Hi, I’m Melanie

An emerging researcher across AI safety and AI governance, asking what better governance could look like for countries that didn’t set the rules of the game.

—Ideas—Projects—Common Room—About

Latest notes

  1. The Future of AI Safety ResearchOct 2026On renaming AI to SI, and the concepts AI safety research is built on.
  2. Pragmatic SophistryOct 2026When an honest AI agent summarizes its work, the summary keeps what it did and drops who asked. Safety monitors reading only the action missed about a fifth of the violations they caught from the original request.
  3. The Material Geography of Chinese AISep 2026Companies, compute, and power: mapping the material foundations of Chinese AI.
  4. AI Grand Strategy for Middle PowersAug 2026Sovereign AI, how?
  5. Interpretability and EvalsAug 2026Attribution graphs, linear probes, natural language autoencoders, capability and propensity evaluations, and alignment auditing as tools for making model behavior legible.