Elon Musk elaborates on "Brain-Machine Interface" for the first time: The ultimate goal is to "merge humans with AI and avoid being surpassed by AI"

Elon Musk's brain-computer interface welcomes new progress. Following the successful completion of the first human brain device implantation surgery, Neuralink will welcome its second human implant recipient. Musk revealed that the company hopes to implant its devices into "high single-digit" patients this year.

Musk has mentioned multiple times before that Neuralink can help paralyzed patients regain mobility. Restoring previous functions is just one goal, Neuralink aims to give people "superpowers," allowing those with implanted chips to have abilities stronger than normal individuals. During a live stream, Musk mentioned:

Neuralink will launch two products. The first product is Telepathy, which helps people with damaged neurons regain physical functions. The second is a brain care device, similar to an Apple Watch with tiny electrodes, implanted to rewrite brain signals and help people better utilize their brains.

Regarding the first implant recipient, Musk stated that only about 15% of the threads in the first patient's implant were functioning properly. Despite this, the patient was able to use the brain-computer interface (BCI) to watch videos, read, play chess, and other electronic games, sometimes spending up to 70 hours per week using it.

Musk mentioned that if you place your hand on the top of your skull, you won't feel any device. Even if someone is bald, you won't really notice it. From the internal contour of the skull, from a physical perspective, the brain doesn't really notice a depression in the skull because there isn't one.

Musk:

Yes, it won't be long before a person with a neural connection device can communicate faster than a fully abled person. So the goal is to give people superpowers. We just want to make it clear that if only one person has a Neuralink chip in their brain. So for those who think we've put a ship in their veins, want to assure you, or its value, you may not believe us, but we haven't put each other in your brain.

Musk:

Alright, welcome to the Neuralink live stream. We will update you on the progress of the first implant recipient and then talk about what changes we have made for the second patient. We hope to implant the next human in about a week. This is our first product, called Telepathy, which allows you to control a computer or phone by just thinking. So let's, in fact, start with some introductions. DJ, you can begin.

DJ:

Hello everyone, my name is DJ Seo. I am an electrical engineer and chip designer, leading several generations of Neuralink designs. I am currently a founding team member and the current president.

Matthew Mcdoul:

I am Matthew Mcdoul. I am a practicing neurosurgeon, serving as a neurosurgeon at Neurolink.

Head of Applications:

I develop some software engineering, trying to figure out how to turn brain activity into cool things in the world Elon Musk:

Let's see, so we'll start the demo. So our first product is similar to what we call Telepathy, which allows people implanted with Neurolink to control their phones or computers by just thinking. Once you can control your phone and computer, you can basically control almost anything, just by thinking. So there's no eye tracking or anything, purely your thoughts.

Elon Musk:

So, this is a very profound device that can help many people who have lost the connection between their brains and bodies. Imagine someone like Stephen Hawking, you know, imagine if he could communicate at the same speed as someone who still has the connection between their brain and body. So, really, they can help millions of people around the world.

This is part of our overall goal, which is to achieve very high connectivity between your brain and your computer, as well as other parts of the world. The ultimate goal is to reduce the civilization risks of AI by establishing a closer symbiotic relationship between human intelligence and digital intelligence. But this will take many years, and we will help address many brain or spinal cord injury issues.

Elon Musk:

For our first product, Telepathy, this is again very profound. If there are damaged or a few neurons that can bridge the gap between the brain motor cortex and the spine, allowing someone to use their body, then it is also possible to bridge the gap in the long term. Again, I think that will be very exciting. You know, in the long run, it is possible.

Then after our second product, we have demonstrated working with monkeys, which will allow a completely blind person to regain sight with two eyes or completely restore their object neurons. So this is something we hope to demonstrate in the future. So this is just to give you an idea of what this device is, a way of caring for thoughts, a bit like a Fitbit or an Apple Watch with tiny wires or electrodes.

These tiny wires are implanted in the brain and they rewrite the electrical signals. So many people think the brain is a very mysterious thing. It is mysterious in many ways, but it does operate with similar electrical signals. So if you can read and write these electrical signals, you can connect with the brain, and the size of these devices makes them the same size as a removed piece of skull. So it's like a few centimeters in diameter piece of skull removed.

After implanting the tiny wires with a surgical robot, we replace it with a device that gives neurons the ability to rewrite completely wirelessly. It is completely wireless, so like I can happily have a neural link now that senses charging. So you can basically have an electromagnetic device that you use to charge the device. So, yes, it's like an Apple Watch.

DJ:

In addition to this being actually a more difficult technical challenge, consider how much brain tissue is limited.

马斯克 (Elon Musk):

So in our case, it will also undergo scanning and other checks. So charging and bandwidth communication is a bigger challenge because it needs to go through something inherent in the skin.

马斯克 (Elon Musk):

So, the first step of our mind control is basically to unlock digital independence for people with policies, allowing them to control computers with their minds without moving their bodies. Our goal is to provide them with the same level of control, functionality, and reliability as when I use a computer, even better than the level of control I have.

Head of Brain Interface Applications:

That's not a high standard for you. He controls this with his brain, so unlike you, you can't see his hands in this video, but he doesn't use a mouse and keyboard, he doesn't think about how to move the cursor.

马斯克 (Elon Musk):

A few days later, the cursor moved here, I think the way he described it is also.

Head of Applications:

Yes, he has more videos on the platform. Be sure to check it out. Yes, he can do this, he is a person who can speak and move his head in real time.

