An octopus has roughly 500 million neurons. For context, a dog has about 530 million, a cat about 760 million. But here’s where the comparison stops being useful: two-thirds of an octopus’s neurons don’t reside in its brain. They’re distributed across its eight arms, each of which contains a neural network complex enough to taste, touch, decide, and act semi-autonomously—without waiting for instructions from the central brain. An octopus arm that has been surgically severed will continue to respond to stimuli, reach for food, and retract from threats for up to an hour. The arm doesn’t know it’s been separated from the animal. It has enough local intelligence to carry on.
This is not how intelligence is supposed to work. Every vertebrate on earth—every mammal, bird, reptile, fish—runs on the same basic architecture: a centralized brain that receives sensory input, processes it, and sends commands to the body. The octopus evolved an entirely different solution. Its intelligence is not housed in its brain and expressed through its body. Its intelligence is a property of the entire organism, with cognitive processing distributed across multiple semi-independent neural centers that coordinate without a strict hierarchy. The last common ancestor between octopuses and humans lived roughly 500 to 600 million years ago—a flatworm-like organism with no eyes, no limbs, and a nervous system barely worthy of the name. Everything the octopus brain can do, it evolved independently from everything the human brain can do. Convergent evolution of complex cognition, separated by half a billion years.
What they can actually do
The behavioral evidence is extensive and, for a mollusk, frankly embarrassing to vertebrates. Octopuses open screw-top jars from the inside. They navigate complex mazes and remember the solution. They carry coconut shell halves across the ocean floor, reassemble them into a shelter when threatened—tool use, planning, and multi-step problem-solving combined in a single behavior. They’ve been observed shooting jets of water at laboratory equipment they apparently find annoying, which researchers interpret as play behavior—activity with no obvious survival function, performed seemingly for the experience of doing it.
They recognize individual human faces and behave differently toward different people. Researchers at the Seattle Aquarium documented an octopus that consistently squirted water at one specific staff member who had done nothing to provoke it, while being docile with everyone else. They learn by observation—watching another octopus solve a problem and then replicating the solution without trial-and-error. A 2023 study in Current Biology demonstrated that some species display individual personality differences in problem-solving: neophilic octopuses (those attracted to novel objects) approached puzzle boxes faster but didn’t necessarily solve them faster than more cautious individuals, suggesting that octopus cognition involves multiple independent cognitive traits that don’t all scale together.
An August 2025 paper in Trends in Ecology & Evolution introduced a framework for understanding tactical deception in cephalopods—the capacity to mislead other organisms through deliberate behavioral manipulation, a cognitive ability previously attributed almost exclusively to primates and corvids. A January 2026 paper in Biological Reviews provided an updated assessment of sentience in cephalopod mollusks, building on the 2012 Cambridge Declaration on Consciousness that specifically included cephalopods among animals capable of conscious experience—the first time invertebrates received such recognition from a formal scientific consensus.
The distributed brain
A 2024 study published in Current Biology produced a three-dimensional molecular atlas of the octopus arm nerve cord, revealing spatial and neurochemical complexity that researchers described as far richer than previously understood. The arm nerve cord isn’t a simple relay cable. It’s a processing center with its own regional specializations, neurotransmitter systems, and computational architecture—a brain in miniature, running its own operations while communicating with the central brain through a bandwidth that appears to be relatively narrow compared to the total processing happening locally.
This distributed architecture means the octopus doesn’t perceive its surroundings, analyze that information centrally, and then issue commands to change color or move an arm. Millions of skin-based chromatophore cells can change color and texture in response to local visual input—even though the octopus’s skin technically can’t “see” in the way eyes do, the chromatophores contain light-sensitive proteins that enable the skin to respond directly to its visual environment. The camouflage isn’t centrally directed. It emerges from the coordinated activity of distributed local processing units, each responding to its immediate surroundings.
The Office of Naval Research funded a $7.5 million Multi-University Research Initiative to build a “Cyberoctopus”—a computational model that simulates the distributed intelligence within the octopus, with the goal of understanding how decentralized inference and decision-making can be leveraged for engineering applications. The research has direct implications for soft robotics, where the octopus’s ability to control a boneless, infinitely flexible body without centralized motor planning is a design paradigm that conventional robotics hasn’t been able to replicate. Related research papers on octopus-inspired technology grew from 760 in 2021 to 1,170 in 2024—a 54 percent increase in three years.
The molecular convergence
Perhaps the most striking finding in recent octopus neuroscience is the discovery that octopus brains and human brains share the same “jumping genes”—transposable elements called LINEs (Long Interspersed Nuclear Elements) that are active in the parts of the brain responsible for cognitive abilities. In humans, LINE transposons are particularly active in the hippocampus, the brain region most associated with learning and memory. In octopuses, the same family of transposons is active in the vertical lobe, the brain region most associated with learning and memory. Two organisms separated by 500 million years of evolution, using the same molecular mechanism in the same functional brain regions for the same cognitive processes.
Researchers at SISSA in Trieste and the Stazione Zoologica Anton Dohrn in Naples described this as “a fascinating example of convergent evolution”—a case where two genetically distant species independently developed the same molecular process in response to similar cognitive demands. The implication is that intelligence isn’t just a lucky accident that happened once in vertebrate evolution. It’s a solution that evolution has found multiple times, through multiple architectures, using some of the same molecular tools.
Why it matters beyond marine biology
The octopus is doing two things simultaneously for science. First, it’s demolishing the assumption that sophisticated cognition requires centralized processing. For over a century, neuroscience operated on the implicit model that intelligence means a big brain running the show while the body follows orders. The octopus demonstrates that distributed intelligence—where local nodes make autonomous decisions, coordinate with neighbors, and produce coherent global behavior without top-down control—can generate problem-solving, tool use, social recognition, and potentially consciousness. This has direct implications for AI architecture, where researchers studying octopus neural systems are designing more flexible robotic networks that don’t rely on a single central processor.
Second, the octopus is the strongest evidence we have that if complex intelligence exists elsewhere in the universe, it probably doesn’t look anything like us. The octopus evolved intelligence on the same planet as humans, in the same ocean, under the same physics, and it arrived at a solution so alien that we’re still struggling to understand how it works. If intelligence can diverge this dramatically within the shared evolutionary history of a single planet, the range of possible cognitive architectures across different planets, different chemistries, and different selection pressures is essentially unbounded. As one University of Washington neuroscientist put it: understanding how the octopus perceives its world “is as close as we can come to preparing to meet intelligent life beyond our planet.”
The octopus lives fast—most species survive only one to two years—and dies after reproducing, often dramatically (the female stops eating to guard her eggs and starves to death; the male enters senescence and essentially falls apart). This is intelligence that evolved without the benefit of long lifespans, cultural transmission, or social learning across generations. Every octopus that opens a jar, solves a maze, or recognizes a human face figured it out on its own, within a life measured in months. Whatever the octopus is, it’s not what we expected intelligence to look like. And that might be the most important thing it teaches us.
We cover octopus cognition alongside mirror neurons, whale communication, corvid intelligence, and the full landscape of animal neuroscience across our Neurozoology course—including why the most important brain on earth for understanding intelligence might be the one with two-thirds of its neurons in its arms.