The Complexity of the Human Brain Versus Artificial Intelligence Ambitions

The complexity of the human brain is paving the way for discussions on the aspirations of artificial intelligence research. Industry expert Aakash Gupta emphasised the immense challenge of mapping even a small segment of neural tissue.

A Decade-Long Scientific Endeavour

In a recent update on X (previously known as Twitter), Gupta referenced a decade of diligent work conducted by researchers at Harvard University in partnership with Google. He suggested that the research findings should significantly humble every AI lab globally.

A Cubic Millimetre: A Monumental Challenge

The focus of the project was on reconstructing a mere cubic millimetre of human brain tissue, which is approximately one-millionth the volume of the entire brain. This effort rivalled substantial industrial research programmes.

Scientists dedicated ten years to mapping this tiny sample. The imaging process alone operated continuously for 326 days, slicing the tissue into 5,000 ultra-thin sections, each just 30 nanometers thick. These were scanned using an electron microscope valued at $6 million.

The Data Challenge

Even after the imaging phase, the task had only just begun. The dataset became so extensive that automated machine-learning systems were essential for stitching the images into a cohesive three-dimensional reconstruction, which Gupta remarked was beyond the capability of any human team to process manually.

Massive Data from a Minuscule Sample

From that tiny fragment of tissue, the researchers identified:

• 57,000 cells
• 150 million synapses
• 230 millimetres of blood vessels

This amounted to 1.4 petabytes of raw data—equivalent to about 1.4 million gigabytes—compressed from a sample smaller than a grain of rice.

Gupta extrapolated that mapping the entire human brain at the same level of detail would generate an estimated 1.4 zettabytes of data. This volume is comparable to the total data produced worldwide in a single year. Storing this information could incur costs in the tens of billions and would necessitate data centre spaces spanning hundreds of acres.

Discoveries That Challenge Existing Knowledge

Beyond the impressive engineering achievements, the researchers uncovered biological structures that remain poorly understood.

Neuroscientist Jeff Lichtman, who led the Harvard project, indicated that the findings highlighted “a significant gap between what is currently known and what remains to be uncovered.”

Among the unexpected discoveries were:

• A single neuron creating over 5,000 connection points
• Axons forming tightly coiled whorls with no clear rationale
• Cell clusters arranged in mirrored formations

Such revelations imply that foundational assumptions regarding neural wiring might be incomplete.

Given the substantial technical demands, scientists are not planning to map the entire human brain immediately. Instead, their next focus is on a mouse hippocampus, which encompasses around 10 cubic millimetres of tissue over the following five years.

A Substantial Step Forward

This modest undertaking represents an increase in scale of roughly 1,000 times compared to the already analysed tissue, positioning it as a critical proof-of-concept for connectomics, the discipline dedicated to mapping neural connections.

A Reality Check for Artificial Intelligence Aspirations

Gupta underscored a paradox prevalent in current AI development: modern neural networks draw inspiration from the brain, yet researchers have not yet fully decoded the wiring of even a small segment of the biological system they aim to replicate.

While advanced AI models rely on vast computing clusters, the human brain operates on approximately 20 watts of power—less than that of a standard household light bulb. This discrepancy raises important questions about efficiency, design, and the true requirements of intelligence.

Gupta argued that the essential message is not to undermine AI’s progress but to acknowledge the remarkable complexity of the biological model, which science is just beginning to explore.