Neuralink: Brain-computer interfaces are entering the takeoff era
Max Hodak, the co-founder of Neuralink, pointed out in an interview that the essence of brain-computer interface is a technology that uses the neural interface of the brain for two-way communication.
Recently, Max Hodak, the co-founder of Neuralink and the current founder and CEO of Science, mentioned in an interview the breakthrough progress of Science in restoring the vision of the blind. In the interview, he also boldly predicted that with the integration and development of artificial intelligence (AI) and brain-computer interface technology (BCI), the first group of humans who can live to be 1000 years old may have already been born.
Breakthrough Progress
In the interview, Hodak shared the breakthrough progress of Science in the field of restoring vision, saying that its core product, Prima, has helped more than 40 blind patients regain their sight. The product also recently conducted a large-scale clinical trial, and the research results have been published in the New England Journal of Medicine.
The technical principle of Prima is to implant a 2mm×2mm micro silicon chip under the retina at the back of the eyeball of blind patients, which is equivalent to a solar cell array. Patients need to wear special glasses. After the camera in the glasses captures the external image, the image is projected into the eye through a laser projector. After the laser irradiates the implant, the solar cells on it absorb the light and stimulate the cells directly above, thereby bypassing the necrotic rods and cones and transmitting the visual signal back to the retina.
The trial was conducted in 17 centers in Europe, and the results were remarkable. Some patients were able to read every letter on the eye chart after the operation. This technology is mainly targeted at patients who are blind due to the loss of photoreceptors, such as age-related macular degeneration and retinitis pigmentosa.
Hodak pointed out, "It helps blind patients create coherent and formed images in their minds for the first time in the world."
In contrast, previous technical solutions, such as the product of Second Sight a decade ago, could only allow patients to perceive sporadic flashes (phosphenes) without being able to combine them into meaningful visual images.
Redefinition of Brain-Computer Interface
Hodak redefined the brain-computer interface (BCI) from the perspective of computer science.
He believes that the brain is essentially an information processing computer enclosed in the skull. All information flowing into or out of the brain passes through a few "cables": the optic nerve (cranial nerve II), the auditory nerve (cranial nerve VIII), and 31 pairs of spinal nerves. And the electrical impulses on these nerves are the API (Application Programming Interface) of the brain.
Hodak pointed out that humans do not directly experience the external world but experience the "world model" constructed by the brain. This means that the walls and lights we see are all fictional generative models of the brain.
When talking about brain-computer interface technology, Hodak believes that its essence is to understand and utilize these APIs for two-way communication.
Moreover, there are many types of brain-computer interfaces, not a single universal product. Hodak pointed out, "I don't think there will be a universal interface that everyone will implant. Different modes are often suitable for different things. For example, using electrical stimulation will help restore functions (such as Prima); using ultrasound may achieve a 'digital sleeping pill' or a 'digital stimulant', etc."
He also put forward the view that "neuroscience is moving towards a grand unification with artificial intelligence." He believes that the "latent space" inside the artificial intelligence model, that is, the internal representation of the data learned by the model, is very similar to the way the inferior temporal cortex of the brain processes information.
Hodak said bluntly, "Neuroscience and artificial intelligence are deeply integrated. Many neuroscientists have switched to artificial intelligence research because it is more convenient to explore these representations in artificial intelligence models than to directly study the human brain." Of course, conversely, artificial intelligence research also provides a new theoretical basis for interpreting and intervening in brain activities through the brain-computer interface (BCI).
Biohybrid Neural Interface
For the next 5 to 10 years, Hodak outlined a blueprint beyond the existing electrode stimulation - the biohybrid neural interface.
Science is trying to cultivate stem cell-induced living neurons in the brain implant, allowing them to "grow together" with the original neurons in the brain. And use "low immunogenicity technology" to hide it from the immune system to avoid rejection, realizing a universal graft without customization.
Hodak compared this concept to the way of neural interface connection in the movie "Avatar", aiming to establish a new ultra-high bandwidth "cranial nerve", such as a nerve bundle directly connected to the Internet, to achieve a deep coupling between human consciousness and machines.
In the interview, Hodak also mentioned the term "conscious machine". He said, "There is a saying that the brain-computer interface is a side story of artificial intelligence, and the goal is to integrate humans and machines. I think it makes sense, but more directly, I think the brain-computer interface is a side story of 'longevity' and 'medicine'. If the ultimate goal of artificial intelligence is a super-intelligent machine, then the ultimate goal of BCI is actually a 'conscious machine'."
"There may be no measurement method that can tell us whether a thing is conscious. The only thing you can be sure is conscious is yourself. If so, to study consciousness, we need to use brain-computer interface technology to experience it personally. Once you understand the underlying physics of the brain supporting consciousness, you will ultimately get a super-intelligent conscious machine, and we can become a part of it through an ultra-high bandwidth connection."
Can Humans Live to 1000 Years Old?
At the end of the interview, Hodak made a remarkable long-term prediction. He pointed out, "The first group of people who may live to be 1000 years old have probably already been born."
He predicted that the adoption of brain-computer interfaces will follow a clear path: first, it will be applied to help severely disabled patients restore basic functions; then, it will be extended to address the functional decline caused by normal aging; finally, when the implant can provide some enhanced capabilities (such as direct access to information and super-sensory experiences), the ratio of risk to reward will fundamentally change, and public acceptance will significantly increase.
Hodak concluded that although biotechnology has developed slowly in the past, it has now entered an era of non-linear acceleration. He optimistically believes that by 2035, although not all diseases can be cured, humans will have new technological means to reshape the interface between humans and machines, thus profoundly changing the meaning and boundaries of life itself.
This article is from the WeChat official account "Science and Technology Innovation Board Daily", author: Zhou Ziyi. Republished by 36Kr with permission.