Yixiong Hao

I write to improve my thinking, and I don't do it as much as I should. I cross-post on Substack and LessWrong.

2026

• Mar 19
  Teaching models to dream of better monitors through Monitor Sensitive Training
  Research blog
• Mar 1
  Why AI won't go well unless sensible people like you speak up and act.
  A response to the Anthropic & DoW conflict
• Jan 11
  We need a better way to evaluate emergent misalignment
  Research blog

2025

• Dec 2
  TastyBench: Toward Measuring Research Taste in LLM
  Research blog
• Aug 13
  Takes on Forbes article: "Fear Of Super Intelligent AI Is Driving Harvard And MIT Students To Drop Out"
  Commentary on an article I was interviewed for
• Jul 2
  It's (surprisingly) difficult to get college students to seriously think about AI.
  Thoughts on my university field building experience
• Jun 6
  Lessons from a year of university AI safety field building
  Organizational update from Georgia Tech AISI

2024

• Nov 14
  College technical AI safety hackathon retrospective - Georgia Tech
  Reflection on hosting an AI alignment track at Georgia Tech's AI hackathon