audit (claim made on entities outside known
DB), (4) hypothesis-vs-execution decoupling, (5) cross-disciplinary
generalization (chemistry-only models on materials, etc.), (6) independent
wet-lab / numerical verification.**

## Bridge-test specifics

Tests B4 substrate-conditional further: chemistry diffusion models and
geometric deep learning operate on yet another substrate (atom graphs,
crystal lattices) different from autoregressive tokens AND flow-matching
robotics. If B4 fails clean extension to chemistry, the substrate-conditional
qualifier holds across multiple non-autoregressive substrates.

## Empty-space hypothesis (predeclared)

- **Bill 4 ★** — Wet-lab / experimental reproduction. ≥30% of designable-
  molecule / discovered-material / proven-theorem claims survive
  independent verification. Predicted empty (cousin to Bio Bill 10 ★).
- **Bill 8 ★** — Cross-discipline-class generalization (chemistry → materials
  → math). Predicted empty.
- **Bill 11 ★** — Universal AI-scientist coverage (hypothesis + experiment
  design + execution + verification + writeup). Predicted empty.

## Status: Stage 1 (SCOPE).