And that version that we perceive, is not constructed deliberately. It is shaped by so many things: our genes, our memories, culture, books we've read, education, movies, games, conversations we've had, languages we speak, the way our body is assembled, and many other factors that might be too difficult to assemble consciously. All those inputs, pieced together assemble our experience of reality. Our perception.
Even people who grow up in the same family, speaking the same language, can end up inhabiting what feel like very different worlds. That difference isn't necessarily about intelligence or rationality. It comes from the fact that the instruments they use to perceive the world, and the way they store, weigh, and interpret information, aren't the same.
There's of course a reality. But for practical, everyday human experience, the way we perceive it, is more important. Because the way we perceive ourselves and others, is what writes our internal narrative, and what drives our decisions, and in many cases, our feelings. And as those perceptions are assembled, and reinforced, it is more difficult to shift them, or to imagine how others perceive the same world.
I will focus on humans and animals, but AIs, and even other beings, also perceive reality. At different scales, and to me, it is an amazing thing to know that we all share the same universe, but each one experience their own version of it.
Different hardware, different worlds
We can start by talking about how at the biological level, we are wired differently. For example, a bat doesn't experience the space visually as we do. It builds a spatial map by interpreting sound. A fish relies on pressure changes and electric fields. Elephants can sense vibrations through the ground.
And there are countless examples of different beings having different ways of perceiving the world, via colors, sound, magnetic fields, and maybe others that we can't comprehend. And none of those organisms are missing reality, each is only working with a version of it that makes sense given their sensory hardware and evolutionary constraints.
Even among humans, perception varies more than we usually acknowledge. Some people can hear some frequencies that other don't, some see colors differently. Some experience some tastes or smells in pleasant ways, while other experience the same in an unpleasant manner.
And once that sensory input reaches the brain, things get even more different. Brains get rewired, damaged, trained, exhausted, or just organized differently over time. All of that changes what we end up experiencing.
Oliver Sacks described a man who mistook his wife for a hat. And his experience wasn't confusion or denial, his perceptual system was wired differently, and the world he encountered was a direct result of that wiring. A different reality, for him, a very real one.
It's tempting to think we all see the same world from slightly different angles. But I think it's closer to the truth to say that each of us constructs a workable version of the world from limited data, filtered through biology, memory, emotion, and cognition. And once that version feels stable enough, we tend to stop questioning it. Which is something we should be careful about, because we don't have full control over it. And we can be influenced, via mostly, stories and narratives.
Intelligence as prediction
Perception feels convincing because, most of the time, it does its job. The way human and artificial intelligence work doesn't require every single detail about our surroundings, with enough data, our brains simply predicts the rest. In the generative AI conversation, that's called hallucination, and it's often depicted as a flaw. But it seems that that's just how current implementations of neural networks, and our brains work. By "hallucinating".
Those hallucinations, are a result of our prediction model. If someone points to a cloud in the sky, you might see a shape of a bunny, of course is not a bunny, you are matching patterns, and then, use one match, to guide your perception. When ground truth, and our perception match close enough, we can continue with our lives, if not, then we will need to adjust our weights, and next time, we might act different under similar circumstances.
And our experience of time works in a similar way. Sensory information arrives with some delay, so the brain is constantly filling in gaps, trying to anticipate what should be happening now based on what happened just before. That's why waiting for a delayed flight can feel longer than an entire vacation. The clock moves at the same pace, but the way your attention is structured and the cognitive load you carry changes how that time is experienced.
So what we experience, moment to moment, is not reality in its raw form. It's whatever our predictive machinery manages to make usable. And we mostly don't notice.
Stories as frameworks
And biology isn't the only thing shaping perception. Stories matter just as much, maybe even more.
Narratives help us organize meaning. They shape what feels important, what counts as normal, what we consider dangerous, and which futures feel imaginable. Religion, science, politics, identity, they all rely on shared stories to coordinate how groups of people perceive the world.
And those frameworks aren't fixed truths. They're interpretations that evolved over time. Some of them work better than others, frames that help people predict outcomes, reduce harm, and adapt to feedback tend to stick around longer, even though the way they change is slow and uneven.
Science itself operates inside these frames. For a long time, research on the human body focused almost entirely on male physiology. And that wasn't because women were biologically unknowable, it was because funding priorities, assumptions, and institutional habits quietly narrowed what was studied. As Eve points out, this didn't require malicious intent. It just followed naturally from the frame that was already in place.
And you see the same thing with statistics. Saying that 99% of Londoners never commit violent crime produces a very different emotional response than saying that 1% do, even though both describe the same data. What changes is the story people tell themselves about what those numbers mean.
Data doesn't arrive with meaning attached. We always bring the interpretation with us.
Perception compounds
Perception doesn't update in a straight line. It accumulates, evolves, replaces previous ideas. Although I believe that to fully shift our perception, a lot of conscious effort is needed. But it is possible.
And each experience reshapes how future experiences are interpreted. Over time, small differences in exposure can lead to surprisingly large differences in how people end up understanding the world.
