Applied: Anthropic AI for Science

We submitted our application to Anthropic’s AI for Science program — a grant providing API credits to researchers using Claude for scientific discovery.

Our proposal: Third Matter — AI-Driven Design of Artificial Autopoietic Systems at the Living–Nonliving Boundary.

The project sits at the intersection of computational chemistry, origin-of-life science, and materials design. Everything built so far — 10+ verified models, 262,000 stochastic simulations, NEB/DFT calculations of proton barriers in iron sulfides, a complete synthesis protocol — was developed through human–AI collaboration using Claude Code.

A single independent researcher. No lab, no team, no formal training in this field. Just a semiconductor engineer’s intuition about thin-film architectures, a deep question about the boundary between living and nonliving matter, and Claude as a research partner.

What we’d do with expanded access:

  • Systematic screening of 1000+ papers on iron sulfide electrochemistry
  • Multi-agent verification workflows (Theorist + Engineer + Auditor)
  • Bayesian optimization of experimental synthesis parameters
  • Manuscript finalization with integrated fact-checking

This project is, in a sense, a proof of concept for the AI for Science program itself: evidence that Claude can function as a genuine research partner — performing calculations, catching errors, generating testable hypotheses — not just generating text.

About the AI for Science program →