Write: geo@qiyanword.cn
阅读中文
I am Feng Xiaoping, a researcher. For ten years I have done one thing: taking changes people call "hard to pin down" and breaking them into structures that can be observed, simulated and argued about.
My work did not begin in a business school or a laboratory. It began with a concrete question: why do models that are highly effective in hindsight fail together at inflection points?
In 2018 I turned that question into a method and called it Latent Pattern Computation: do not start from an existing hypothesis; compute the patterns and strategies hidden inside a complex background, so that changes not yet named can surface first. What changed was not the analytical technique but the order of questions — look at what the world is growing into, then decide which theory explains it.
In 2023 I pushed the method to the level of the agent and began forming the Taiji Intelligence framework: the sensing logic of passive detection — receive passively, cross multiple points, infer a position from weak signals — together with a modelling approach we call "three-dimensional resonance" internally, and complex-systems theory, so a system might move from responding to problems toward generating them. The point was not more compute. It was accepting that intelligence is not the approach to a pre-existing truth, but the continuous reorganization of a system in an open world.
In 2026 I organized these threads into a public research line: business intelligence and ecological evolution as the subject, philosophy, theory, tools and agents as one loop — and the process written out in the open, not kept in a drawer.
Most of my time now goes to three things: reading signals nobody has organized, writing patterns as simulations that run, and testing a judgment with the people who actually have to make the decision.
None of them was a change of subject. Each pushed the same question one level further: see the pattern, grow the agent, organize the system.
A method that computes hidden patterns and strategies from a complex background, and a shift from the forecasting paradigm to the generative one. First claim: make the pattern visible before explaining it.
Using the sensing logic of passive detection together with complex-systems theory to move an agent from answering fixed questions toward generating its own — so that it reorganizes itself in an open world instead of approaching a pre-stored answer.
Business intelligence and ecological evolution as the subject; philosophy, theory, tools and agents as one loop, opened to firms and investors who want the method tested.
Markets, organizations and technologies are temporary condensations in a process. Measure them as entities and you measure the residue of the last round. Research has to follow the evolution, not the snapshot.
Every method has a shelf life. Where a paradigm stops working is where research should begin. I do not defend a theory; I answer for whether it can still generate new questions.
An indicator that cannot be questioned is rhetoric. The Wind-Change Index, the scenario framework, the simulation platform — each has to be falsifiable, or it does not belong in a decision room.
Putting these in writing saves both sides three months.
The problem has to be affecting a live decision. Hypothetical topics, or projects that exist to justify a budget, I decline.
Before the first meeting I need data, interview notes and your existing judgment. Without material I can only say generic things.
Deliverables are several scenarios, trigger signals and a sequence of decisions. If you want one answer to "what happens next year", this is not the place.
Sensitive data can be anonymised or withheld, but method and reasoning default to being written up.
Yes. I work in my own name, without an institutional affiliation or a fixed team. When a project needs more hands, I invite collaborators per project.
Methodologically the same thing. The generative market hypothesis needs real decisions to be tested, and firms need models that explain change. The difference is the deliverable: a paper, a whitepaper, a tool, or a workshop.
Only when there is research value. If the ask is an industry report or a strategy deck, I will point you to teams who do that well.
One 60-minute conversation plus a problem memo under two pages. If both sides think it is worth continuing, we design the project.
Firms and investors with a live decision and a willingness to open material; also a small number of academic exchanges and writing invitations. The test is simple: is this question worth a year of serious work?
Two doors: the research agenda and the notes are open to read. When you want to talk, write.