Neuralink is already attempting to transmit vision directly to the brain; the project is called Blindsight, and its goal is to use a brain implant to try to restore some level of visual perception to blind individuals. However, there is a specific detail that makes this technology particularly interesting. The concept involves creating vision without relying directly on the eyes. When we see something, the eyes are merely the starting point of the process: the retina converts light into electrical signals, which travel along the optic nerve to the regions of the brain responsible for visual processing; it is only there that these signals begin to transform into what we perceive as an image.

Blindsight aims to explore the other end of this pathway: a camera would capture the surroundings, an external device would convert that input into data, and the Neuralink implant would directly stimulate the visual cortex. In other words, the system would attempt to deliver an image straight to the brain. In 2024, the FDA granted Blindsight "breakthrough device" designation, which could facilitate the company's interaction with the agency during the development process.

This does not mean the technology has been approved for clinical use, and there is another important detail to consider: if it works, we should not expect someone to immediately regain vision comparable to our own. One of the foundational elements studied in this field is phosphenes—perceptions of small dots or flashes of light produced when specific regions of the visual system are stimulated.

The great challenge lies in controlling many of those stimuli and transforming them into patterns the brain can interpret as useful information. While this might initially allow for the distinction of shapes, contours, or objects—or help someone navigate their surroundings—the underlying principle is far more profound. With Terry, we saw a machine attempting to learn how to interpret signals generated by the brain; with technology like Blindsight, the dynamic is reversed: the machine must figure out how to produce signals that the brain can interpret.

If we eventually master both directions—brain-to-machine and machine-to-brain—we will have created something vastly different from the interfaces we use today. We will no longer be talking merely about controlling computers with our thoughts; we will be discussing technology capable of directly exchanging information with the human brain. This raises a final question: currently, these technologies are primarily developed to restore lost functions—such as movement, speech, or vision—but what happens once we perfect this process? Will we continue to use brain interfaces solely to restore human capabilities, or will we begin using them to create capabilities we never possessed before?