This is an essay I put off for half a year. I kept waiting for my AI to gather enough context to help me write it.
If my thoughts, choices, and even the next thing I am about to say were already written into a causal chain I could not leave, what would effort, regret, hope, and creation mean?
Early this year, I listened to WhynotTV Podcast #4, an interview in which guest Jiayi Weng discussed his model-training work at OpenAI. Near the end, the conversation turned to determinism.
Weng said he leaned toward believing that the world is determined. Human thoughts and choices may also sit inside a causal chain that is already fixed. He was not talking about a god or someone writing a script for each person. He spoke in terms of states, rules, causality, and computation.
My first reaction was: how did someone who studies models and computing systems end up talking about determinism?
The word fate had always sounded a little mystical to me, like everything was meant to happen under some hidden plan. I had never been very interested in that idea. Hearing a researcher take free will seriously made it feel like a different question.
I did not fully understand him. Still, the thought took root.
AI gave me an engineering way to approach it
A few months later, I was writing about AI and work. I wondered whether AI, once deeply involved in people's jobs, could extract tacit skills that used to pass only from a master to an apprentice.
Why can a good salesperson, project lead, or operations manager make a judgment so quickly? We call that experience, intuition, or Sense. It used to be hard to teach. People learned by doing the work for a long time.
When AI works with someone over time, it sees more than the final slide deck. It also sees how that person makes judgments, chooses, gives things up, and hesitates.
When those processes, once confined to the mind, start leaving traces, I find myself asking a radical question: what can never be quantified?
If experience, intuition, and Sense can be extracted, what remains in me that could never be computed?
I remembered Weng's interview from earlier in the year. I thought I was beginning to understand how he had arrived at determinism.
LLMs have had a strong effect on how I think about this. They gave me an engineering metaphor. In a very rough account, a language model has parameters formed through training, reads the current context, and predicts the next token.
The strange part is that this step-by-step prediction can produce complex behavior: writing, programming, reasoning, and judgment. After talking with AI for long enough, sometimes it has a good sense of where I am going before I finish the first sentence.
It is hard not to make a dangerous comparison.
What about me?
If I treat a person's genes, body, family, education, experiences, personality, and every piece of information entering the brain right now as one enormous context, could what I call a choice be the next token generated by this system in its current state?
Perhaps I feel that I choose freely because I cannot see my weights or replay the inference that just happened.
The metaphor helped me get closer to the question. I do not treat it as an account of how the human brain works or as proof of determinism. I only began to think that things which once looked almost like a soul might also grow from states and computation.
For the first time, I understood fate as a chain of initial conditions and rules, with no god of destiny writing a script.
The pitch brought me back to what is happening
If everything is determined, why make an effort? If my decision to try is already decided, what does it mean to change my life?
I spent some time following that thought. Whenever it carried me too far, though, everyday life pulled me back.
I go play football.
When I run, the wind passes my face. My feet press into the grass and my body has weight. When I pant, air really enters my lungs. When I kick the ball, I feel the impact against my foot.
In that moment, whether the universe is deterministic seems less important.
Even if everything billions of years ago had already determined that I would be on this pitch today, the wind still crosses my face. I still have to play the game myself.
I get tired and excited. I curse myself when I miss a shot, and I can stay happy for a long time after a beautiful goal.
A theory can explain an experience while the experience keeps happening. Knowing that water is H₂O does not make thirst go away. Knowing that happiness can be explained does not cancel it.
Even if we someday prove that the universe is entirely determined by cause and effect, that explanation cannot live this second for me.
Sometimes I borrow an imaginary view from outside the universe and use it to dismiss what is happening in front of me. But I live here. I can feel the wind. That is real.
Once I read a prediction, it enters the context
Later I pushed the question another step. I found a harder problem: I am inside the system I am trying to compute.
Imagine a machine that can predict me. It tells me, you will raise your right hand in ten seconds.
After hearing that, I could think: fine, I will raise my left hand instead.
A determinist could keep going and say that my decision to raise my left hand was already part of the calculation.
Fine.
