"The most incomprehensible thing about the world is that the world is comprehensible."
A public disagreement between AI researchers Yann LeCun and Demis Hassabis has reopened a long-running debate on whether human intelligence can be described as "general".
In a recent podcast appearance, LeCun said the idea of general intelligence, when used to mean human-level intelligence, is flawed.
LeCun argues that "there is no such thing as general intelligence", saying the term is largely used to describe human-level intelligence, which he believes is a mistake. Human intelligence, he said, is "super specialised", shaped by evolution to handle the physical world and social interaction efficiently.
While humans navigate real-world environments and deal with other people well, LeCun pointed out that they perform poorly at many structured tasks, like chess, and are outperformed by other animals in several domains. This, he said, shows that humans are not broadly general but highly specialised.
"We think of ourselves as being general, but it's simply an illusion because all of the problems that we can apprehend are the ones that we can think of," LeCun said.
Hassabis responded that LeCun was conflating general intelligence with universal intelligence. "Brains are the most exquisite and complex phenomena we know of in the universe (so far), and they are in fact extremely general," he wrote in his post on X.
He argued that while no system can escape the no free lunch theorem, a general system can still learn any computable function in principle. "In the Turing machine sense, the architecture of such a general system is capable of learning anything computable given enough time and memory," he said, adding that human brains and AI foundation models are "approximate Turing machines".
Hassabis also rejected the idea that human performance in narrow domains undermines generality. Referring to chess, he said it was notable that humans invented the game at all and reached elite levels of play.
LeCun later said the dispute was largely about terminology. "I object to the use of 'general' to designate 'human level' because humans are extremely specialised," he wrote in his response.
He argued that intelligence should be judged not just by theoretical capability but by efficiency under limited resources. "For the vast majority of computational problems, [the human brain is] horribly inefficient," he said, citing time and memory constraints in tasks such as chess.
To support his argument, LeCun used an analogy from deep learning, noting that while a simple neural network can approximate any function in theory, it becomes impractical for most real-world problems.
He also pointed to biological limits, arguing that the number of functions the human brain can represent is vanishingly small compared to the space of all possible functions. "Not only are we not general, we are [also] ridiculously specialised," he said.
LeCun concluded by noting that humans mistake this specialisation for generality because most possible functions are incomprehensible. Quoting Albert Einstein, he wrote, "The most incomprehensible thing about the world is that the world is comprehensible."