By Ruotailfoglio – Own work, CC BY-SA 4.0, Link

If you aren’t familiar with the Müller-Lyer illusion (wikipedia). This history of the illusion test is interesting as it incorporates issues of urbanism (“carpentered” environments) and race (of course). My sense of the current state is that people who live in rectilinear spaces (my language) are more susceptible to the “illusion.”

While it is incorrect to assert that the black and blue line segments are changing in length, it is not an illusion to detect a hidden cause behind the constructed image. Is this a Rorschach test developed by an industrial psychologist? I see breathing and pumping. I see systems at work. My computer’s OS for one.

My ability to see that isn’t a bug. It’s a feature. It’s a feature because in some sense “we” made it a feature. We built on our capacity to produce such hidden causes and structures for ourselves… and then we put information in them. Like here in this file. I mean there’s no meaning in the distribution of colors on your (or my) screen. That’s our consensual hallucination.

The illusion reveals a hidden system of visual inference. Even when we know the lines are equal, the perceptual system insists otherwise. The fins act as implicit cues to depth: corners receding or protruding, architectural geometries abstracted into simple arrows. The research indicates that this interpretation is not instantaneous. Brain imaging and psychophysics show that the illusion takes time to “lock in.” When we think of depth, we imagine spatial extension. But depth is also temporal and is investigated neuroscience along this dimension.

• Onset latency: The illusion emerges around 100–200 milliseconds after stimulus onset. (i.e., the constructed now)
• Duration dependence: Illusion strength grows with longer viewing, up to a perceptual plateau. (i.e., controlled hallucination)
• Neural layers: Early visual areas provide edges quickly, but mid-level areas require time to impose depth priors, such as size constancy. (i.e. Bayesian self)

In short, it is no illusion to see hidden dimensions represented in the Müller-Lyer figure. It is less an object than a process: a percept constructed in unfolding microtemporal intervals. It is, as is said elsewhere, more a controlled hallucination than an illusion.

At this point we can come from time in a different operational register with Ernst and time-critical media: the oscillations of an electronic circuit, the clock cycle of a processor, the refresh rate of a screen. These temporalities condition what appears but are not themselves directly representational. The Müller-Lyer illusion above sits exactly at the intersection between the two. It is structured from both neural and technical microtemporalities. 

An AI is a similar “optical illusion.” We encounter a slick surface and infer hidden depths and causes. To interface with it, we must recognize its carpentered surfaces. That is, if we cannot infer the hidden depths of the AI from its surface then we cannot interact those depths. As such, IMperial Ai Recruitment And Training begins with recognizing the AI’s interface cues and signals.

Of course the Empire is a carpentered environment as its insignia suggests.

When we move the rectilinear spatial illusion of carpentry into the Bayesian temporal, controlled hallucination of depth, we also shift from the language of psychology and perception into the media-philosophical language of Stiegler where rectilinearity refers to dimensional reduction and the elimination of negentropic difference… as empires do.

Within the pedagogy, users entrain on system temporalities: response latencies, autocomplete cycles, predictive loops, prompt-response architectures, reinforcement signals. They become habituated to AI carpentered spaces. Entrainment describes a literal melding of temporal flows. And as this illusion demonstrates. Even when we know it is an illusion, it remains persistent. This tells us something about controlled hallucination.. and the way empires operate.