But once I know the prediction, it becomes an input. It enters my context and participates in the next version of me.
The future being determined and my being able to read the future in advance are separate questions. Even if the world has one fixed trajectory, I may not be able to obtain a complete script for myself from inside the system.
The moment I read the script, the fact that I read it becomes part of the story.
Whether the script is already written is one question. Whether a character can read it is another. I think of this as a kind of fate we cannot compute from inside.
AI may describe me and affect me
I have started to connect this question with my years of using AI.
I keep asking AI to review, model, and evaluate me. I give it my projects, judgments, confusion, and a large number of conversations. Then I ask: what kind of person am I? What patterns do you see in my thinking? Why do I make these choices?
AI gives me a model of myself. Sometimes it feels frighteningly accurate.
But I can see that model. Once I see it, it starts to change me.
AI tells me I have a clear habit of asking the question behind the question. I look at that description and think it sounds right. The next time I face a problem, I may do it more deliberately.
So is AI describing me, or has it started to help shape me? I cannot say.
I have learned to be wary of especially neat stories about personal growth. AI is good at arranging a life into a smooth line. It can reconnect past events until they all seem to point toward today.
It sounds good. It can also be dangerous. Real lives are rarely that tidy. Worse, once I believe the story, it starts to have causal force.
I keep reminding myself not to let a beautiful story about myself pass for an accurate model. The person called me keeps drifting.
I may not choose all my weights
Only then did I begin to form an account of my own.
I do not really believe there is a self completely outside the past and its causes. If we keep asking why a choice happened, we can almost always find something that led to it.
I can still understand another kind of freedom. A system can observe itself, build a model of itself, then doubt and revise that model.
I remain inside the causal chain, but I am also a node in that chain that can produce feedback.
To keep using the LLM metaphor, I may not choose all my weights, but I can keep changing my context.
That may also explain why, over the past few years, I have chosen things that look rather random: learning something I may not need yet, meeting someone I otherwise would not meet, or doing something without knowing what it is for.
A person who lives in a stable context for a long time may keep generating a more stable version of themselves. I still want to see other possibilities. So I sometimes give the system new input and see what it produces.
Creation can still surprise me
This brings me to a word I love: creation.
I really enjoy making something new. It might be a paragraph, a game mechanic, a new way to say something, or a sudden connection between two things that seemed unrelated.
In that moment I can feel clearly that the world has gained something that was not there before.
Determinism can explain this completely. It can say that all the inputs from the past finally combined into an outcome at this moment. I accept that.
But before the thing appears, I do not know what it is either. To me, it is still creation. It is still a surprise.
I still do not know where quantification ends
If you ask me now whether I believe in Weng's determinism, I do not have a clean answer.
I have a strong intuition that the world may be determined. LLMs gave me my first engineering view of how complex abilities that once seemed close to a soul might grow from parameters, states, context, and computation.
As I began to study how AI extracts tacit human abilities, I also started to think that many things we call intuition, experience, and Sense may simply lack a good description for now.
I still have a radical question: what can never, in principle, be quantified?
I do not know. I want to find that boundary.
As for fate, this is about as far as I have got. Maybe everything was already determined. Maybe it was not. I do not know.
Even if it was, I cannot stand outside the universe. Every attempt to understand myself becomes input to the next version of me. From where I am, life is still unfolding.
I still have to make decisions and live with their results. I still hope, regret, and create.
More than that, I am still living inside it. I feel the wind when it crosses my face. I pant when I am tired from running. I am happy when I score a goal. When I love someone, I really love them. When I lose something, I really hurt.
Those things do not become false because a larger explanation might exist.
I have come to think that fate and living seriously can coexist.
If life is a sequence of tokens that has already been determined, and someone tells me what the next token will be, that sentence itself becomes my next input.
If everything has already been decided, why live seriously?
Because even if the answer already exists, I have not lived far enough to reach it.
On the way there, the wind will still come. I will still feel it.
I want to see it for myself.
---
Source conversation: WhynotTV Podcast #4, with the determinism discussion beginning around 1:52:48.