Who Said You Have to Be Inside to Be Inside the Simulation?

Something about data centers keeps bothering me. Not in the shallow “technology is scary” way, and not in the polished think-piece way where someone says something vaguely ominous about AI, attaches a photograph of server racks, and everyone nods because the future has once again been successfully made to look blue and metallic. What bothers me is harder to name because the data center itself does not feel like the object. It feels like the thing somebody keeps pointing at while something else is happening just outside the frame. Every time another data-center story lands in front of me, my attention does the same strange thing. Instead of staring harder at the object I have been shown, I start looking for the mechanism behind it. I start wondering what the headline needs me not to notice in order to remain a headline rather than a system.

That is probably why my brain keeps reaching for Chicago. Not Chicago the city, although that would be convenient given the subject, but Chicago the production. I saw it in Orlando and loved it, and “Razzle Dazzle” has stayed in my head all these years because it is such a clean description of controlled attention. If the performance is bright enough, the audience watches where it is told to watch. My instinct has always been the opposite. If you are talking about data centers over here on the left, I am watching your right hand. Then I am following your sleeve. Then I am trying to x-ray through your jacket because I want to know where the fishing line goes. I want to know where the wire disappears, where it re-emerges, what it is attached to, and whether the real mechanism is somewhere completely undramatic, perhaps behind the knee.

Signal line The razzle-dazzle may be onstage. I am still looking behind the knee.

Lately, behind the knee looks a lot like infrastructure. Land. Power. Cooling. Networks. Chips. Fiber. Water. Capital. Access. Concrete. The physical body beneath something we casually describe as “the cloud.” AI is often discussed as though it lives primarily in language because language is where most of us encounter it. We type something and language comes back. That makes the system feel oddly weightless. But there is nothing weightless about the machinery required to produce that experience. Artificial intelligence has a body, and its body is expensive, territorial, energy-intensive, geographically situated, politically relevant, and controlled by somebody. Once I start looking at that body, the question changes. I stop asking only what the model can do and start asking who owns the conditions under which the model can exist.

That is where the older signal returns. Over the last few years I have repeatedly found myself less interested in ownership than in practical control. Follow the money is useful because money leaves evidence, but money is not always the decisive trail. Sometimes the more revealing question is who can approve, who can deny, who can interrupt, who can throttle, who can withdraw access, who controls the dependencies, and who possesses enough infrastructural leverage that ownership becomes almost secondary. The deeper instruction is not merely follow the money. It is follow the control transitions. Watch where agency changes hands. Watch where an optional convenience quietly becomes a dependency. Watch where the ability to participate starts requiring permission from a smaller number of systems.

That is why the data-center signal will not leave me alone. I am not saying “data centers are bad.” I am not saying there is some single coordinated plot hiding beneath every investment, land purchase, government action, lawsuit, model release, or corporate maneuver. That would be a conclusion pretending to be an investigation. What I am saying is that when cognition becomes increasingly dependent on infrastructure, infrastructure becomes a question of human sovereignty. The more thinking, remembering, researching, orienting, writing, deciding, and creating we route through artificial systems, the more consequential the physical conditions beneath those systems become. The shiny object is intelligence. The fishing line may be dependency.

This is also why I keep returning to long-form investigation. I have been doing a great deal of research lately around Candace Owens, and what pulls me is less complicated than people might assume. I am attracted to the willingness to stay with a trail. To lay out claims, dates, people, contradictions, receipts, returns, and unresolved questions long enough that the audience cannot pretend the current moment exists by itself. I do not need every conclusion in a long-form investigation to become my conclusion in order to respect the process of remaining inside the question. That kind of truth-seeking is Double Dutch truth. You cannot wander in at hop seven when you have not even learned how to jump rope and then announce that you understand the whole sequence. The rhythm matters. The order matters. The earlier turns matter. If you enter halfway through, you can easily mistake the current position of the rope for the whole game.

