Key Takeaways from Elon Musk's Neuralink Presentation: Solving Brain Diseases and Mitigating AI Threat
"Hello world." Elon Musk's secretive brain-machine interface company, Neuralink, has finally given us an update.
Yesterday at 8 pm PDT, the company held an event that was live-streamed on YouTube and is now available on their website.
While "it will take time," Musk said during his keynote, he believes the company will allow us to solve brain diseases, stop AI from "leaving us behind," and ultimately, "achieve a symbiosis with artificial intelligence."
All you need to do is have tiny neuron signal-reading threads inserted into your brain by a robot. Here are the main takeaways from the Neuralink event.
Neuralink wants to hire the best talent in the world
"The main reason for doing this event is recruiting," Elon Musk said at the start of his presentation. "We want the best talent in the world, and that is a primary purpose for this presentation." They are looking for experts in neuroscience, mixed signal chip design, optics, machining and other fields shown in the presentation.
The Tesla, SpaceX, SolarCity and Neuralink man made sure to emphasize that Neuralink won't achieve their potential overnight. "There is an incredible amount we can do", but "all of this will occur quite slowly," he said.
Neuralink aims to understand and treat brain disorders
It wasn't just Musk who spoke at the event. Neuralink President Max Hodak, another one of the speakers, revealed how he initially didn't believe Neuralink was possible before being convinced by Musk.
"You have to be very careful telling [Elon Musk] that something's impossible. It better be limited by a law of physics or you're going to end up looking stupid," Hodak said.
Musk outlined Neuralink's ambitious plans throughout his talk. He suggested that, in the future, the neurotech company will be able to solve all sorts of congenital diseases and brain disorders with their chip that will be inserted, by a robot, into the brain.
"There is an incredible amount we can do to solve brain disorders," Musk claimed.
Brain-reading robot-inserted threads
The way that Neuralink will do this is by having a robot insert tiny threads into your brain via a very small incision. These threads are about a tenth of the cross-sectional area of a human hair, Musk said — roughly the size of a neuron.
"The robot looks through a microscope and inserts each electrode specifically, bypassing any blood vessel. The way the threads are inserted causes no noticeable trauma."
Neuralink's specially built robot will "very delicately apply" these electrode-width threads. The operation would involve just "a 2mm incision, which is dilated to 8mm, and the chip is placed through that." It's so small, that "you can basically glue it shut. You don't need a stitch. It's not stressful."
Neuralink want to eventually make the operation process as simple as LASIK eye surgery, using localized anesthesia. However, as per the New York Times, Hodak detailed the current requirements for general anesthesia and a series of holes being drilled through the skull.
Importantly, the interface to the chip is also wireless. "You have no wires poking out of your head," Musk said, comparing the technology to Bluetooth.
Their chip is "a thousand times better" than the current best system
"Everything that you perceive, feel, hear, think, it's all action potentials and neural spikes (impulses from neurons)," Musk said. That is what Neuralink's system picks up on. "Our goal is to record from and selectively stimulate as many neurons as possible across diverse brain areas."
For this, they designed a chip "which is 4 by 4 millimeters and has a thousand electrodes." The current best equivalent would be an FDA approved Parkinson's deep brain simulation chip, which would have an order of magnitude of ten electrodes, Musk said.
So, if true, the first version of the Neuralink chip, unveiled at the event, is capable of housing a thousand times more electrodes than the current best system.
They hope to mitigate the existential threat of AI
Neuralink's second goal is a little more abstract, however, it is focused on the growing concern of many experts on the dangers of AI.
"After solving a bunch of brain-related diseases, there is the mitigation of the existential threat of AI." This Musk says, is part of Neuralinks's plans for "creating a well-aligned future" where we won't become AI's plaything.
The company has set out "to help secure humanity's future as a civilization relative to AI."
"Even in a benign AI scenario," Musk said, "we will be left behind."
"And hopefully it is benign," he emphasized — Musk is one of many other experts and technologists to have signed a petition on regulating against the development of killer AI.
The Neuralink founder feels the best we can do is to keep up with AI rather than letting ourselves fall behind. "With a high bandwidth brain-machine interface, I think we can actually go along for the ride," he said. "We can effectively have the option of merging with AI. This is extremely important."
Neuralink "is not mandatory"
Ever hip to meme and popular culture, Elon Musk showed he is keyed into people's fears that Neuralink, and any other big AI company for that matter, could lead to a Matrix or Skynet-style apocalypse — as well as the real fear that it could lead to an immense centralization of power.
Musk clarified, with a knowing smile: "this is not a mandatory thing. This is a thing that you can choose to have, if you want it."
Neuralink's chip will act like a "tertiary layer to the brain"
At the event, Elon Musk spoke of the importance of the collaboration between the limbic system and the cortex in our brains. Neuralink, he says, will effectively add a third layer to this collaboration — "a tertiary layer, which is the digital superintelligence layer."
However, we "already have this layer", he countered: "it's your phone, and your laptop. The constraint is how well you interface the input and output speed."
"Output speed is especially slow because it's most people typing with thumbs these days. Input speed is much faster due to vision. The thing that will ultimately constrain our ability to be symbiotic with AI is bandwidth," Musk explained.
They hope to start human tests by the end of next year
Finally, Elon Musk said that Neuralink aims to start human tests soon.
"We hope to have this, aspirationally, in a human patient, before the end of next year."
The neurotech company also released a white paper detailing their research after the presentation. It is yet to be peer-reviewed.
Could the singularity be near? The first hurdles, for Neuralink at least, are FDA approval, human testing, and approval from the scientific community at large.
An interview with Dr. Birgül Akolpoglu allows IE to dig deeper into the potential, limitations and misconceptions of biohybrid microrobots for medical use.