Nick Statt reports in The Verge:
In the neuroscience community, there is a consensus that enhancing both AI and human cognition are complementary goals. "The thing that would demand the greatest minds of our generation, i(s) the ability to co-evolve with artificial intelligence.” Elon Musk has begun putting together a team to explore human augmentation, first for medical purposes and inevitably for human enhancement.
For Bryan Johnson, the founder and CEO of neuroscience startup Kernel, the question is when, not if, we all have computer chips inside of our brains. Kernel, founded last fall with more than $100 million of Johnson’s own money, is trying to better understand the human brain, so that we may one day program it to improve.The company is focusing first on medical applications, to gain a deeper understanding of the diverse and complex ways the brain can fail. Eventually, Johnson would like to move toward augmenting the organ to make us smarter and healthier and pave the way for interfacing directly with computing devices.Johnson, who made his fortune selling his payments company Braintree to PayPal for $800 million in 2013, doesn’t have past experience in neuroscience. He is, however, riding a new wave of interest from Silicon Valley. There is a growing fear, among some futurists and other Silicon Valley elite, that humans will develop a crippling dependence on machines and software that continue to rapidly accelerate beyond our capabilities and understanding. This is a fear not necessarily shared by the neuroscience community, which is less focused on enhancing human intelligence, at least right now, than they are on treating people with Alzheimer's and helping paraplegics regain movement.Yet the goal of Kernel, ultimately, is to allow humans to outcompete or at least co-evolve alongside machines — by becoming a little digital themselves. Kernel has made some big claims: promising to improve neurodegenerative disease, for instance, to help pave the way for improving cognition. But for the last decade, brain implants have only dealt with movements, and have typically only been used in paraplegic people beyond experimental medical trials and stimulation devices for conditions like epilepsy.“We know if we put a chip in the brain and release electrical signals, that we can ameliorate symptoms of Parkinson's,” Johnson tells me. “This has been done for spinal cord pain, obesity, anorexia… what hasn’t been done is the reading and writing of neural code.” Johnson points to the programming of yeast cells and CRISPR gene editing as examples of breakthroughs that apply the principles of computing to living organisms. “What I wanted to do was work with the brain the same way we work with other complex biological systems like biology and genetics.”Of course, our understanding of genes is much farther along that our understanding of the brain. “Frankly, the technologies we have for interacting with the brain are blunt tools at best,” says Blake Richards, a neuroscientist and assistant professor and the University of Toronto who focuses on how the brain modifies itself and learns from experience. “Most neuroprostheses involve dropping a big array of electrodes into the brain.”This makes Johnson’s vision sound both difficult and distant, with a laundry list of scientific obstacles standing in its way. He will need more money — he’s currently declining outside investment but may take venture capital funds in the future. The project also requires time, perhaps decades, to achieve anything close to Kernel’s cyborg vision, which currently resides only in fiction. But despite these hurdles, Johnson is intent on starting now with Kernel as one of the early leaders in an emerging hybrid field, one that blends the cash-flush, experimental spirit of Silicon Valley with the most cutting-edge neuroscience research.Brain hacking, so to speak, has been a futurist fascination for decades. The idea that we will, inevitably, have chips in our brains and ways to interface directly with computing devices has been a staple of the most seminal cyberpunk works, from William Gibson’s Neuromancer to Masamune Shirow’s Ghost in the Shell to the Wachowskis’ The Matrix. The reality, however, is far more complicated — and dangerous. Very few people in the world have multi-electrode arrays implanted inside their skulls today. Those who do only undergo the invasive surgery required as a last resort, to alleviate the symptoms of severe neurological conditions or as a way to restore movement to paralyzed patients or allow amputees to move prosthetic limbs.Richards is skeptical of any company promising advancements that require invasive surgery. “People are only going to be amenable to the idea [of an implant] if they have a very serious medical condition they might get help with,” he adds. “Most healthy individuals are uncomfortable with the idea of having a doctor crack open their skull.”Johnson is first to admit the difficulties Kernel must reckon with to even begin working on these types of technologies, principally the idea of working exclusively with patients who have severe neurological conditions. He says that working with brain implants is a requirement right now. “There’s no tech that exists in the world that allows you to be outside the brain and gain access to critical data,” he says. “You need to be inside the brain, inside the skull.” Down the line, Kernel would like to explore less invasive ways of working with the human brain.Yet even then, moving beyond the medical field and into the realm of improving cognition requires a significant amount of scientific progress, Richards points out. “We understand very little about the human brain compared with what we understand about the mouse brain,” he says. “Almost all of our data on the human brain comes from epileptic patients, which is problematic for understanding how the brain works at large.”To really understand the brain, Richards adds, will take years of work. We’ll need to hone how we gather data from the brain — itself a challenging task with its own complications — and improve our understanding of how the brain carries out core functions. From there, researchers will still have to work within the confines of ethical medical trials and regulatory boundaries that restrict how and to what effect we can work on human brains. As it stands today, Richards says, we don’t even yet have have a thorough grasp of how the brain does everyday tasks like storing information we can recall later or letting us conjure conversations from years in the past. “The computations and algorithms carried in the brain are still largely mysterious to us.”These challenges haven’t stopped Johnson from setting his sights on neuroscience as the next frontier. While companies have in the past tried to make commercial headway in the field of neuroprosthetics, Johnson is focusing instead on investing in research that may yield new insights into the brain. He may be one of the first to pour a Silicon Valley fortune into the field, but he suspects others will