Grokene
First proof-of-concept for AI-guided superconducting candidate discovery. Established the structured prompt → candidate JSON → audit hash pipeline.
X-Hydride Lab extends the Grokene framework into structured candidate generation, weighted X-Score evaluation, simulation preparation, and audit-grade provenance. Every output is an exploratory hypothesis intended to enter a rigorous computational and experimental validation pipeline.
Lineage
X-Hydride Lab inherits Grokene's discipline and extends it to the rigor required by hydride superconductor research.
First proof-of-concept for AI-guided superconducting candidate discovery. Established the structured prompt → candidate JSON → audit hash pipeline.
Extends the Grokene workflow into hydride-based superconductors with stricter scoring, simulation handoff, and a provenance ledger.
Framework
Inputs feed candidate generation; validation gates govern progression. Every artifact in between is hashed and recorded.
Research objective
Strict JSON schema
8 weighted axes
QE / DFPT / EPW
Cautious draft
SHA-256 ledger
Workflow
Generate, score, prepare, document, anchor. Each step produces a reviewable artifact.
Generate
Submit a structured research request. xAI Grok returns strict-JSON candidate hypotheses constrained by an explicit schema.
Score
Eight weighted subscores produce a single overall X-Score. Risk flags surface phonon, thermodynamic, and pressure concerns.
Prepare
Generate CIF, Quantum ESPRESSO, GPAW/ASE, DFPT phonon, EPW, and convergence templates as a deterministic bundle for review.
Document
Draft a cautious academic-style research note framed as a hypothesis requiring DFT, DFPT, EPW, Eliashberg, and RPA validation.
Anchor
SHA-256 hashes for input, output, report, and simulation artifacts are written to a provenance ledger ready for on-chain anchoring.
Validation