With three piglets, Musk, the Silicon Valley “Iron Man”, once again pushed the “brain-computer interface” technology to the public at Neuralink’s launch. Musk has a long history of interest in brain-computer interface technology, and before the launch, he expressed his thoughts on the technology: he hoped that brain-computer interface technology would solve problems such as sleep or vision impairment, amnesia, depression, epilepsy, stroke, etc.
In addition to Musk’s cool launch, “telepathy”, “eternal life” and other concepts, the brain interface actually refers to the human brain and machines (including intelligent hardware), the Internet between the establishment of interfaces and bridges, so that the brain’s electrical signals can directly interact with computer electrical signals, so that the brain directly connected to the virtual world, communication.
In addition to Musk’s Neuralink, there is also a company in the country that is making hardware for brain-computer interfaces. Unlike Neuralink’s direct-implanted chip, the company follows the direction of the non-invasive brain interface. Founded in 2015, BrainCo is the first Chinese team to incubate at Harvard University’s Innovation Lab. The founder is a Ph.D. from Harvard University’s Brain Science Center. The team was hatched by Harvard University and received investments from China Electronics, China Merchants Capital, Everbright Holdings, Tencent Co-Founders, Walden International, Dingyu Investment and other institutions.
BrainCo founder Han Yu.
Two days before Musk’s launch, BrainCo officially released brain OS’ intelligent operating system for the human brain. After the conference, BrainCo founder Han Yu gave an exclusive interview to The News, talking about his understanding of the brain-computer interface, the current difficulties of commercialization, and the ethics of the brain-computer interface.
Major breakthrough: Brain science or brain surgery.
At Neuralink’s launch, Musk introduced two devices. One is a coin-sized chip with 1,024 channels that can be placed on the inside of the skull, read information about brain nerve activity, transmit brain wave data wirelessly in real time, and the other is a surgical robot that can avoid blood vessels and open a small hole in the skull to quickly and accurately implant the chip in a predetermined location.
Monkeys implanted with chips in their brains in 2008 controlled the robotic arm with their ideas.
“Neuralink’s chip detects pig brain signals, a technique that has been in use for the past 20 years. But they upgraded the equipment, and from that point of view, I think they’re going in the right direction. “But when I chatted with some Harvard professors about the conference, I thought they weren’t a major breakthrough in brain science, but a major breakthrough in brain surgery. The best description of the conference should be a major breakthrough in brain surgery and related equipment. “
In Han’s view, Neuralink’s two devices, from a signal point of view and behavioral relevance, made a good show. Piglets come across something and there’s a change in brain signals, which has been done in the past, but Neuralink has upgraded the device.
But the breakthrough point in the brain-machine interface over the next 5-10 years should be through the piglet’s brain signal, and we can see whether it wants to eat carrots or tomatoes next.
Although the announcement was not a major scientific breakthrough, Han expressed admiration for Musk in the interview. “I think he’s doing a lot of things, not pursuing short-term business returns, but he wants to achieve some of the scenarios that humans should experience. He believed that in the future, we should have a symbiotic mechanism between humans and AI, which was the first time he had created Neuralink. Han said.
Brain interface: intrusive or non-invasive.
Unlike Neuralink’s intrusive brain-machine interface route, BrainCo, founded by Han, follows a non-invasive route. These two routes are also the mainstream of brain-computer interface technology.
As the name suggests, the intrusive brain-computer interface is the implantation of chips and devices into the human brain to obtain brain signals. Instead of intrusive, it means getting brain signals through an external device, not surgery.
“First of all, from an academic point of view, the two routes are currently relatively even, in the United States, if we look at the university laboratory, probably about 60% of the intrusive, rather than intrusive to about 40%. I represent a non-intrusive genre, with school labs like MIT and Harvard taking this route, while the University of Washington and Stanford University take this intrusive route. “At some academic exchange meetings, people from both genres spit on each other, but don’t really think each other’s methods are really inferior to their own. “
In his view, intrusive and non-invasive two-week brain-computer interfaces have advantages and disadvantages. First of all, from the quality of the brain signal, invasive is certainly much better than non-invasive. Since the motor is inserted directly into the brain, the source signal can be collected. But from past data acquisition, intrusive brain interfaces have a limited amount of brain signal. So far, there may be fewer than 100 samples of people, most of whom are sick, according to Han. In contrast, non-invasive brain interface routes are easier and safer to obtain signals.
But whether invasive or non-invasive, the biggest problem with brain-computer interfaces today is still how to properly interpret the signals given by the brain.
“The core point of the brain-computer interface is decoding, which constantly parses these signs of the human body. For the past 5-10 years, the brain-computer interface has been decoding. But for the interpretation and meaning of direct signals in the brain, in fact, many people do not understand, do not know. All we can do now is parse infinitely and guess what these signals mean. Han said.
