Octopuses Learn by Watching Other Octopuses—Even Though They Barely Interact

Octopuses can watch another octopus solve a problem once and immediately replicate it—without ever meeting that octopus again, likely never interacting at all, and certainly without anyone teaching them. This shouldn't be possible.

The conventional wisdom says observational learning evolved in social animals. Primates, elephants, dolphins—creatures that stick together, raise offspring collectively, and navigate complex group hierarchies—these are the animals that benefit from watching and learning from peers. Intelligence, from this view, is the price of admission to society. Solitary animals, the thinking goes, have no evolutionary incentive to develop this skill. An octopus in the wild doesn't have a buddy system or a mentor figure. It's born, it grows up alone, and it dies in less than two years without ever raising young or maintaining a family unit. By all rights, observational learning should be completely absent from the octopus skill set.

Yet according to recent research, octopuses possess precisely this ability. Studies have documented that octopuses can learn by observing other octopuses performing tasks, even when the observer has no social bond with the performer and no expectation of future interaction. Researchers have shown that when one octopus watches another solve a puzzle or overcome a barrier, the observer later attempts similar solutions—suggesting it has internalized what it witnessed. This capability was once believed to be a vertebrate-exclusive trait, a marker of the kind of social cognition that only animals living in groups would ever need to develop.

The mechanism underlying this remains partly mysterious, but the evidence is solid. According to findings on octopus intelligence and problem-solving skills, these animals demonstrate sophisticated cognitive abilities that far exceed what their solitary lifestyle would seem to require. Their brains are organized entirely differently from vertebrate brains—their neurons are distributed throughout their eight arms rather than centralized—yet they still manage to encode, retain, and apply information gleaned from watching others. This suggests observational learning isn't actually a product of social evolution at all, but rather a general-purpose cognitive tool that any intelligent animal might develop, regardless of whether it has a reason to use it.

The explanation likely hinges on the fact that intelligence, once it reaches a certain threshold, becomes a multi-tool. An octopus's remarkable brain evolved to solve the immediate problems of its environment: hunting, hiding, navigating complex terrain, manipulating objects. A brain that sophisticated doesn't need much additional circuitry to copy what it sees another brain doing. The capacity to learn by observation might be an accidental byproduct of general intelligence rather than something specifically selected for. In other words, octopuses didn't develop this skill because they're social—they developed it because they're smart, and smart brains can do many things whether or not those things were strictly necessary for survival.

This raises an unsettling question about intelligence itself: How much of what we think of as "social cognition" is actually just what brains do when they're sufficiently complex? If a solitary octopus can learn from watching strangers it will never meet again, maybe observational learning isn't really about society at all. Maybe we've been confusing the context in which intelligence developed with the actual machinery of intelligence. The octopus is a living reminder that evolution sometimes builds more powerful tools than a given animal strictly needs—and that things we thought required community might actually just require thinking.