Searching protocol for "model reduction"
Shrink TLA+ models for faster checks.
Optimize TLA+ specs for clarity and performance.
Cut AI costs without sacrificing quality.
Route tasks smartly, cut AI costs by 30x.
Add new problem models to the codebase.
Propose new problem models or rules.
Compress LLMs for efficient deployment.
Shrink LLMs, boost performance.
Select and optimize embeddings for semantic search.
Aggressive v3 performance targets.
Generate Pydantic v2 data models.
Quantify social & environmental impact, drive ESG.