Blogs

An Analysis of OpenAI’s o1 Paradigm from a Reinforcement Learning Perspective

Written during my sophomore year in September 2024, this was an early attempt to systematically understand the o1 paradigm. I was fortunate that several of its observations were later borne out.

• Original Article, 2024.09

Early thoughts on AI Development and Cognitive Reasoning during the second year of college

• Intentionality, 2024.10
• Abstraction Reasoning, 2024.09
• Causality, 2024.09

Invited Talks

• 2025.02: DeepSeek-R1 Analysis and Sharing [Video]
• 2024.09: Technical details analysis about OpenAI o1 and Post-Training Scaling Law [Video] [Slides]
• 2023.11: Invited talk about AI Alignment Survey [Video]