📝 Blog & Thoughts
From Self-Play to PSRO: Why Fictitious Play Still Matters for Understanding Multi-Agent Training
Self-Play is not a single algorithm, but a training loop centered on opponent distributions. This post connects Fictitious Play, Double Oracle, and PSRO through one unified question: how should an ...
From AlphaGo to *Longzhong Dui*: High-Quality Reasoning Under Limited Information
AlphaGo can crush human grandmasters on a closed board, yet it cannot produce a *Longzhong Dui*-style open strategic analysis. This post compares MCTS with human decision-making and argues that the...
From Self-Play to AGI: A Plausible Path and Six Unresolved Fundamental Problems
Self-Play has driven breakthroughs in closed environments, but the leap toward AGI requires far more than scaling. This post outlines six interlocking problems that any self-play system must solve ...
The Dialectic of Simplicity and Complexity: Historical Evolution and Practical Unity of Decision-Making Thinking
Human understanding of the world has always oscillated between two extremes: one is the reductionist impulse to explain everything with the simplest laws, and the other is the systems perspective t...
Welcome to My Blog!
Hello! This is my first blog post.