Today, a Habr post flashed past in my feed: “How to Make Your Plant Work for You: a DIY Analog Bioprocessor.” The author sticks four electrodes into the fleshy leaf of a jade plant — that very “money tree” — hooks them up to an Arduino, and gets stable nonlinear readings: with the LOW, LOW combination, the device honestly outputs 15–19; with HIGH, HIGH, 572–576. Those two states can already be interpreted as logical AND and OR. You get a bioprocessor: bio because ions are doing the computing, analog because the transformation is smooth, processor because the output depends on the input according to a function.
The author himself honestly warns that he did it for fun and to learn new skills. But the question hangs in the air: what, exactly, is happening here that even looks scientific? Can we seriously talk about a plant as a computing device?
Spoiler: behind this “windowsill gag” lie 150 years of science, including the first bio-signal ever instrumentally recorded in physiology, Nobel-level ambition, a scientific scandal with 36 signatures, and spinach that can detect explosives.
Electricity in plants was measured before electricity in the human heart. In 1873, the English physiologist John Burdon-Sanderson placed electrodes on a Venus flytrap during a hunt and recorded a characteristic voltage spike — the action potential of living tissue. This is considered one of the first electrophysiological recordings in history, period: Waller would not produce the first human electrocardiogram until 1887. In other words, before humanity heard its own heart, it heard a flytrap deciding whether to eat a fly.
Seven years later, Charles Darwin entered the game. In The Power of Movement in Plants (1880), experimenting with the roots of seedlings, he concluded that a signal of irritation travels from the root tip to the growth zone — and put it bluntly: the root tip “acts like the brain of one of the lower animals,” receiving impressions from the sense-organs and directing the movements of neighboring parts. Darwin planted both ideas that would torment science for a century and a half: the plant as sensor and the plant as “brain.”
The brightest figure in this story is Jagadish Chandra Bose, a physicist from Calcutta. At the turn of the 20th century, he built the crescograph — an instrument that mechanically magnified the movements of plant tissues millions of times and recorded them on a smoked glass plate. With it, Bose registered tissue pulsations, which he called the “heartbeat of plants,” as well as reactions to poisons, electricity, and touch. In 1902, he published Response in the Living and Non-Living, in which he showed that plant tissue and a piece of metal respond to stimulation with strikingly similar electrical curves, drawing the radical conclusion that no sharp boundary exists between the living and the nonliving.
Western science received Bose coolly: too many bold generalizations, too little rigor, too much “the plant feels.” But he left behind a method for good — electrodes in tissue, a galvanometer at the output, and a plant speaking the language of volts.
The irony is that Bose was also a pioneer of radio waves. He experimented with millimeter waves, built his own coherer, and publicly demonstrated wireless signal transmission in Calcutta in the mid-1890s — alongside Marconi and Popov, and in some cases before them. He lost the patent race — more precisely, he simply never entered it — but remained in history as one of the first researchers of the millimeter-wave band, the very one 5G uses today. The physicist who figured out how to make plants speak in electricity also anticipated modern communications.
Now for the mechanics. On the inner surface of a Venus flytrap’s trap are sensitive trigger hairs. Touch activates mechanosensitive ion channels; they generate an action potential — and the trap snaps shut in a hundred milliseconds, one of the fastest movements in the plant world.
There is a catch: one action potential is not enough. It needs two, arriving about 20–30 seconds apart. This filters out false triggers: a raindrop can cause one impulse, but not two in succession. In essence, the plant implements a logical AND gate with a timer.
And there’s more. In 2016, a group of researchers showed in Current Biology that the flytrap can count. Two action potentials — snap shut. The third — begin sealing and synthesizing digestive enzymes. The fifth — activate pumps that draw sodium ions out of the prey. The plant “decides” how deeply to digest by counting its own electrical impulses. In essence, it is a finite-state machine with five states, its counter powered by bioelectricity.
The sensitive plant adds memory to the picture. When its leaves fold, the signal spreads through the tissues at roughly a centimeter per second — 4 orders of magnitude slower than a nerve impulse, but for a plant, that is lightning. And in 2014, Monica Gagliano’s team published a paper in which Mimosa “grew accustomed” to harmless falling water droplets and stopped folding its leaves — and, judging by the authors’ data, remembered the lesson for at least a month: after 28 days, the habituation could still be reproduced. The methods were later criticized, the reproducibility challenged, but the idea of “memory without a brain” has hung firmly in the air ever since.
