The Turk

I recently finished Tom Standage’s “The Turk: The Life and Times of the Famous Eighteenth-Century Chess-Playing Machine.” I want to share some points that made me think.

The Turk was a chess-playing machine invented in 1770 by Wolfgang von Kempelen to impress the Austria-Hungary empress. Kempelen’s motive in building the Turk was simple. He disliked another invention presented to the empress so much that he set out to prove “anything you can do, I can do better.” This natural emotion of comptetiveness remains the basis of scientific research today, especially in AI.

For over 85 years, the Turk toured Europe and the Americas, defeating almost everyone. It played against Napoleon, Benjamin Franklin and Philidor. It was exhibited to Catherine the Great and Marie Antoinette. It inspired Charles Babbage and Edgar Allan Poe. For a hoax, its longevity was remarkable. It is hard to grasp how they kept its secret for so long.

Hiding in Plain Sight

And yet, the explanation was simple. Of course there was a human chess master hidden inside the cabinet, Occam’s razor was accurate. Many people guessed the truth, but it wasn’t proven until the last owner of the Turk gave away the secret. People wanted to believe in magic.

That desire has not disappeared. Today, LLMs are often treated as if they reason in a human sense. But there is no such thing. It is just pattern matching on a scale that is so massive, it is beyond our comprehension. Huge amount of computation and data are not hidden, they are out there. But, like the Turk, the impressive interface conceals the comparatively crude mechanism. Deception, in this sense, is not a bug, it is a feature. Of course, the engineering is genuinely clever (for both the Turk and the LLMs), but the magic lies in presentation.

Asking the Wrong Questions

The Turk was an episode in a recurring pattern: the printing press, the loom, the Industrial Revolution, calculators, computers, neural networks, LLMs, and now agents… Each wave arrives sooner than the last, each amplifying human physical or cognitive ability. The curve of development with respect to time is exponential. I feel like we are at a point where the slope is so steep that it seems almost vertical. We are probably at 89° now, but it is not hard to imagine crossing 90° one day (it would imply time travel in some sense)!

When change accelerates like this, we tend to ask the wrong question (me too). What matters is not whether a machine can truly think, but what it can do. If it performs a task well enough, the distinction between genuine cognition and a convincing simulation becomes practically irrelevant. This echoes Alan Turing’s idea: if a machine’s responses are indistinguishable from a human’s, the “illusion” of intelligence is functionally sufficient.

This is also why fearing that AI will replace humans misses the broader pattern. It will indeed replace humans (including software engineers), just as the printing press replaced scribes and calculators replaced human computers. Yet society adapts! We no longer lament that steam engines displaced manual labor, we take their productivity as a given. The real question is not whether replacement happens, but how we reposition afterward.

Perhaps, instead of obsessing over whether machines imitate us, we should invert the perspective. For centuries, machines have imitated nature. Maybe the next step is enhancing biological capability itself, augmenting ourselves rather than merely building better stand-alone tools. Smart glasses already does this in some sense. What if we remove even the visible device?

First, chess was solved, and now coding. What’s next? We are poor at prediction, especially in exponential eras, yet we continue to guess. I predict we will have a breakthrough in brain–computer interface and neural augmentation within the next 10 years. If that happens, the next “Turk” will not be a box with a hidden human, but a seamless fusion of biological and computational intelligence.

Published: Feb 15, 2026.
Last edited: Feb 15, 2026.