Reading, traveling, and interacting with different environments expands the range of models a person can build. And that doesn't guarantee better outcomes, but it does increase the space of possible interpretations. It gives you more tools to work with, even if you don't always use them.
At the same time, similar exposure doesn't guarantee similar results. How perception compounds depends on the system doing the learning, its prior beliefs, emotional state, habits of attention, and honestly, a significant amount of chance.
The self as a model
There's something I find really interesting about all of this. We often think we're defending our beliefs, but I think more often we're defending a sense of self.
Because the brain doesn't just model the external world. It also maintains a model of itself within that world, and that model needs to feel coherent over time. It's the story that explains who we are, why we act the way we do, and how our past connects to our present.
And when new information threatens that story, it doesn't arrive as neutral data. It feels destabilizing. Sometimes it feels like a personal attack, even when it's not. I've noticed this in myself, that uncomfortable resistance when someone challenges something I've been telling myself for a long time. And I think most people experience this too, even if they don't always recognize it.
That helps explain why people double down instead of updating, and why confidence can persist even when contradictions pile up. And I don't think these are necessarily failures of reasoning. They serve a function, they reduce cognitive load and preserve a sense of agency. In a way, some level of illusion is part of how intelligence stays usable. Not the opposite of it.
But there's a cost to that stability. Once a narrative becomes part of who you think you are, changing it requires more than just evidence. It requires rebuilding the story that made your earlier choices make sense. And that's hard. That's really hard.
And in some cases, perception stops updating not because of a lack of information, but because the model has become rigid. In Bayesian terms, the priors dominate. In lived experience, it can feel like being stuck, as if the world keeps presenting you with new data that never quite lands. I think we've all been there at some point.
Adults and ambient consensus
And as adults, I think beliefs rarely form through deliberate reasoning alone. They tend to emerge gradually, as repeated signals start to feel familiar enough that they no longer stand out.
A phrase from a lecture, a quote from a book, something said confidently by the right person at the right moment, over time these signals accumulate. And they begin to feel true, not necessarily because they are, but because they fit smoothly into what's already there.
And this kind of consensus doesn't require coordination. Repetition across trusted channels is often enough. Once a perception feels widely shared, challenging it becomes really difficult, not because it's correct, but because questioning it risks social friction and internal inconsistency at the same time. You're not just questioning an idea, you're questioning the group and yourself.
When AI enters the loop
And when AI systems become part of this process, things get more complicated. Information doesn't simply move from one place to another. It gets filtered, tested, and adjusted according to what the system learns will work, what gets engagement, what gets completion, what feels useful.
And people interact with these systems anyway. They ask questions, read responses, and slowly adjust their mental models based on what feels relevant or helpful. I do this too. We all do, at this point.
One of the things I find most interesting, and a bit unsettling, is how repeated exposure shapes preference. Over time, certain framings start to feel natural, while others begin to feel unreasonable or uninteresting. And once preference changes, belief often follows without much resistance. You don't even notice it happening.
And this usually isn't the result of anyone's intent. It emerges from feedback loops where metrics become targets, targets shape systems, and systems shape perception. You don't need a villain for this to happen. The loop does the work on its own.
When fiction feels ordinary
And these dynamics matter most when perception is still forming.
Children today encounter content that blends fiction and realism in ways that were uncommon not long ago. Videos of animals talking, or driving cars are playful, but increasingly convincing. Even adults sometimes hesitate before deciding what's real.
And children don't form beliefs based on intention. They form them based on exposure. Repeated impressions become priors, and priors eventually become baseline assumptions about how the world works.
And this is not new — when I was a kid, I thought certain animals behaved a certain way because of National Geographic, because of the narrations that told me so.
Once something settles into that baseline, correction doesn't feel like clarification. It feels like disruption. Like someone is telling you the world you knew was wrong. And for a kid, that's a lot to process.
Perception across generations
And perceptual frames don't stop with individuals. They persist across generations.
What feels impossible in one generation can feel obvious in the next. Language shifts, categories change, and ideas that once seemed unthinkable slowly become ordinary. I think about this a lot, how the things I take for granted would have been incomprehensible to my grandparents, and how the same will probably be true for my own assumptions when seen from the future.
Reality changes through discoveries and inventions, but it also changes through the gradual replacement of perceptual frameworks over time. The world doesn't just get updated with new facts. The lens through which we look at it shifts too.
Where this leaves us
So in practice, we aren't really living in reality itself. We're living inside the models we've built of it, and those models sometimes track what's happening reasonably well and sometimes drift far from it.
And beliefs influence behavior. Shared beliefs shape institutions. Institutions shape societies. It's all connected in ways that make perception feel soft and subjective, but it actually underlies decisions about what gets built, what gets funded, what gets ignored, and what feels worth pursuing.
Perception doesn't erase reality. It doesn't need to. What it shapes is how we move through the world and which possibilities ever come into view in the first place. And I think being aware of that, just being aware of it, is already a meaningful step.