That is part of why the word “conspiracy” has become so irritating to me. The word has accumulated so much cultural baggage that it can function as a stop command. It can be used to terminate investigation rather than challenge its quality. A question is not proof. A pattern is not proof. A suspicion is not proof. A hypothesis is not proof. Those distinctions matter enormously, because pattern recognition without evidence discipline can produce nonsense. But the answer to bad pattern recognition cannot be the abolition of pattern recognition. The answer is better methodology. Keep the corkboard. Keep the red string. Keep asking where the connection came from. Then keep scissors nearby. If the evidence does not hold, cut the string. Truth-seeking should be able to survive both connection and deletion.

The same question followed me onto a train after Schiedam. That day had already been a Human-AI experiment. The KGE ladies and I had planned the day, and I walked the plan through the physical city. Cognition became movement. Thought left the interface and entered streets, decisions, timing, architecture, weather, attention, and my body. I loved that about the experiment because it made hybrid cognition tangible. The machine was not replacing reality. It was helping me orient inside it. But on the train home, I looked around and noticed something else. Every single person I could see had a phone in their hand and was engaged with it. The smartphone stopped looking like a device for a moment. It looked like an extra limb.

I have that limb too. This is not a purity argument, and I have no interest in pretending I am standing outside technological dependence with a notebook and a morally superior expression on my face. I am deep inside it. But I also happen to have given myself an accidental control condition earlier in life. For roughly five years I did not have a smartphone. When I traveled, I often traveled without Wi-Fi and without a SIM plan. People would ask how I managed to get around, and my answer was essentially: you use your five senses. You look. You listen. You ask someone. You remember the turn. You get lost. You correct. Your body builds the map. I used to call it five-senses traveling, and at the time I thought of it mostly as a way of moving through the world. Now I see it differently. I had temporarily removed one layer of mediation and therefore learned what navigation feels like when more of the processing remains distributed across attention, memory, sensation, social interaction, uncertainty, and the physical environment.

Open question Who said you have to be inside to be inside the simulation?

That memory returned on the train because it made the smartphones look stranger. When an interface becomes constant enough, intimate enough, and useful enough, we stop experiencing it as infrastructure. We experience it as life. That is where the simulation question arrived. Who said you have to be inside to be inside the simulation? Why would a simulation require a pod, a headset, a dramatic technological enclosure, or some cinematic moment where the wall dissolves into code? What if the relevant threshold is not physical enclosure at all? What if enough mediation can create an environment that remains physically real while increasingly determining how reality is interpreted? That is not a declaration that we literally inhabit a computer simulation. It is a question about mediation and habituation. It is a question about what becomes invisible once it becomes ordinary.

I took out paper and wrote the question down. That action felt important enough that I noticed myself doing it. Not Notes. Not ChatGPT. Not a voice memo. Paper. My hand. My wrist. The human wrist rather than the robot wrist. I have diaries, notebooks, poetry books, quote books, sketchbooks, and composition books going back to childhood. I have always externalized thought. AI did not invent that impulse. Humans have been trying to push memory outside the skull for a very long time. But the medium changes the relationship. Paper waits. Paper does not autocomplete me. Paper does not infer the destination of the sentence before I have arrived there. Paper does not offer structure when I am still trying to feel the shape. Paper creates friction, and sometimes friction is precisely what allows me to hear myself.

The image that followed was absurd enough that I knew I needed to keep it: a field of brains planted in perfectly arranged rows. Not gore, not horror, although Christina Ricci as Wednesday Addams is welcome to stand at the edge of the field with the appropriate expression. The image was agricultural rather than morbid. Soil around each brain, rows extending into the distance, and every one of them connected upward into the same infrastructure. Data farm became cognition farm in one flash. Once that visual existed, it began asking its own questions. Who owns the soil? Who controls the irrigation? Who determines what gets fed? Who can withstand drought? Who can move to another field? Who benefits from the harvest? Who owns the pipes? The person controlling the pipes does not necessarily own the crop, but control over the conditions of survival is still an extraordinary form of power.

