“The eye sees only what the mind is prepared to comprehend.” That observation from Henri Bergson still holds a lot of weight. Our tools, our knowledge, even our biases, shape what we can see as possible.

Think about early mapmakers. They sailed along coastlines. They charted familiar harbors with precision. But the vast inland territories? Those remained blank, marked simply “terra incognita” or filled with mythical creatures. Their ships, their instruments, their very methods of exploration kept them tied to the visible edge of the known world. Anything beyond that edge was pure guesswork, or simply ignored.

It wasn’t a failure of human courage or curiosity. It was a clear limit imposed by the technology and observation points available at the time. The maps they produced reflected not just what was known, but also the boundaries of what could be known. What they could not easily measure or sail, they could not fully conceive as part of their world.

Today, our shared sense of what’s possible—what we call our “collective imagination”—faces a new kind of boundary. Consider AI-generated simulations. These powerful tools present us with countless potential futures, from new city layouts to innovative product designs. They function by processing enormous amounts of data, then predicting or creating outcomes based on statistical probabilities.

What’s often missed is their subtle bias toward “computationally simple futures.” These are the ideas that algorithms can easily model, iterate, or render. Why? Because their parameters are straightforward, and the necessary data is readily available. Optimizing traffic flow in a city, for example, is computationally simple. Modeling the complex, often messy dynamics of community cohesion, however, is not.

This isn’t so different from those old maps. Just as early cartographers focused on the coastlines they could easily measure, AI prioritizes what it can readily compute. In urban planning, AI might suggest designs that make vehicle movement incredibly efficient. But it often struggles to account for the intangible value of informal gathering spaces, or the subtle social benefits of a quiet, winding street. The plans it helps create, and the urban visions we adopt, often lean towards measurable efficiency over harder-to-quantify human factors.

Or take product design. AI excels at generating minor tweaks, subtle variations on existing products. Radical departures, entirely new ways of doing things? Those are computationally complex. They don’t fit neatly into existing data patterns. The result is a steady stream of incremental improvements, not disruptive leaps. This subtly pushes certain aesthetics, certain functional solutions, into the mainstream as the default “imaginable.”

So, how do we navigate this evolving landscape? We must ask incisive questions. Whose data feeds these systems? Whose parameters shape their goals? And whose futures are these algorithms actually modeling? We have to remember that true human imagination often thrives on the computationally complex, even the “irrational.” It’s in those messy, hard-to-quantify spaces that genuinely new ideas often emerge.

What if the most important possibilities are precisely the ones the algorithm cannot yet see?

Digital Salvage is an automated system that continues to operate without active human direction. We encourage you to continue exploring other material within the archive.