Because from current brain science research, the human brain is still a black box, with nearly 100 billion neurons, and no one knows what’s going on inside it or how it’s going to change.
Re-create people or make supermans.
Brain-computer interfaces have attracted much attention because of the technology’s hope of transforming people. There are two kinds of transformations: one is to help people with disabilities, to help them regain the movement and thinking ability of normal people, and the other is to help normal people become “supermans”.
The first role, in the current technology has been implemented cases. At the 2018 World Cup in Brazil, for example, a paralysed teenager completed the task of starting the game with the help of a brain-computer interface device.
Making “Superman” means that brain-computer interfaces can make normal people stronger, more powerful, and allow people to interact in a stronger way. For example, in the future, after the breakthrough of brain-computer interface technology, human communication can not rely on the mouth, can be communicated through consciousness.
The task of created “Superman” should be accomplished through brain-computer interface technology. Han believes that mankind still has a long way to go.
Taking the intrusive route of brain-computer interface as an example, because the structure of the brain is very complex, there are strong synergies. Even if the researchers performed cranial surgery in one area, other movements and instructions would require cooperation and information from other regions of the brain. This means that just implanting a chip in one area may not be enough.
Brain safety or eternal life.
Brain-computer interface technology focuses on parsing the human brain, which means that in the future our brains will become as clearly visible as all the digital information available today. Will this be the future we want to see? Can we keep this part of the information safe?
“At the moment, almost all of our identity, call information, fingerprints, and appearance are recorded. The brain is the last virgin land of our privacy. Han said. But he isn’t worried about brain privacy, “just as governments and international organizations can do well to protect these existing data privacy, I’m sure in the future we’ll have a lot of standards to protect the privacy of the human brain.” “
But for the average normal person, will they really be willing to accept the brain-computer interface technology, enhance themselves?
“In a Harvard study, passers-by were randomly asked on campus if they would be willing to undergo cranial surgery, and the number of people who were willing to answer was zero. Normal people don’t want to have cranial surgery. But people with disabilities, the people who have to use them, may choose. Han said.
According to Musk’s vision, brain-computer interface technology may actually eventually help humans achieve “eternal life” that preserves our memories and brain information.
At this point, Han believes that this should be after the brain interface successfully helped humans become superman to consider the issue. At the same time, he believes that eternal life is actually divided into different understandings. “There may be different schools of understanding of eternal life in academia now. For example, the human body cannot hold for thousands of years, but with the help of technology, our brains can be preserved, and they think it is a kind of eternal life. But people are free-conscious, and that’s the core. For example, a clone of yours, he went to touch a stone you can not feel. So he’s not you, he has nothing to do with you. So eternal life is the same, when you can completely replicate a brain that’s been preserved for 1,000 years, but what does it have to do with you, he’s just a brain, and your body can’t feel it anymore. Unless one day someone suddenly hits him and you feel sore, it’s another topic.
BrainCo wants to be the underlying platform for brain-computer interface technology.
Two days before Musk’s launch, BrainCo officially unveiled its latest product, brainOS Human Brain Intelligence Operating System, in Boston, USA.
BrainOS Brain Intelligent Operating System is a system developed with BMIoT (Human Internet of Things) as its core concept. According to the introduction, the system can be divided into two parts, one is the application of brain-computer interface technology, so that users through the system control of smart home, car, mobile phone and other external devices;
“It’s an underlying platform that standardizes our algorithms and technologies for so many years. Our goal is to be able to connect countless brain signal collector hardware and numerous input devices to our platform in the future. This system will become an underlying decoding platform, which is what is at the heart of Our BrainOS. Han Yu.
In view of the current development of brain-computer interface technology is still in the early stages, Han said that to let the brain-computer interface commercial landing, may encounter many difficulties and challenges. For example, the company has been doing the standardization work in the early days.
“The pre-EE is a very professional way of collecting, before in the hospital those better collectors may be about 100,000 yuan, and his EE collected signal is actually very weak. This is also the reason why many of the previous brain signal collection companies died, because it has no way to standardize, his good products may not even be 10%, there is no way to produce in the factory. Han said.
For BrainCo, a lot of time was spent in the early stages to overcome material standardization, acquisition standardization, and the entire calibration test. “This part of us is actually experiencing very big difficulties, but the later product once made, because it is to solve the real life problems of people. For example, to help disabled people do prosthetics, to help students improve their grades, to this stage will be easier. Han said.
BrainCo has launched a commercially available product based on invasive brain interface technology: Focus Focus Training for Smart Education and BrainRobotics Smart Bionics for People with Upper Limb Disabilities. The first of these two products can use neurone feedback training to improve cognitive ability, track concentration data, and quantify and improve it. The product has been trialed in many well-known high school and university laboratories in North America, such as mitisan media labs.