follow in his quest to transform the brain as a computing platform, even if it takes years of research and billions of dollars of investment. For Johnson, those stipulations are just part of the deal. Money has always been a means to an end for the 39-year-old entrepreneur. After he sold Braintree to PayPal, Johnson decided that what he did next had to have the maximum positive impact possible. So he began talking with friends, experts, and fellow tech industry contemporaries, trying to discover where and for what his wealth could be best used to explore.After talking with hundreds of people, Johnson says he decided that neuroscience had the most potential. “Intelligence is the most precious and powerful resource for humans,” he says. “We’ve always built these tools, starting with the rock, thermostat, calculator. Now we have AI. Our tools and [digital] intelligence are increasing at great velocity. On the flip side, human intelligence is just about the same as it’s always been.So Johnson enlisted the help of some of the best scientists in the field to start looking into neuroprosthetics. These are devices implanted within the skull that mimic, substitute, or assist functions of the brain, ranging from controlling the motor cortex to preventing the onset of seizures. Johnson’s idea, at least at first, is to have his team at Kernel explore and better understand core brain functions like information recall, memory, and neuronal communication.To do this, the company is developing its own hardware and software to try and alleviate the devastating effects of neurological and degenerative diseases like epilepsy, dementia, and Alzheimer's. It’s being aided greatly by the research and expertise of Theodore Berger, a professor of biomedical engineering at the University of Southern California. Back in 2002, Berger’s research proved that it was possible to use software and mathematical modeling to replicate the hippocampus, which is the part of the brain responsible for memory and its eventual degradation. Nearly a decade later, Berger’s lab at USC used a chip implanted inside the brain of rats to restore lost memory and improve information recall.Now, Berger splits his time between USC and Kernel as the startup’s acting chief science officer. Kernel itself, now a little more than 20 employees, operates out of Los Angeles, near Berger’s lab where the team can collaborate with the biomedical engineers there and observe the scientists’ work. Kernel plans to gather data from human trials, with an implantable medical device not unlike the one used in Berger’s animal trials back in 2011.To help Kernel and aid in its longer-term efforts, the company has also scooped up Kendall Research Systems (KRS), a spin-out of the Massachusetts Institute of Technology that focuses on neural interface devices for use in research and clinical trials. As part of the deal, Kernel is bringing on KRS founder and CEO Christian Wentz. Johnson has also courted some other big names in the neuroscience field from the MIT community. Ed Boyden, a professor of biological engineering and brain and cognitive sciences at MIT, has signed on as chief scientific advisor. And Adam Marblestone, a neuroscientists who focuses on improving data collection from the brain, is now Kernel’s chief strategy officer, having worked in the past with Boyden’s Synthetic Neurobiology Group.“I can’t agree more that these things are all possible,” says Chad Bouton, a biomedical engineering veteran of the Battelle Institute and now the vice president of advanced engineering and technology at the Feinstein Institute of Medical Research. “What I often say is we are trying to figure out how to crack the neural code in the human body. If we can crack the neural code, then we can unlock so many doors.”Bouton says that we’ve already made substantial progress in figuring out how the motor cortex drives the function of limbs. “We can crack the code in the motor area of the brain,” he says. “But if we could crack the code in the rest of the nervous system, and understand these messages passing back and forth, we would be able to better diagnose and treat diseases.”In the future, however, Johnson has grander ambitions beyond medical treatment. He wants to use these implants — and hopefully, one day, make the process of receiving them less invasive — to augment human intelligence. He envisions a world where the human brain is made smarter, faster, and more creative. Most importantly, however, Johnson sees a world where humans, and not just machines, improve over time.Artificial intelligence may soon displace millions of jobs and render obsolete the livelihoods of everyday workers or, in the minds of some more outlandish technologists, induce a doomsday event for the human race. This is another driving force behind the creation of Kernel.“I think if humanity were to identify a singularly thing to work on, the thing that would demand the greatest minds of our generation, it’s human intelligence,” Johnson says, “specifically, the ability to co-evolve with artificial intelligence.”It is for this reason that Tesla and SpaceX CEO Elon Musk has begun putting together a team of his own to explore the possibilities of human augmentation, first for medical purposes and inevitably for human enhancement. Last week, Musk dropped hints of his interest in human enhancement by telling a crowd at the World Government Summit in Dubai by saying that “we will probably see a closer merger of biological intelligence and digital intelligence.” His new venture, however, remains relatively under wraps for now, with a public announcement sometime soon.“As far as I know, Elon and I are the only two pursuing this from a commercial perspective,” Johnson says. “That’s fantastic. I’m so happy that he’s in the game.” Johnson notes that the number of calls he’s received from interested investors has increased since low-key chatter of Musk’s plans began circulating in the Bay Area late last year.Even in the neuroscience community, there is a general consensus that enhancing both AI and human cognition are complementary goals. “The current success in AI came out because of their mimicking of the ways the brain operates,” says Richards, who himself studied AI before transitioning to neuroscience research. “There’s a building cross-talk between AI and neuroscience whereby AI takes inspiration from neuroscience and neuroscience takes inspiration from AI. Slowly but surely we’re working toward a broad theory of intelligence, both artificial and natural.”Whether Kernel helps the humanity achieve that broad theory — and goes even further beyond — will largely depend on how it decides to use Johnson’s money, and whether the hurdles of scientific progress impede the founder’s bold vision of the future. “We’re entertained by Black Mirror, but outside of that, we’re not discussing [human intelligence] as a populace,” Johnson says. “I’m trying to get the best minds of our generation in government and tech and media to talk about this problem. Brain science is the new rocket science.”