In the mid-2000s, a group of physiologists — Stefano Mancuso, František Baluška, and others — announced the birth of “plant neurobiology”: a discipline studying plant signaling systems which, in their view, are functionally analogous to nervous systems. Plants, they claimed, are capable of learning, memory, and problem-solving, while their root networks exhibit something akin to swarm intelligence.
The response was swift. In 2007, Trends in Plant Science published a manifesto signed by 36 scientists, with the expressive title “Plant neurobiology: no brain, no gain.” Plants have no neurons, no synapses, nothing resembling a central nervous system, the authors wrote; all this terminology is anthropomorphism that misleads and harms science’s reputation. Baluška and his co-authors responded with a mirror-image article: “the gain is more than the name.”
The dispute settled nothing, but that is precisely what makes it interesting. “Plant intelligence” is not a fact but a framework of interpretation. On one side stands strict materialism: there are ion channels, action potentials, hormonal signals, and no “consciousness.” On the other, a cautious analogy: distributed sensing, local decisions, memory without a central organ. Both sides are right on their own terms, and that is exactly why the subject has endured for 2 decades.
In 2016, Michael Strano’s lab at MIT published a paper in Nature Materials on “plant nanobionics.” Carbon nanotubes were embedded in spinach leaves: when the plant’s roots absorb nitroaromatic compounds from groundwater — typical components of explosives — the nanotubes begin fluorescing in the near-infrared range, and the signal can be read by an ordinary smartphone camera. A sensor plant that grows itself, feeds itself, and silently signals that the ground is mined. And there was more: in the same experiments, an infrared “communication channel” was built into the leaves. Essentially, the plant was taught to receive and reflect a signal visible from the outside with a regular camera. In effect, it was the first step toward a plant transmitter.
Then, in 2021, Nature Electronics published a paper on a “plant actuator”: a Venus flytrap was connected to conformable electrodes, and researchers learned to make its trap close on command — so delicately that a “robot hand” with a flytrap at the end could grasp fragile objects without damaging them. New Scientist dubbed it a “flytrap cyborg,” controlled from a smartphone. Here, the plant is not a computer but an actuator: a ready-made pneumatic-hydraulic drive honed over millions of years, needing only to be plugged into an outlet.
Interestingly, this line of research has a reverse side too, which we have already covered: in the 2000s, “Physarum computers” were all the rage — slime mold as a living network optimizer, “designing” the Tokyo subway system. Then it turned out that reproducibility was shaky, and the slime mold’s “genius” was a property of the task, not the organism. The story of plants as computers is the same plot in miniature: as sensors and actuators, plants work; as “processors,” they are only a metaphor.
Let us return to the Habr bioprocessor. Between the two electrodes stuck into the leaf is living tissue: an ionic conductor with nonlinear resistance that depends on ionic composition, moisture, membrane polarization, and applied voltage. With the LOW, LOW combination, the current through the leaf is minimal — the tissue behaves like a high-resistance resistor, hence the modest 15–19. With HIGH, HIGH, the voltage polarizes the membranes, conductivity rises — and the device honestly shows 572–576. Two stable states — there’s your “logic.”
The crucial point: the “processor” here is not the plant but the scheme of interpretation. The plant is a nonlinear element and a sensor, while the decision that 15–19 means “zero” and 572–576 means “one” is made by a human, as in any analog computer — from hydraulic machines to water-based logic circuits. This is neither deception nor a stretch: that is how all analog computing works. Usually, the nonlinear element is made of silicon; here, it is made of a leaf that also photosynthesizes.
What is especially pleasing is that the post’s author repeats exactly what Bose did — makes a plant speak in volts and amperes. The only difference is that Bose had a crescograph with mechanical magnification in the millions, a laboratory in Calcutta, and a knighthood, while here there is a $10 Arduino and the urge to see “what happens.” Science does not always move from large to small; sometimes it simply returns to the kitchen.
The plant as computer is a beautiful and almost always false metaphor. The plant as measuring device and actuator is real engineering, 150 years old. Between those two poles lies the field’s entire history: from the first bio-potential ever recorded in history — yes, from a flytrap — through Bose’s crescograph and the “plant neurobiology” scandal to spinach tapping nanotubes when explosives are hidden underground.
And the tastiest thing about this story is its final irony. We search for extraterrestrial intelligence, build neural networks, argue over consciousness in octopuses — while the most alien, most unlike-our-own way of processing information has been standing on the windowsill for a billion years. It does not think as we do. It simply counts to five — and only then eats the fly. Maybe that is why it is so hard to come to terms with it: it has its own logic, and it is not about us but about survival.