Neuralink, Elon Musk’s ambitious project to wire up the brain to computers, stepped out of the shadows Tuesday evening.
In a detail-laden presentation at the California Academy of Sciences’ Morrison Planetarium, the tech entrepreneur explained how his foray into brain-machine interfaces could pave the way for a symbiotic relationship with artificial intelligence.
Clinical trials could start as early as next year: “We hope to have this, aspirationally, in a human patient, before the end of next year. So this is not far,” Musk said.
Pricing wasn’t announced in the presentation, the primary purpose of which was recruiting. Neuralink is hiring in the areas of robotics, materials, electrochemistry, micro-fabrication, histology, mixed-signal chip design, optics, and more. The company’s job board on its website lists eight engineering openings and a talent acquisition position.
Little was known about Neuralink prior to the presentation, bar a multi-page explainer published on WaitButWhy in April 2017. At that time, it seemed the company was exploring a variety of methods for linking up brains and machines, and it would initially focus on healthcare benefits as a way of funding further research.
“I feel like I’m in Transcendence,” Musk joked, in an otherwise science-focused event that lacked the party vibe of some Tesla events. “Actually, I was in Transcendence!”
Over two years later, and that vision has come into sharper focus. Neuralink has a product, a means of wiring to the brain, and even an iPhone app to plug into the existing world of hyper-connected technology. Company employees noted the rate of advancement versus existing solutions.
“Elon has this incredible optimism where he’ll pierce through these imagined constraints,” Max Hodak, president of Neuralink, told the audience. “You have to be very careful telling him that something’s impossible.”
Here’s what we learned:
6. Neuralink Has a Product
First, a quick primer. Scientists have developed a number of ways to interface with the brain, which reads the electrical firing of neurons in different ways. Some are non-invasive, like EEG scans that use a helmet, but they can be rather imprecise. Others are invasive like ones used in surgery, but the probes can be large and cause issues. Neuralink has opted for an invasive approach that uses a small chip to read the brain, with minuscule probes weaving their way through.
The product is called the N1. It’s a chip that sits in a hermetic package, which fits into a cylinder measuring eight millimeters in diameter by one-fourth of a millimeter tall. Each chip measures four millimeters by four and uses 1,024 electrodes. By comparison, designs used for Parkinson’s today can use just 10 electrodes.
All electrodes have read and write functionality. The probes are five microns thick, three microns thinner than a red blood cell and 95 microns thinner than a human hair. The design enables the probes to get close to neurons to detect spikes, and the team believes that the probes can rest 60 microns away from a neuron to detect the spikes.
In initial setups, Neuralink places four N1 chips in a patient, three in motor areas and one by the somatic sensory cortex. They’re then wired to an inductive coil near the ear that connects to a link that sits on the outside of the skin. The link contains the battery and Bluetooth to power the system, making it possible to remove and upgrade the firmware without actually touching the sensors again.
Neuralink went through a variety of prototype designs, including ones with a USB-C port. As the goal is to make it as safe as possible for surgery, the team had to compromise on more ambitious designs with triple the probes:
5. It Could Reach Patients Very Soon
As mentioned before, the first trials will focus on healthcare. The company aims to host the first-in-human clinical study trial before the end of next year, focusing on patients with quadriplegia due to C1-C4 spinal cord injury.
This will use the four-chip setup to enable patients to control their smartphone using their brain. Through that, they can control a mouse and keyboard on a computer through a Bluetooth connection.
Timescales will vary depending on regulatory approval. Musk previously stated in April 2017 that it may be around eight to 10 years before it’s available to people without disabilities.
And yes, as previous evidence suggested, the company has been using animals in its testing. President of Neuralink Max Hodak said that “we wish that we didn’t have to work with animals,” explaining how the firm takes careful consideration over its approach to tests. Musk noted that the team enabled a monkey to control a computer with his brain.
4. Surgery Will Be Like Lasik
These probes are incredibly fine, and far too small to insert by human hand. Neuralink has developed a robot that can stitch the probes in through an incision. It’s initially cut to two millimeters, then dilated to eight millimeters, placed in and then glued shut. The surgery can take less than an hour.
The goal is to make the insertion about as complex as Lasik eye surgery, making it easy to link up with machines.
3. It Has an App to Bring It All Together
It uses an iPhone app, designed in partnership with MetaLab, to interface with the neural link, using a simple interface to train people how to use the link.
“You have no wires poking out of your heard; very important. It basically bluetooths to your phone,” Musk said. “We’ll have to watch the app store updates for that one, make sure we don’t have a driver issue.”
No word on an Android version yet.
2. A Brain App Store? It’s Possible
One of the most intriguing comments came during the question-and-answer session, where an audience member asked about third-party software running on the pod. With read-and-write abilities, it’s potentially a tricky area of development.
“Conceivably there could be some kind of app store thing in the future,” Musk said.
Hodak noted that any creations couldn’t use an ad-supported model. While ads on phones are mildly annoying, ads in the brain could be a disaster waiting to happen.
1. A Symbiotic Relationship Is Still the Goal
Neuralink may be initially focused on healthcare benefits, but Musk noted his goal is still to link up humans with A.I. Musk compared it to using a smartphone, except making it a more direct link instead of telling the brain to move fingers to interact.
“This is going to sound pretty weird, but [we want to] achieve a symbiosis with artificial intelligence,” Musk said. “This is not a mandatory thing! This is a thing that you can choose to have if you want. I think this is going to be something really important at a civilization-scale level. I’ve said a lot about A.I. over the years, but I think even in a benign A.I. scenario we will be left behind.”
Details around the economics of the setup are still sketchy, but Musk joked that “if you want to be symbiotic with A.I., I think it’s safe to say you could repay the loan with superhuman intelligence.” Perhaps a funny suggestion, but research suggests that intelligence does not always predict financial wellbeing.
Far from wiring up and worrying about the details later, the really smart move may be to wait and see how Neuralink develops further over the coming years. Based on Tuesday’s presentation, it could be a fascinating ride.
Watch the full Youtube video below.