I'm a student in Taiwan with no money, no institution, and no credentials.
But I have a few questions I can't stop thinking about — and I suspect I'm not the only one.
There have been three eras of discovery.
The first belonged to genius. Newton, Einstein, Darwin. One mind, one intuition, one breakthrough. Rare. Unrepeatable. You can't manufacture a genius.
The second belongs to compute. AlphaFold. GPT. Feed the machine enough data and let it search the solution space faster than any human ever could. Powerful — but bounded. Machines can only search inside the space of questions humans already know how to ask.
The third era has no infrastructure yet. It belongs to collective intuition: the idea that the next important insight might be sitting half-formed in someone's head right now — someone with no credentials, no funding, no institution — and what's missing isn't intelligence or curiosity, but a place where that idea can survive long enough to become rigorous.
That's what I'm building.
The academic system is slow, bureaucratic, and allergic to ideas that don't fit neatly into existing categories. AI can now answer almost any known question instantly. But the questions that actually matter — the ones at the edge of what humans understand — still have no good home.
There are people everywhere with serious, strange, half-formed ideas about how reality works. They're in Discord servers, Reddit threads, ArXiv comment sections. They have no place to think out loud with other people who take them seriously.
I wanted to build that place.
Not a Q&A site. Not a paper repository. Not another social network optimized for engagement.
A place where hard questions survive long enough to be taken seriously.
An open collaboration platform for problems that AI can't solve alone and institutions are too slow to touch.
The asymmetry of entry is intentional.
To ask a question, you need to have demonstrated rigorous thinking first — measured by your reputation score (see below). This isn't gatekeeping. It's signal. A question from someone who has spent months stress-testing their own reasoning is a different kind of question than a casual prompt.
To contribute, anyone can participate. No credentials required. Only your reasoning matters.
The core mechanic:
Problems attract contributors who form branches. Each branch develops its own hypothesis, challenges others using first principles, and evolves through adversarial discussion. Branches are not threads — they are competing lines of thought. You don't reply to a branch, you develop it. Or you open a new one.
No one guarantees an answer. The process itself is the point.
Every contribution — every hypothesis, every challenge, every branch — is evaluated against three rules. Not because we enforce them top-down, but because the community scores contributions based on them:
1. State your assumptions explicitly. What are you taking for granted? Say it out loud. Hidden assumptions are where bad thinking hides.
2. Make every step falsifiable. If a step in your reasoning can't be challenged, it's not reasoning — it's assertion. Every intermediate claim must be the kind of thing that could, in principle, be shown to be wrong.
3. Make a prediction. Any serious hypothesis must eventually say: if I'm right, then under these conditions, we should observe this. No prediction means no hypothesis. Just a story.
These three rules mean you don't need domain expertise to evaluate a contribution. A literature student can tell whether a physicist has stated their assumptions. A high schooler can tell whether a step is falsifiable. The quality of thinking becomes legible across disciplines.
Your reputation has three dimensions — one per rule:
- A-score: Your assumptions have been built upon by others. They were clear enough, honest enough, to serve as a foundation.
- F-score: Your challenges have changed the direction of a discussion. You found the hole.
- P-score: Your predictions have been validated — or at least seriously engaged with — by subsequent contributions.
These are separate because different people are good at different things. Some people are exceptional at surfacing hidden assumptions. Some are precision instruments for finding unfalsifiable steps. Some have rare predictive intuition. The platform values all three. The combination of all three, across many people, is what produces rigorous collective thinking.
Reputation gates one thing only: the right to post new problems. This threshold is high by design.
I believe that emergence — the appearance of complex behavior from simple interactions — follows the same mathematical structure whether it's happening in a market crash, a biological system, a consciousness, or a conversation.
I don't know if that's true. It's a hypothesis, not a fact.
But if it is true, then a platform with enough participants, enough complexity, and enough genuine intellectual friction might itself produce emergent results that nobody predicted — including me.
The key word is rigorous emergence. Every large social platform already produces emergence — memes, consensus, moral panics, movements. But those systems optimize for engagement. The emergent behavior reflects that. Thinking Together optimizes for rigor. The hypothesis is that what emerges will be different in kind, not just in degree.
Thinking Together is a tool for attacking hard problems. It's also a living laboratory for observing how human collective intelligence behaves when you stop optimizing it for engagement and start optimizing it for rigor.
Foldit solved a protein structure problem that had stumped scientists for fifteen years. The people who solved it weren't biologists. They were gamers who didn't know it was supposed to be impossible.
I want to build the place where that keeps happening.
These are the problems I want to think about first. They're not here because I have answers. They're here because I've been stuck on them and I think they're worth being stuck on together.
They were not chosen at random. There is a hidden structure between them: Problem 1 and Problem 3 are two sides of the same question, approached from opposite directions. Problem 2 is the substrate that might reframe both. I don't know what will emerge at their intersection. That's the point.
Problem 1 — Design a communication system that is not human
Not a simplified human language. Not a code. A genuine attempt to design a system of meaning-making that doesn't inherit any of the structural assumptions of human cognition or linguistics. What are the minimum conditions for communication to exist at all? What would it look like if we built from those conditions up, without assuming a speaker, a listener, or a shared context?
Problem 2 — Integrate known physics into a fully holomorphic framework
The use of complex numbers in quantum mechanics is treated as a mathematical convenience — a tool that works without a clear physical justification. My intuition is that this is backwards: that reality might be fundamentally complex-valued, and what we observe is a projection onto the real axis. If that's true, the discontinuities and additional assumptions that plague current physical theories might dissolve into the analytic structure of complex space. I don't know if this is right. I want to find out.
Problem 3 — Build a framework for communicating with non-human minds
Not science fiction. A serious attempt to ask: what would it take to establish meaningful communication with an intelligence that doesn't share our evolutionary history, our sensory apparatus, or our concept of language? What can we learn from existing research on animal cognition, information theory, and the structure of meaning that could ground a real attempt at this?
I can't build this alone. I'm a student. I have no resources.
I'm looking for people who want to think about hard things seriously — and institutions who believe that the next important idea might come from someone with no credentials, no funding, and no patience for how slowly the official system moves.
If that's you, or if that's your organization:
This is an open invitation.
Everything here is and will always be open source. No patents. No paywalls. No institutional capture. Whatever gets built here belongs to everyone.
This README follows its own three rules. So it needs to make predictions.
If the three-dimensional reputation system works the way we think it does, then contributors with high A-scores should produce contributions that generate more downstream branches than contributors with low A-scores. Contributors with high F-scores should produce challenges that demonstrably redirect discussion. These are measurable. We intend to measure them.
If the platform itself behaves like a complex system with genuine emergence, then at some point it should produce a result — a connection, a hypothesis, a synthesis — that nobody on the platform predicted or intended. We don't know when. We don't know what form it will take. But we believe it will happen, and we will document it when it does.
If we're wrong about both of these things, this experiment will have failed in a useful way. That's fine. Failed experiments that are honest about their assumptions are more valuable than successful ones that hide them.
If you want to talk, build, or fund this:
Open an issue. Or find me.
Started in Taiwan, 2026. By someone who couldn't stop thinking.