- Neuralink, the neural tech company founded by Elon Musk in 2016, held a live demo Friday showing off its ability to read the brain activity of a pig with a surgically implanted chip that transmits data wirelessly.
- Musk described Neuralink's AI-powered "brain chip" as "a Fitbit in your skull with tiny wires" and predicted the technology could one day give people telepathy, cure paralysis, or enable superhuman vision.
- But scientists who watched the demo said Neuralink didn't show off anything that innovative, and that it's still a long way from using the chip in humans or delivering on Musk's ambitious claims.
- Here's what you need to know about Friday's demo, the technology and company behind it, and what the future holds for "brain-machine interfaces."
- Visit Business Insider's homepage for more stories.
What's the technology Neuralink is developing?
Neuralink, a neural tech startup Musk co-founded in 2016, has been working on a "brain-machine interface" consisting of a tiny chip implanted in the skull that can read and write brain activity.
Musk described the chip, which Neuralink showed off Friday during its live demo, as being something like "a Fitbit in your skull with tiny wires."
The coin-sized chip connects to ultra-thin, flexible wires — each one is around 5 microns thick, about 20 times thinner than a human hair — that contain a total of 1,024 electrodes and fan out within the brain. Those electrodes can read or, theoretically, write brain activity by sensing or stimulating neurons, all while transmitting the data wirelessly via Bluetooth-like radio waves so researchers can analyze it.
Neuralink has also developed a precision robot to surgically install the chip underneath the skull and feed the wires into the brain as to not damage it by accidentally puncturing blood vessels — a procedure Musk claimed takes just hours and leaves nothing but a small scar.
What does, or could, Neuralink's chip actually do?
Musk has, unsurprisingly, made some wild claims about the technology's potential.
According to his predictions, it could: create "symbiosis" between the human mind and computers; "shed some light on consciousness;" allow people to "save and replay memories;" cure paralysis, blindness, memory loss, and other neurological diseases; enable "superhuman vision," or give people the ability to summon their Tesla telepathically.
"Yes, this is increasingly sounding like a 'Black Mirror' episode," Musk acknowledged during Friday's demo.
While neuroscientists said Neuralink appears to be a long ways off from delivering on some of Musk's more ambitious promises, they're optimistic that brain-machine interfaces could in the not-too-distant future be used to treat some neurological issues like spinal cord injuries or Parkinson's, or be used to control prosthetics.
"Diseases where we understand the [neurological] circuit and what exactly that circuit is doing are potential applications," Dr. Jason Shepherd, an associate professor of neurobiology at the University of Utah, told Business Insider.
But other neurodegenerative disorders, like dementia, are more difficult to tackle, he said, because "there's complex behavior, there's complex learning and memory or aspects that are not regulated by just one area of the brain."
What did Musk show off Friday?
Neuralink treated viewers to a bizarre and slightly dystopian demo where it used its chip to analyze the brain activity of pigs, which are often used in research because their brains more closely resemble the anatomy of human brains.
Friday's demo involved several pigs: "Joyce," who didn't have a chip implanted; "Dorothy," who had one but has since had it removed, (which Musk said was important because it showed they could remove it if you change your mind or want to "upgrade"); "Gertrude," who has had a chip for two months; and several others with "dual" chips implanted.
Attendees then got to see and hear "real-time signals" from Gertrude's Neuralink, which was configured to detect "spikes" of neural activity in her snout so that the screen lit up and sounds played as she sniffed food around her pen.
Musk also showed a video one of the pigs on a treadmill where the company tried to predict the position of the pig's limbs by sensing its brain activity, which it claimed it did almost perfectly.
Of course, working on pigs is still several major steps removed from human subjects, both in terms of technical hurdles and regulatory approval. On that front, Musk said Neuralink had gotten "breakthrough device" approval from the Food and Drug Administration, which can help speed up the research and development behind new medical technologies.
Finally, Neuralink's team provided some limited technical details about the materials used in the chip and wires as well as the speed and resolution at which the chip can read and write data.
There was a lot hype around the demo — was anything actually new?
Not really, at least as far as what people could glean just from watching the demo, though scientists and engineers would no doubt love to see Neuralink publish more research on its chip, materials science, and algorithms to support some of the hype Musk has created.
"All the technology that he showed has been already developed in some way or form," Shepherd said. "Essentially what they've done is just package it into a nice little form that then sends data wirelessly."
Neuralink's work builds on work that neuroscientists and bioengineers have been doing for decades, who in recent years have made breakthroughs such as mind-controlled robotic arms.
"If you just watched this presentation, you would think that it's coming out of nowhere, that Musk is doing this magic, but in reality, he's really copied and pasted a lot of work from many, many labs that have been working on this," Shepherd said.
Still, Musk's enthusiasm for the topic and deep pockets will likely generate additional interest, both from investors and the public, which could be a boon for other neuroscience work, he added.
Scientists have been impressed before with some of Neuralink's breakthroughs, most notably the "sewing machine" it developed to insert the ultra-thin wires into the brain.
"Doing this stuff by hand, these are very fine things … it's very hard to have a steady enough hand to do these things manually," Andrew Hires, an assistant professor of neurobiology at the University of California, told Business Insider in July 2019 after Neuralink published new research and claimed it had gotten a monkey to control a computer.
What challenges do Neuralink and other brain-machine interface researchers still need to overcome?
Depending on what they want to use the technology for, researchers and tech startups working on brain chips still face a range of challenges.
A big one is developing materials that can be safely inserting into a human brain for long periods of time without deteriorating themselves or causing infections.
Others include making chips that can read data quickly enough and at a high enough resolution to be useful, designing algorithms that can accurately interpret the data, and designing a way to insert the wires deeper into the brain so the chip can access more of its activity.
There are also major security, privacy, political, legal, and ethical issues that are bound to arise. Obviously, implanting a chip in someone's head that reads their raw brain data — and lets them control their physical movements, sight, smell, or hearing via a mobile app, uses Musk said Neuralink is pursuing — could go very wrong.
"A lot of these kinds of companies are sort of stuck in between the development stage and the application stage," Shepherd said. "And from my view, the development stage of this is I think very premature, there's still a lot of work to be done."
Neuralink could also face some of its own, unique challenges. STAT News reported earlier this week that several former Neuralink researchers expressed concerns about a chaotic culture within the company characterized by a clash between tight deadlines and the slower pace typical of scientific research.
They told STAT that the company is using tech's favored "move fast and break things" approach to building a medical device, and that "the strategy is very immature at times."