Musk is versatile. The American entrepreneur has given life to many projects in various fields, each of which, if successful, can change the life of mankind. Neuralink stands out even among these projects. If we succeed in realizing it, not only the world around us will change, but also we ourselves. But will it be possible?
Warning: This article was published in Naked Science # 44, June 1, 2019. Published on the site in its original form. Has not lost its relevance.
Musk is constantly intriguing: soon we will be presented with something interesting - something that we could not have imagined before. Until that happens, we decided to figure out what Neuralink is, what to expect from it, and why all this is needed.
Neuralink: the beginning
Still shrouded in secrecy, Neuralink was founded in July 2016. Its headquarters are located in San Francisco, and the stated goals of its activities are research in the field of medicine, and more precisely, the development and production of implantable neurocomputer interfaces. To make it clearer: the company was formed to create a "technology of the future" that can connect the human brain to a computer. The main goal is to create an effective brain-computer interface.
Elon Musk is one of the founders of Neuralink and is part of the company's expert group. As Musk himself told Tim Urban, author of the blog Wait But Why and perhaps the main promoter of the entrepreneur's idea, he personally met with thousands of specialists from different fields to form a project team. The names of the participants who were among the first to be invited are known.
This is Vanessa Tholos, an engineer and flexible electrode specialist at Livermore National Laboratory. Ernest Lawrence; Timothy Gardner, a professor at Boston University, known for implanting electrodes into the brains of birds to understand how they sing; Philip Sabes is a professor at the University of California, San Francisco who studies how the brain controls movement. The recruitment of specialists continues now, the company's website informs about the search for the most talented experts from various fields. By the way, information on staff recruitment is all that the portal can please.
Musk first spoke about the idea of creating a wireless brain-computer interface in the summer of 2016 at the Vox Media conference. This interface has been referred to as neural lace. The very concept of neural lace comes from the science fiction stories of futurist writer Ian Banks. In his books, neural lace is a kind of cobweb-like device that is implanted into the human brain and provides symbiosis with machines.
According to The Wall Street Journal, Musk was going to finance the project on his own, including with borrowed funds for his shares in other companies. Another investor can be the Founders Fund of Peter Thiel, the creator of the PayPal payment system.
Penetration into the brain. How it was?
Actually, what can a brain and a computer have in common? Answer: at least electricity. Back in 1849, the Swiss scientist Emile Heinrich Dubois-Reymond found evidence for the presence of electrical potentials in living tissues. And, in particular, he proved that the brain - like nerve and muscle tissue - is capable of generating electrical signals. In 1875, electrical activity in the brains of animals, independently of each other, was discovered by the English physiologist and surgeon Richard Caton (he studied the brains of rabbits and monkeys) and the Russian physiologist Vasily Yakovlevich Danilevsky, who worked with dogs.
In 1924, the German physiologist and psychiatrist Hans Berger, using a galvanometer, for the first time recorded on paper in the form of a curve the electrical signals generated by the human brain. He suggested calling such a recording an electroencephalogram.
For the first time, an experiment using a neurointerface was performed by the American cyberneticist, neurophysiologist and psychiatrist Gray Walter in 1963. He was one of the first to implant microelectrodes in the brain. For the experiment, electrodes were first implanted in patients in various areas of the cerebral cortex. During the experiment, they had to press a button that toggles the slides of the projector. Meanwhile, Walter was recording the corresponding brain activity. Finding the area of the cortex responsible for the reproduction of this muscle pattern, he connected it directly to the projector, and disconnected the button from it.
The patients did not know that the button was disabled, they continued to press it, and the slides were switched anyway. Now the control of the projector was carried out directly by the brain - and faster than a person could press a button.
Neurologist and inventor Philip Kennedy first implanted electrodes in the brain of a paralyzed patient in 1998. A few months after the operation, his patient learned to move the cursor around the computer screen and type.
Today, brain-implanted electrodes are used to mitigate the side effects of Parkinson's disease, epilepsy, and other neurodegenerative diseases. Such devices are already used by about 150 thousand people with Parkinson's disease. Electrodes are implanted deep into the brain and generate regular electrical impulses.
It should be noted that there are three ways to "enter the brain." Invasive (from Novolatinsk invasivus, invado - "I go inside"), that is, microelectrodes are placed directly into the cerebral cortex. Semi-invasive - electrodes are placed on the hard or arachnoid meninges. In the skull, the dura mater is directly adjacent to the bones of the skull. In total, the brain has three membranes. The third, the "deepest" one, directly adjacent to the brain, is called soft. And the last, most gentle way is non-invasive: sensors for measuring the electrical potentials generated by the brain are placed on the scalp.
At the same time, it is obvious that the most difficult and risky method is invasive. It is used only in cases when there is no other way to help the patient.
