Last night, Elon Musk’s Neuralink held a media event to show us all what they’ve been doing for the last couple of years. After much secrecy and tons of speculation, they finally gave us a peek at what they’re doing. They’ve made amazing progress, but still have a long way to go before achieving their stated goals. However, along the way, they’re going to be doing a lot of good for suffering people.
This is a complex topic, and this article will only give a very basic overview. In the interest of brevity, I’m going to leave a lot out. For those wanting to learn more about the topic and go in depth, I highly recommend reading Wait But Why‘s super long deep dive on the topic and then watching the whole recorded livestream from last night‘s event. If you read all of that, you’ll have a pretty good idea of what’s going on and why it matters.
For the rest of us who don’t have time to read something the length of a book and then watch hours of video, I’ll do my best to summarize what’s going on and why it matters for the future of our species.
The “Why” Of Neuralink
Over the course of the last year, Elon Musk has dropped a hint here and there about Neuralink. The above video from the Joe Rogan Experience gives a quick synopsis of the “why” behind Neuralink. In short, artificial general intelligence (or “AI”) presents a great risk for humankind. Many works of fiction explore the bad possibilities, including the Terminator movies and many others. If AI gets far more intelligent than humans, we would be at a distinct disadvantage. At best, we might end up like “pets”, while at worst we could go extinct.
To help increase mankind’s odds of surviving and thriving, Musk proposes an “if you can’t beat ’em, join ’em” approach. By upgrading our brains and making AI and computers a part of ourselves, we could achieve symbiosis with superintelligent AIs, ensuring humanity’s survival…hopefully.
To understand how this is possible, we have to look at the brain a little. At the smallest level, the brain is made of neurons, which (put entirely too simply) pass electrical impulses around amongst themselves to do what the brain does. At the largest level, we have a primitive “mammal” brain that deals with basic life functions and instincts, and a more advanced “neocortex” wrapped around the outside that deals with more intelligent and “cerebral” tasks the brain performs.
Outside of our skulls, we have a third layer, or as Elon Musk says, a “tertiary layer”. It’s our digital selves that exists here on the internet. We connect to it through our computers and smartphones, store information in it, and can basically do a lot of superhuman things like talk to people around the world, share images, and have our private data stolen to influence elections. Compared to even the President of the United States decades ago, we all have greater access to information and even power.
But there’s one big problem–bandwidth!
I type at 80-100 words per minute, and I’m faster than most. When I’m on my smartphone, I take two pieces of meat and tap it against some sensitive glass to slowly type. We can read a lot faster than we can write, but the connection between us and our digital self is not far off of the old 56k dialup days in terms of speed of information transfer. If we can’t get our biological selves into faster communication with our digital selves, we can’t hope to compete with artificially intelligent beings that exist totally as digital selves.
What Neuralink Is Doing To Fix This
This is where it gets complicated. Really complicated.
In short, they have to send and receive electrical signals from different parts of the brain, and that’s anything but simple. The brain is an extremely complex organ, and just poking wires into the brain (if you can do that safely), doesn’t put out a signal that you can send to a computer. You also have to send signals into the brain, but the brain wouldn’t know what to do with signals from your computer (assuming they didn’t just kill or severely injure you). Even if you learn the electrical language of the brain, there’s a host of medical issues that can come up with something as complex as the brain.
Finally, not least of all, there’s the issue of people’s distaste for surgery and experiments that could look like a horror movie. We can’t just cut people’s heads open and put in a bunch of wires. To make it something people would even want to do, we have to make the process simple, pain free, easy to heal from, and minimally invasive. If the process of getting a brain-computer interface installed resembles being assimilated by the Borg, you’ll find few volunteers.
Neuralink has been working on all of these issues for two years, and have come a long way. The result? The N1 Implant concept.
In short, they have developed an application-specific integrated circuit (or ASIC) that can help interpret electrical signals from over 1,000 sites in the brain, all at the same time. Attached to the bottom of the chip (and hermetically sealed to it) is a ribbon of small “threads” that are smaller than a human hair.
When removed from its protective packaging, you can see the chip and the “threads” compared to a penny for reference. Once ready, a small 2mm incision is made in the patient’s skin and a very small hole is made in the skull. This does not require loss of hair or large scars. A special surgical robot is then used to unwrap the threads from their sleeve and place them extremely accurately into the brain.
Once these sites are chosen, the machine is more capable than any human hand of placing them in just the right spot, avoiding blood vessels and other things that can cause injury to the brain. When placed, it appears like the below image:
Several of these chip systems are placed at different parts of the brain to send and receive electrical signals, which are interpreted by the chip at the end of the threads. When done, the surgeon secures the chip to the skull and glues the small incision shut for healing under the scalp. Wires go to a site behind the ear, which can wirelessly communicate to a bluetooth module.
Finally, you can use an iPhone to interact with the implants:
It’s Going To Be Years Before Availability
The system is nowhere near ready, though. They’re planning some human trials by the end of next year, but will only start with people who would stand to benefit the most from a basic neural interface. The first goal (other than safety) is to help people who can’t move or who have other serious nervous system issues. This will give them a chance to start improving lives while the system is still very basic and can’t do much.
From there, they’ll move to treating simpler and simpler problems people have until the system is very mature. At that point, they’ll start making it available for people who want it on an elective basis, or in other words, people without medical problems who just want it because they want it.
The technology still has a lot of hurdles to overcome, but the presentation shows that they are making real progress. Not only are they working on solving the basic scientific problems to implement the technology, but they’re keeping things in mind that would impede later progress, by designing for those problems from the beginning.
Featured image credit: Neuralink