Write: geo@qiyanword.cn
阅读中文
Feng Xiaoping · Business ecosystems, investment decisions, AI Research & Writing
I study how markets, organizations and agents generate new possibilities through interaction. When forecasts keep failing, I work on the other side: regenerating the question, the option, the judgment.
Markets do not wait to be described. They form, temporarily, in the interactions of their participants. The object of study moves from what is to how it is being generated.
Start from noise, not from a hypothesis: compute the patterns hidden in weak signals, then let agents grow their own responses — latent patterns to Taiji Intelligence.
Wind-Change Index, scenario generation, multi-agent simulation. Philosophy has to become code that runs, numbers that can be read, assumptions that can be argued with in a meeting.
They are not three academic divisions. Theory supplies hypotheses, application supplies evidence, and agents push both into the open world.
Business Ecosystem Intelligence
Firms, consumers and regulators grow their own structures through interaction. I simulate them to find niches, and the threshold where wind-change starts.
Investment Decision Intelligence
From predicting the market to generating opportunity: define a problem nobody has framed, then create options that can be tested.
Cognitive Science & AI
Agents in open worlds cannot live on pre-training. They learn while acting, evolve with people, and let system-level cognition exceed the sum of its parts.
A project usually starts with a vague anxiety — "this industry seems to be shifting". These four moves turn it into something that can be discussed and tested.
No conclusion first. Compute co-occurrence and deviation across job posts, supplier quotes, patents, community language — find patterns that have not been named yet.
Write the pattern as something that can run: who the agents are, what the rules are, where feedback closes. The first use of a model is to be disproved.
Do not deliver a forecast. Deliver three paths that can evolve, each with trigger conditions and early signals, so a decision has a foothold while you do not know.
Review real decisions to calibrate the model. Every deviation becomes input to the next round, and the tool fits the work better each time.
Long essays quarterly, short fragments as they come. Everything is open to read.
Efficient-market theory treats price as a summary of information. The generative view treats price as the residue of interaction. Why rising forecast accuracy can come with falling decision robustness.
Volatility describes how hard prices shake, not whether the structure is being replaced. The index combines entropy, dispersion of behavior and network reconfiguration — and states the three cases where it fails.
Consulting sells certainty; clients need an order of action under uncertainty. After one strategy workshop I replaced the entire deliverable — including the model I had built for six years.
The purpose of intelligence is not to solve the problem it was given, but to keep discovering and posing more essential ones.
Core proposition
I do not write generic strategy reports. The precondition is a real confusion that is already affecting a decision, and a willingness to let research enter the room.
With firms, investors or universities, testing the generative market hypothesis against real data and decisions.
One or two days around a live decision: the full scenario-generation process, and a list of trigger signals to take away.
Enter the decision cycle each quarter for calibration and review, so the organization grows its own judgment.
Replies usually within two working days. If the direction is wrong for me, I will say so and point you elsewhere.