Randy See

Research, with room to change

From a thought to a better question.

Small, invited studies that turn questions into evidence, and evidence into better questions.

Open experiments in my dashboard →
  1. 01 / THE PROCESS

    Start with a question

    An observation, a possibility, a thought that does not quite fit.

  2. 02 / THE PROCESS

    Make the hypotheses clear

    Ideas from me, other researchers and AI, with their origins and assumptions visible.

  3. 03 / THE PROCESS

    Write the conceptual account

    Explain what the idea claims, what is uncertain, and what could challenge it.

  4. 04 / THE PROCESS

    Run a bounded experiment

    Agree on the method, invite participants, and keep a record of what actually happened.

  5. 05 / THE PROCESS

    Report the evidence

    Write up the method, findings and limits—including negative or inconclusive results.

  6. 06 / THE PROCESS

    Revise the framework

    Let the evidence change APD and the next questions. Then repeat.

A place to try things carefully.

I’m building a small research program around Adaptive Possibility Dynamics: exploring ideas, making them testable, and revising my understanding through evidence.

Experiments begin with invited participants. You keep one account and a reusable background profile, while each experiment has its own questions, agreement and progress. Sign in first; Randy can then enable individual studies for your account. Personal invitation links are also supported.

The first planned study compares theoretical accounts of a person, starting with personality-test results and everyday examples. Participants assess which anonymous descriptions fit them best. The hypotheses may come from me, other people or AI synthesis.

Feeling described is one kind of evidence. It does not, by itself, prove a theory or establish how someone will act in every situation. These experiments are research, not diagnosis.

I prepare AI prompts and responses manually using subscription services. Each study explains which service receives selected answers and how its data will be used. The participant area keeps the steps connected; it does not independently run AI agents.

Questions about participation? Get in touch. You can also read the APD article.