# AI Capability Ledger

> Source-linked dated observations of capability advances and unresolved questions.

Canonical: https://ainextia.com/capability-ledger/
Publisher: https://ainextia.com/

## Evidence records
- 2026-11-05 — AI-assisted model R&D acceleration (Independent analysis): AI tools are helping frontier researchers complete parts of model development faster Unresolved: Will the acceleration persist across labs and the most difficult research stages? Original source: https://metr.org/blog/2026-11-05-anthropic-ai-rd-acceleration-assessment/
- 2026-11-04 — Scaled automated vulnerability investigation (Company reported): A controlled agent pipeline can triage and reproduce numerous security findings in an authorized research environment Unresolved: How many fixes are accepted, deployed, and independently validated? Original source: https://openai.com/index/a-thousand-fewer-paths-to-root/
- 2026-11-04 — AI-assisted laboratory experimentation (Company reported): Models can assist in planning and data processing across stages of a biology experiment Unresolved: Will productivity gains reproduce in other labs? Original source: https://www.anthropic.com/research/investigating-vault-particles-with-claude
- 2026-11-02 — Proof of inference research (Research prototype): Cryptographic methods can potentially verify model identity without sharing weights Unresolved: Can the computational overhead become practical? Original source: https://www.anthropic.com/research/provable-inference
