What I saw in the Neuralink Show and Tell Video

A summary of Elon Musk's fascinating Neuralink presentation.

---

Introduction

In a November 30 "show and tell", Elon Musk provided an update on his coming ambitions for the Neuralink team.

In case you're not familiar with Neuralink, they are producing a neural implant for surgical insertion into the biological brain. The implant provides a high bandwidth input/output platform for the brain through the use of wireless communications. In order to accomplish that, the team has also had to create a robotic surgeon that can complete the insertion safely and in a repeatable fashion.

In this video, Musk provides an overview of the project's current status as well as its short and long term goals. He says that the main purpose for this presentation is recruiting, and asks potential applicants to visit the link, neuralink.com/careers. He then hands the presentation off to engineering staff who go into a variety of details about the implant technology as well as the surgical and manufacturing techniques.

In the following sections, I'll cover some of the highlights that I picked out from the video. I definitely recommend watching Musk's overview at the beginning, as well as the last two sections on "Vision Applications" (1:37:09) and "Spinal Cord Applications" (1:46:42).

Here is the full video:

Background and Overview

One of the engineers in this show and tell presentation (DJ) says that he was motivated to get into this field by watching The Matrix, where Neo was able to learn Kung Fu by data transfer. After seeing that film, he says that he wanted to bring the capability to fruition in the real world.

According to the presentation, Neuralink is working on a device that will be a high bandwidth i/o interface for the human brain. Addressing concerns about mixing biology and synthetic devices, Musk argues that we are all - already - cyborgs because of our constant connectivity by way of the ubiquitous cell phone.

He says that this sort of connectivity is necessary for modern life, but suggests that it is too constrained by the i/o speed of our devices. In order for (augmented) human intelligence to stay ahead of an AI superintelligence and avoid eventual extinction, Musk argues that it is necessary to vastly increase the speed of our connectivity.

Musk also divides the project goals into short term and long term goals. In the short term, he says that the goal is to, "Solve debilitating brain and central nervous system ailments". For the long term, his goal is, "Alignment with artificial general intelligence". To me, this second goal is especially interesting in the context of some of his recent Twitter activity. For example, a couple days ago he tweeted

The intelligence of this hive-mind will improve significantly as signal/noise, effective cross-linking of tweets & speed of tweets all improve

And some time ago, he also tweeted

Buying Twitter is an accelerant to creating X, the everything app

Is it possible that he sees Neuralink and his Twitter acquisition as two prongs of a strategy to "help humanity compete with AI"?

At any rate, the talking point that the press has run with from his talk is Musk's claim that they are working with the FDA and expect to be about 6 months from starting clinical trials for the device in humans. For early applications, he lists two: giving sight to the blind, and restoring movement for paralyzed patients.

Early applications: Restoring sight and restoring mobility

With regards to restoring sight, one of the presenters - Dan - suggested that Neuralink has three characteristics which make it well-suited to the purpose. First, the device can inject current to stimulate cells in the brain's vision centers; Second, it can have an enormous number of electrons - which is important for high-resolution vision through a visual prosthetic device; And finally its electrodes can be inserted deep into the brain where they are needed in order to be effective.

Dan showed a fascinating video of a monkey who was trained with rewards for eye movements to a spot on a screen where a flash was observed. After training, the same eye movements could be observed when the appropriate cells in the monkey's brain are stimulated, even though no actual flash appeared on the screen. This suggests that it is possible to "project" vision into the brain, but it also made me wonder about the risk of malicious introduction of false perceptions.

On the topic of restoring movement, another engineer named Joey presented. Joey noted that paralyzation is frequently caused by a loss of signaling capability in the nervous system between the brain and the paralyzed portion of the body. The hope is that by implanting Neuralink devices on both sides of the signaling break, communication can be restored and the paralyzation can be reversed.

Joey showed a fascinating video of his own where a pig was observed to be receiving limb stimulation by way of a Neuralink implant, and the operator was able to induce the pig to contract or extend its legs. Beyond the technology, another interesting aspect of the video was that the pig had the ability to withdraw consent for experimentation simply by stepping outside a box of blue tape that had been marked on the ground. If the pig is inside the box, then it is receptive to receiving stimulation, if it steps out of the box, then experimentation halts.

As with the vision application, however, the possibility of malicious use occurs to me here, too. If the chip can be remote controlled from outside the body, then I'm not sure what the capability for abuse might be.

Engineering Notes

The team focuses on three axes for its production efforts: safety, scalability, and access to brain regions (ability to perform diverse functions). The current generation product is the M1 chip, which is roughly the size of the quarter with about 1,000 channels connecting to the brain. The hope is that it will scale up to 16K channels and that it will be able to operate for a full day without interruption for charging. The device is designed to have low power consumption, wireless communications, and wireless charging. Because it is fully implantable, it is basically invisible to observers.

An interesting facet of the testing is that they created an environment to accelerate device aging. When the device operates in that test environment, it ages at roughly four times the speed of an implanted device, so they are able to project its performance over longer spans of time than the device has actually been in existence.

Robotics Notes

The surgery required to insert these devices is very precise. The needle that does the insertion is only 40 microns in size, which is on the same or lower order as a human hair. Additionally, the number of neurosurgeons available in the world is fairly small.

Therefore, in order to have the surgery be safe, repeatable, and scalable, it was necessary to automate the insertion. This is demonstrated at 41:43 in the demo video. Basically, a series of threads are connected to the brain cells, and the implant chip is attached to the skull.

Some demo videos

These are demos that are floating around YouTube. Some are shortened excerpts from the November 30 video, others have been out there for a while.

• Monkey playing pong

• Monkey typing "telepathically"

• 2020 pig demo

• Charging demo

Conclusion

As with most technologies, this seems to come with a tradeoff of risks and benefits. The capability to restore sight for the blind (or even to provide sight to people who never had it) would be amazing, as would the possibility of restoring motion for the paralyzed. I'm sure there are a huge number of other benefits that are also possible. But, the idea of having a wireless connection to our bodies that could be exploited by hackers, opperessive governments, or other malicicious actors is worrisome.

One aspect of this that I haven't seen any commentary about is how it relates to Musk's plans for Twitter. In light of the tweets in the introduction, I definitely suspect that the two could be highly interrelated. Since taking over Twitter, two of Musk's areas of focus have been "free speech" and "faster performance". Both of those make sense in terms of a hypothetical competition with emerging AI superintelligences. Free speech would be important because it is necessary for fully informed decision making; and faster performance would be important so that the Neuralink augmented human intelligences could communicate with each other and reach equilibrium points at faster speeds.

I have often thought of the Steem blockchain in terms of an emerging hybrid intelligence between humans and machines, so it's not surprising that someone would think of Twitter the same way. In that context, the interesting dynamic that Steem has over Twitter is the ability to use blockchain rewards as a mechanism for guiding reinforcement learning.

Please let me know your thoughts on this fascinating topic!

Update: I later noticed another article on the same topic. For more, see Elon Musk hopes to implant his chip in a human brain in six months/Elon Musk espera implantar su chip en un cerebro humano en seis meses from @mauromar.

---

Thank you for your time and attention.

As a general rule, I up-vote comments that demonstrate "proof of reading".

---

Steve Palmer is an IT professional with three decades of professional experience in data communications and information systems. He holds a bachelor's degree in mathematics, a master's degree in computer science, and a master's degree in information systems and technology management. He has been awarded 3 US patents.

---

^{Pixabay license, source}

Reminder

Visit the /promoted page and #burnsteem25 to support the inflation-fighters who are helping to enable decentralized regulation of Steem token supply growth.