StudyAI showed that AI ideas appeared long before computers — from Heron to Geniac

StudyAI has compiled a vivid retrospective showing that the idea of artificial intelligence emerged long before electronic computers. Long before neural networks, people were already trying to create mechanisms that looked like "thinking" machines and performed actions according to given logic.

A Dream Older Than the Computer

The main point of the article is simple: if we consider AI not just as a program on a chip, but as a system that follows an algorithm and imitates intelligent behavior, then its history does not begin in the XX century. The author draws a line from ancient automatons to mechanical toys and educational machines that operated without processors, screens, and cloud infrastructure. These devices did not think in the modern sense, but they showed that people had long been trying to turn rules, sequences, and feedback into machine behavior.

One of the earliest examples is an Egyptian proto-automaton from the Middle Kingdom, depicting a figure of Hathor with a simple internal mechanism. Later, Hero of Alexandria took the idea to a completely different level: his constructions could perform scenes, move characters, and even dispense holy water using a principle similar to a vending machine. In the Middle Ages, Arab engineers picked up the torch: Al-Jazari described complex automatons, including a handwashing device that offered soap and towels on its own. All of this looked like mechanical theater, but already with clearly defined logic of actions.

Mechanics as Algorithm

The most famous example from the collection is the Antikythera mechanism. It is usually called the first analog computer, but in the text it is presented more broadly: as an ancient machine that turned astronomical observations into useful recommendations. The device tracked celestial cycles, helped predict eclipses, and calculated dates. Essentially, we have before us a compact decision-making system on gears: the user inputs initial data, and the mechanism produces a meaningful result. For an era without electronics, this looked almost incredible.

The collection is important not because of its exoticism, but because it shows several different models of machine behavior. Some devices reproduced a pre-set scenario, others calculated cycles and helped with forecasting, still others reacted to obstacles almost like autonomous robots. These were not just toys for the public, but practical experiments with logic, prediction, and control. This is why the historical examples read today not as curiosities, but as early versions of engineering ideas familiar from modern robotics and AI.

• Hero of Alexandria's automatons performed scenes and worked according to a pre-programmed scenario
• The Antikythera mechanism calculated astronomical cycles and provided dates for important events
• Leonardo da Vinci's mechanical knight could move, sit, stand, and raise its visor
• Geniac taught Boolean logic, played tic-tac-toe, and even assembled simple musical sequences
• The Soviet "Umka" navigated obstacles and didn't fall off the table thanks to pure mechanics

A special place is occupied by Leonardo da Vinci's mechanical knight. Based on surviving sketches, it was built on cables, pulleys, and transmissions that mimicked human biomechanics. The machine could move limbs, change body position, and perform a set of spectacular actions at a carnival at the court of Ludovico Sforza. For the modern reader, this is more of an ancestor of humanoid robotics than a Renaissance curiosity. But such projects break the familiar picture in which the history of AI begins only with the mathematical models of the XX century.

Toys Before Neural Networks

As we get closer to modern times, the collection moves from spectacular automatons to devices that already resemble computational technology. Particularly notable is Geniac—a mechanical educational computer from the mid-XX century with light bulbs, disks, switches, and jumpers. It had no memory, it worked on combinational logic, but it allowed you to assemble circuits, play, solve simple tasks, and literally reprogram the machine's behavior with your hands in a matter of minutes. This is an important bridge between the age of automatons and the time of mass cybernetics.

No less interesting is the Soviet device "Umka"—a toy car that could navigate obstacles and not fall off a table without cameras, lidars, or depth sensors. Its behavior was provided by a probe, a movable support, and a wheel installed at an angle. When the car hit an obstacle or approached the edge, mechanical feedback changed its trajectory. By modern standards, this is a very simple system, but that is precisely the strength of the example: useful autonomous behavior can be obtained even without computational power if an engineer properly set up the rules for reacting to the environment.

What It Means

StudyAI's collection shows well: the history of AI is not only about neural networks, GPUs, and large language models. It is a long evolution of ideas in which people have spent centuries trying to teach machines to perceive signals, follow algorithms, and respond to the world without direct human intervention.