In this episode of 42章经, Hao Zhe, creator of the 30k+ star Vibe-Trading project, discusses his philosophy on AI trading. He argues that in finance, decisions are cheap and plentiful, but *actions* are expensive and require human responsibility. The core value of AI is not to generate more trades, but to build a *verifiable and auditable bridge* from cheap decisions to costly actions, augmenting subjective investment with transparent research.
Summarized by Podsumo
AI trading is not about predicting prices or fully automated trading; it's about creating a *verifiable channel* between cheap decisions (AI analysis) and expensive actions (human trading).
The biggest shift in the Vibe-Trading project was from a tool for quants to an AI-powered *workbench* for all investors, focusing on explainability over black-box signals.
A key insight: The value of *exclusive data* is decreasing; the real edge comes from *data alignment*, hypothesis verification, and risk management (Harris).
AI trading is a *reflexive* game; strategies are quickly arbitraged away, requiring constant adaptation. The focus should be on combining conditions across multiple data points, not single signals.
Even if AI models become perfect, humans must remain in the loop to maintain *agency, responsibility, and a sense of purpose* in their work and life.
For beginners, the first step is to define your risk profile, set up a separate account, and use AI to explore markets or strategies you haven't tried before.
"In finance, decisions are cheap and almost free; what's truly expensive is the action. AI trading is about building a verifiable and auditable channel between these two."
— Hao Zhe
"The role of data is not to give you the answer, but to let you verify whether your hypothesis is right or wrong. Many times, a 'bad idea' just lacks the right data."
— Hao Zhe
"The most important question for the future is not whether AI can replace humans in trading, but whether humans need to be involved to maintain a sense of direction and meaning in their lives."
— Hao Zhe