AI NEXTIA / CAPABILITY LEDGER

What AI can do. What we can actually verify.

An evolving, evidence-based record of demonstrated capability—not a list of predictions disguised as facts. Every entry has an original source, an evidence label, and an unresolved question.

01 / EVIDENCE RECORDS

Watch the capabilities develop.

Starting observations from November 2026. These records are curated manually; they are not a comprehensive live map of every AI breakthrough.

ENTRY 001
Independent analysis

AI-assisted model R&D acceleration

AI tools are helping frontier researchers complete parts of model development faster

Still to establish

Will the acceleration persist across labs and the most difficult research stages?

AI research · METR assessmentOriginal evidence ↗
ENTRY 002
Company reported

Scaled automated vulnerability investigation

A controlled agent pipeline can triage and reproduce numerous security findings in an authorized research environment

Still to establish

How many fixes are accepted, deployed, and independently validated?

Cybersecurity · OpenAI technical reportOriginal evidence ↗
ENTRY 003
Company reported

AI-assisted laboratory experimentation

Models can assist in planning and data processing across stages of a biology experiment

Still to establish

Will productivity gains reproduce in other labs?

Science · Anthropic case studyOriginal evidence ↗
ENTRY 004
Research prototype

Proof of inference research

Cryptographic methods can potentially verify model identity without sharing weights

Still to establish

Can the computational overhead become practical?

AI reliability · Anthropic FlashProveOriginal evidence ↗

WHY A LEDGER?

A release announcement shouldn't disappear into yesterday's feed.

The next step is longitudinal monitoring: repeat measurements, cross-source comparisons and status changes from research prototype to dependable product.

Explore research ↗