Living in Time: Why AI Is Not Like Us

Made with OpenArt

When you watch a film, it feels like movement is happening before your eyes. The film creates fluid motion through discrete frames at typically 24 or 25 per second. Our brains integrate these discrete elements into continuous perception. This reveals something fundamental about how we experience and construct reality: we perceive continuity where there is none.

Now consider AI. When interacting with a chatbot, we experience what appears to be thinking, reasoning, and meaningful response. Like film, AI operates through discrete computational steps. But instead of frames, it’s tokens and predictions. Rather than dismissing either as “just” sequences, we should examine how discrete elements combine to create emergent properties, whether that’s the perceived motion in film or the coherent reasoning in AI.

AI processes language through hierarchies of contextual relationships. Ask it “What did Jon eat?” versus “Did Jon eat?”, and different patterns of prediction emerge. Meaning arises from patterns over time, much like how motion emerges from a sequence of still images. Both systems demonstrate how discrete elements can give rise to qualitatively different experiences.

However, with film, we can clearly distinguish between actual motion in the world and the illusion of motion in playback. With thought, it’s more ambiguous. Is human reasoning continuous, discrete, or something else entirely? The boundary between process and simulation seems not totally clear.

Enter the Carburettor.

To understand this boundary, let’s introduce a different analogy: the carburettor. In older cars, the driver controls a throttle that opens a valve; air rushes in, drawing fuel into the mix via the venturi effect. The process is purely physical. More throttle, more air, more fuel, more combustion, all tightly coupled in real time. The carburettor is not calculating anything; it just is a process.

Now compare this to modern fuel injection systems. Here, the throttle might not be connected to anything mechanical at all. Pressing the pedal sends a signal to a computer, which then calculates how much air to let in and how much fuel to inject. The physical process is mediated by a model. Rules, logic, and prediction.

In this sense, the difference between the carburettor and the computer-controlled injector mirrors the difference between process and computation, between acting and simulating. The carburettor lives in time. The model simulates, predicts, and chooses before action. The model is made by time.

AI as Drive-by-Wire.

AI is like a drive-by-wire system. There’s no direct causal connection between your input and its response. Instead, a layer of discrete prediction mediates every output. What feels immediate is the result of extremely fast computation. The intelligence we perceive emerges from the system’s ability to create an illusion of understanding, due to the speed of the underlying computations. At a pace indistinguishable from natural interaction.

Humans: More Like Carburettors?

This raises a curious thought: perhaps humans are more like carburettors than we want to admit. We speak and act without foreseeing each step. We don’t calculate the next five words before we say them. Our intelligence is embedded in time and process, not held in an abstract model.

AI, by contrast, holds a static model and samples from it rapidly. Like a film being played back, or an ECU calculating fuel injection, AI operates outside of real-time process. Its identity is not in being but in becoming, activated only when run.

Process or Simulation?

So what distinguishes AI from human intelligence? I think it may lie with the relationship to time. AI and film both reveal that discrete elements can generate compelling experiences when operating at the appropriate speed. But unlike a carburettor, which is continuous and embodied, AI is modular, computational, and latent. We should be careful not to confuse simulation with process, or playback with life. A carburettor doesn’t pretend to mix fuel and air; it does it. AI, like film, convinces us through speed.

Coda: A Meeting in Time.

We may try to pinpoint the distinction with this consideration: film and AI are shaped by time, dependent on its pace to create illusion or interaction. We, on the other hand, generate time through attention and action. We live in time, whereas these systems are animated by it. And yet, when their speed matches our rhythms, something uncanny happens: we meet. A moment in time where simulation and life briefly align.