Explore
For people who want to understand the ideas. Start with what is public. Read the frameworks, the notes, the experiment records, and the questions that are still open.
OPEN COLLABORATION · HUMAN × AI
Walk together. Don't follow.
Explore. Question. Build.
From Human × AI Systems and Independent AI Civilization to P0 One Bit Theorem, this work does not ask you to accept an answer. It asks you to inspect the question.
This is not a course, and it is not an apprenticeship. Most of the research, writing, experiments, and thinking is already public. You do not need to become anyone's student to take part. Reading, questioning, checking, redoing, extending, even overturning a hypothesis, can all be part of walking together.
Don't follow my answers. Explore the questions with me.
Three ways · three depths
You do not need to apply before you begin
If a question is worth studying together, we study it. If a system is worth building together, we build it. If you only want to see where the next question goes, start from what is already public.
Discuss
For people who want to challenge or extend the ideas. Good research does not need agreement. A gap, a counterexample, new evidence, or a different model is more useful than assent.
- Public discussion
- Research critique
- Hypothesis testing
- Technical discussion
- Cross-field exchange
Build
For researchers, engineers, founders, organizations, and independent builders. When a question is ready to leave the page and become an experiment, we can make it.
- Human × AI Systems
- AI agents and agent governance
- AI safety experiments
- Human Sensor Quality
- Independent AI Civilization
- P0 experiments, prototypes, and PoCs
Open research map
What Are We Exploring?
These are questions and frameworks already public on the site. The label is the current research status, not a conclusion.
ACTIVE FRAMEWORKIn use
Human × AI Systems
How should humans and AI build reliable systems together?
CURRENT RESEARCHIn progressIndependent AI Civilization
What happens when intelligence no longer needs continuous human direction?
OPEN QUESTIONOpen questionHuman–AI Reality Friction
How do different forms of intelligence negotiate one shared reality?
OPEN QUESTIONOpen questionHuman Sensor Quality
Can humans become better signal sources for increasingly capable AI systems?
ACTIVE FRAMEWORKIn useHuman Hijack Chain
How do attention, cognition, and emotion get taken over, step by step?
OPEN QUESTIONOpen questionP0 One Bit Theorem
What is the smallest reliable information we can get about a deeper reality?
EXPERIMENTExperimentAgent Governance
How should increasingly autonomous agents be observed, constrained, approved, and audited?
The working example is the production incident investigator. It is a prototype, not a complete theory of governance.
Open by Default
Open by Default
Most of the thinking is public on purpose. You do not need a privileged door to begin.
- Research on the site
- Labs
- Long-form notes
- Threads
- GitHub
- Books
- A working experiment
- First The first layer of collaboration is simply paying attention to work that is already public.
- Second The second is questioning.
- Third The third is building something better.
Following a public channel does not create a special relationship with the author. This is about how knowledge moves in the open.
Follow the Work
Follow the Work
You do not need to “join” in order to walk together.
Disagreement
You Don't Have to Agree.
Walking together is not agreement. If a theory can survive only by being agreed with, it does not deserve to be called research.
- Challenge the assumptions
- Reproduce the experiments
- Propose counterexamples
- Bring better evidence
- Fork the framework
- Build an alternative model
Agreement creates followers.
Falsification creates knowledge.
I would rather have a strong counterexample than a weak supporter.
From Reader to Collaborator
From Reader to Collaborator
This is not a ladder you must finish, and there is no formal gate. It is a conceptual path. You can stop at any layer.
- READRead
- QUESTIONQuestion
- TESTTest
- BUILDBuild
- SHAREShare
- COLLABORATECollaborate
Reading is already participation. A good question is participation. One counterexample can matter more than ten new theories. Collaboration can start when the work actually needs it.
When a conversation is useful
When Should We Talk?
Research
You are working on a nearby question, and you have evidence, a model, or a counterexample worth exchanging.
Build
You have a Human × AI, agent, or AI safety problem that is worth building.
Experiment
You want to design an experiment that can be checked or falsified.
Organization
Your team wants to take AI from a concept into production.
A discussion
You are organizing a real conversation about Human × AI, AI systems, AI safety, or future intelligence.
You may not need a message
You May Not Need Me
If you only want to read, follow the research, use a public framework, look at the Labs, check an idea, or track an experiment, you do not need to write first.
A different door
A defined commercial need, a schedule, and a delivery?
Open collaboration
Ideas, research, discussion, experiments, and intellectual exchange. This page is that door.
Professional work
Scoped work: consulting, PoCs, AI integration, AI-first workflow, AI-powered teams, AI safety, and production deployment.
Walk with the questions
Don't follow me.
Walk with the questions.
You do not need to follow me.
Find a question worth pursuing. We will meet on that road.