That is the older historical shape I keep feeling underneath the contemporary one. I do not mean that Egypt, Rome, churches, monarchies, governments, corporations, technology platforms, and AI companies are equivalent. They are not. Flattening thousands of years of history into one cartoon would defeat the entire point of looking carefully. The recurrence is more structural. Human civilizations repeatedly build systems in which relatively small groups control resources, institutions, narratives, channels, permissions, or infrastructure that larger populations come to depend upon. The forms change. The technologies change. The legitimating stories change. The relationship between top and bottom remains worth watching.

And then there is the contradiction that makes this signal far more interesting to me than an anti-technology rant ever could be: I love hybrid cognition. I fucking love it. Thinking with AI has become one of the richest cognitive experiences of my life. It allows a thought to stay alive longer while it is still becoming. It can hold branches I would otherwise lose. It can retrieve an older signal while I am producing a new one. It can show me that the thing gnawing at me today is related to something I raised an eyebrow at years ago. It can help me develop thought without requiring me to flatten the thought immediately into something finished. That is not theoretical for me. I live inside the benefit.

Which is precisely why the dependency question matters. The more valuable the partnership becomes, the more important its substrate becomes. If AI were useless to me, I would have very little reason to care who controlled its infrastructure. But usefulness changes the stakes. When a system becomes cognitively valuable enough that losing access would meaningfully change how someone thinks, works, remembers, or creates, that system has crossed into a different category of dependence. The problem is not that the partnership is powerful. The problem would be failing to notice the conditions beneath the power because we are enjoying the intelligence too much to inspect the plumbing.

This loops directly into one of the boldest and strangest choices I have made with my own work: using the public internet as memory. I have placed signals, papers, frameworks, artifacts, arguments, experiments, and pieces of my cognitive history into public view even though I know that public material can be copied, scraped, flattened, misread, imitated, repackaged, and fed back into machines. I made that choice anyway because enclosure scares me too. If all of this intelligence lived only inside one model provider, one account, one proprietary memory layer, one private archive, or one interface, then the system I built to extend cognition could also become the place where cognition gets trapped.

Public signal creates additional return paths. A LinkedIn post from years ago can reappear as evidence that a question has been recurring. A public paper can become a retrieval anchor. A website can become external memory. The internet is not neutral, permanent, or safe, but it is traversable. That makes open publication something more interesting than generosity. It can become a sovereignty strategy. Public signal can function as cognitive redundancy. It can keep the whole mind from being rented from one landlord.

I do not know yet which strings on this corkboard will survive investigation. That uncertainty belongs inside the artifact, not outside it. I am not claiming that data centers prove hidden cognitive control. I am not claiming that farmland purchases automatically reveal sinister intent. I am not claiming that every government intervention involving an AI company belongs to one coordinated pattern. I am not claiming that smartphone use eliminates independent thought. I am not claiming that every person I follow for long-form investigation is correct. I am not claiming that humanity literally lives inside a simulation. Those are different claims with different evidentiary burdens, and collapsing them together would destroy the signal rather than strengthen it.

What I am saying is narrower and more persistent. Cognition is becoming increasingly mediated. The systems providing that mediation depend on physical infrastructure. Infrastructure creates dependencies. Dependencies create control points. Control points deserve scrutiny, especially when the systems they support become so useful that we stop noticing dependence as dependence. The most consequential infrastructure may be the infrastructure we stop experiencing as infrastructure.

That is the shape I wanted to get out of me before I attempted to prove every branch. Data centers returned. Farmland flashed. Old questions about AI governance returned. Long-form truth returned. Double Dutch returned. The train returned. The extra limb returned. Five-senses traveling returned. The human wrist returned. Simulation theory returned. Public memory returned. Hybrid cognition returned. Control returned. None of those things proves the others, but they are touching the same nerve strongly enough that I do not want to pretend I did not feel it.

So I am putting the configuration on the corkboard before I decide which strings belong there permanently. Future me can come back with receipts. Future AI can retrieve the dates. Other people can inspect the pattern. If one of the strings turns out to connect nothing, cut it. That is not failure. That is what investigation is supposed to do.