In general, the brain is much more complex than a computer. Most scientists consider it to be the most complex object known to science. Even the most advanced computer has orders of magnitude fewer operating units than the human brain has neurons. Scientists note that we still know little about how brain neurons interact with each other, and methods for studying the brain are not yet perfect enough.
What Neuralink wants to surprise
So what, in these conditions, does Elon Musk want to surprise us? According to the WSJ, Neuralink will develop devices that will be implanted in the human brain and will be able to provide direct contact with computers and other equipment.
At first, Neuralink plans to release implants for the treatment of brain diseases: Parkinson's disease, epilepsy, paralysis and others. But still, the main goal is the improvement of people.
However, Neuralink has yet to prove the safety and efficacy of its technology in applied medical problems - the treatment of people. But if it succeeds, then the company will create brain implants to improve cognitive abilities. A person will be able to directly connect to a computer - without input-output devices. No keyboards, mice or joysticks.
At the same time, we are not talking about complex and expensive operations at all. As Musk told Tim Urban, the plans are to create a simple automatic system for quickly implanting electrodes into the brain. The operations will be as simple as, for example, laser vision correction.
There is no exact data yet on which devices Neuralink will produce. We tried to find a few clues in a speech by Musk at the Code Conference, held in California in June 2016.Then he said that the neurointerface should become "a digital layer above the cerebral cortex." Moreover, its components do not have to be surgically implanted: this can be done by injection into the neck, and then the interface components themselves will enter the brain with the blood stream. It was clearly a question of brain implants of the "stentrode" type.
They are being developed under the Reliable Neural-Interface Technology program funded by DARPA (Defense Advanced Research Projects Agency of the US Department of Defense). Conventional stents - the basis of stentrodes - are used to cleanse blood vessels. Stentrodes are flexible stents that act as electrodes. They are delivered to the brain along with the blood stream. Having reached its destination, the "stentrode" expands to fix its position, after which it begins to read signals from about ten thousand neurons located in the neighborhood, and transmits them through a thin wire to an external microcomputer. With proper signal processing, they can be decoded and used - for example, to control an exoskeleton. Such use will greatly facilitate the life of paralyzed people.
What is it all for?
Helping people whose condition is far from the norm is a noble goal. But the ultimate consumer of Neuralink products is all of us. Why do we need it?
As Musk said at the Vox Media conference when he announced the concept of neural lace, the technology will lead to a symbiosis of people and machines, and help humans avoid enslavement by artificial intelligence. The creation of artificial superintelligence, a machine that turns out to be smarter than a person, is a matter of time. And such a machine poses a potential threat to humanity.
In this case, the coexistence of mankind with such a superintelligence on the same planet can only be compared with the interaction of a highly developed civilization and a less developed one. Our history shows that it must end badly for the second.
Elon Musk has repeatedly warned of the danger posed by machine intelligence. But since it is impossible to stop its appearance, the entrepreneur sees the solution to the problem in giving people the opportunity to become part of artificial intelligence themselves.
As Musk said at the 2017 World Government Summit in Dubai, without a direct connection to a computer, a person can lose control over artificial intelligence and become a useless link.
Where is Neuralink now
The company does not tell the public about the details of its work. But still, in the world of science, research results are usually brought to the attention of the scientific and general public. So, on March 14 this year, an article “Sewing machine for minimally invasive neural recording” appeared on the bioRxiv preprint server, among the authors of which the knowledgeable public found the names of famous members of the Neuralink team - in particular, Philippe Sabes.
It should be noted that a "preprint" is a scientific article that has not passed the "peer review", a mandatory procedure before publication in scientific journals. Because this process can be slow, authors use the bioRxiv web service to make their manuscripts available before peer review, thereby allowing other scientists to view, discuss, and comment on their results immediately. Well, you and me too. However, be warned that articles on bioRxiv may contain errors or information that has not yet been accepted or approved by the scientific community.
So what achievement are scientists talking about? The article describes the technology for creating a brain interface that works as a thought-reading system. The experiment was carried out on rats. Scientists inserted electrodes into the animal's brain in a way similar to the principle of a sewing machine. In an experimental rodent, experts removed a part of the skull and introduced microelectrodes into the exposed area of the brain using a needle.
Moreover, it took only a few seconds to implant one electrode.This is much faster than other methods. The electrodes were connected to a small board attached to the back of the animal's head to record brain signals.
True, not all animals were able to fix the device for a long time. But one test subject was lucky. Thus, the article contains the results of two months of observations of the brain reactions of a rat, in whose brain two dozen electrodes were implanted.
So far, the general conclusion is that further research is needed. However, the developments can become an alternative to modern neurosurgical techniques and open the way to a new generation of developments in the field of robotics, electronics and artificial intelligence.
In an interview with Tim Urban in 2017, Musk promised to create neural lace “in eight to ten years,” that is, by 2027 at the most. So there is time. And, apparently, on the way to this goal, the new company of the famous entrepreneur and inventor will remind us of itself more than once.