Head of Applications:

Yes, if you join this live broadcast, you can ask him questions. He will tell you how it feels.

Head of Applications:

Also, I don't think I've played "Civilization" myself. I think this is actually not an easy mode. This is for experts.

Head of Applications:

The key is that during the live broadcast, it is a challenging task to play the most difficult mode, and he can do it while moving.

DJ:

Another game he likes to play is chess. I think sometimes he is actually playing speed games, which is confusing.

Head of Applications:

It takes high-fidelity control and speed control to win.

Another cool thing about our device is that we can use it anytime, anywhere, on a plane, in flight, while trading very cool cash means. Our device can also unlock things that previous participants could not unlock. For example, we can connect him to a game console and play "Mario Kart" with friends and family, happy to see them playing together years later.

Head of Applications:

He likes to play with the device and independently use the browser to watch videos, read, and play games every day. The key metric we care about is to ensure that our device is truly useful, basically the number of hours we use the device every day and every week. We track it once a week after surgery. In a few weeks when he is not too busy or traveling, he can even complete 70 hours of device fusion in a week. This is amazing. Of course, he wants to use it more, but needs to run some sessions. Sometimes he needs to sleep, and of course, occasionally needs to charge the device. Hopefully, we can improve this over time.

Neuralink Executive:

I think, I mean, it's not obvious to people watching this, like he's controlling a regular MacBook. It's not like some limited, efficient thing with very few choices. Just like you can do anything on a MacBook Pro. In fact, it's exactly the same Neuralink executives:

Perhaps another interesting point is that on the first day, he used BCI, then controlled it. He was able to break the previous word record controlled by the cursor just by using his brain. Recently, he even doubled it and was able to exceed our engineers' neural link by about 10%. You can be sure that we are doing very well in this competition and very fast. If you want to see how well you can do, you can do it on our website, which is a very addictive game.

Musk:

Yes, this is a very simple game. You just need to click on the blocks. Even though it sounds silly, it can actually be quite addictive. And especially if you get a low score, you want to have no slow way, I mean, anyone who wants to try this, I suggest going to Neuralink. See if you can break Nolan's record. You will find that it is actually very difficult to do.

Musk:

This is actually the first version of the device, with only a small number of electrodes working. So this is really just the beginning, but even at the beginning, it's twice as good as the world record.

Musk:

This point needs to be emphasized. You know, the media has a habit of saying the cup is 10% empty, but in reality it is 90% full. So I think Neuralink has made quite a big achievement, the first patient, the first device, is twice the world record in all range computer bandwidth. This is really an amazing, amazing great result, and it will only get better from now on.

Musk:

The potential is to ultimately reach the terabit level. This may be a long-term goal to improve brain-machine interface bandwidth. If you consider how low the bandwidth is typically between humans and suggestions, the average bandwidth is extremely low, but what I'm talking about is less than one bit per second in a day. If there are 86,400 seconds in a day, then in very rare cases, the number of bits you output to any given device is less than this number. So this is very important for AI, for humans, symbiotic AI is the ability to communicate quickly. Yes, I can follow.

Application Director:

Yes, to emphasize again, these performances are at a very high level, with about 15% of his channels effectively threaded, so we want to alleviate the problems that lead to this situation.

Application Director:

The brain is a fascinating organ. Share with you some secrets about the brain in any typical brain surgery. A small amount of air is introduced into the skull. This is because neurosurgeons like to have as much space around the brain as possible. So there is a little-known control mechanism that allows the brain to expand or contract based on where you locate carbon dioxide. But typically, neurosurgeons shrink the brain by lowering carbon dioxide. In future surgeries, what we want to do is keep the carbon dioxide concentration fairly normal, or even slightly elevated, which will keep the brain at a normal size and shape during surgery.

Application Director:

This should eliminate the bubbles we saw in the first participant. We believe that when the bubble migrates below the implant, that air pocket may have caused some thread slack, pushing the brain off the implant So this is easily solved.

Elon Musk:

Another consideration we want to focus on is that for our upcoming participants, the brain has very complex folding creases. It is a layer-by-layer neural work table that covers the surface of the brain, folding into this strange shape. The brain's folds delve deep into the brain, followed by neurons. If we insert near one of these folds, there may be very useful information encoded in the neurons there, and we may end up running parallel to some of the neurons we are most interested in, completely avoiding them. To avoid this possibility, we will insert closer to the middle of the fold in future participants, ensuring that we cross the layers of interest, specifically the 5th cortical layer.

Daniel Zhang:

I also think those tiny wires that Musk mentioned are important, they are just a small part similar to human hair. They are very flexible because, you know, the brain is constantly moving, and you want the electrodes to move with the bricks, resulting in fewer marks. Human neurosurgeons, no matter how talented Matthew is, are actually unable to manipulate them, right? So we created a surgical robot that can precisely target any three-dimensional space, X, Y, and C, with micrometer precision, while avoiding certain areas so as not to interfere and cause immune reactions. So we actually have the technology to place them precisely where we want them.

Application Supervisor:

Yes, it's truly a remarkable achievement to see the brain surface after the robot inserted all the electrodes into the first participant without a drop of blood.

Yes, I'm sorry. What most people may not realize is that the brain seems a bit undifferentiated. So if you look at the cortex, it looks like a bunch of folds, you know, maybe it is. Just looking at a picture of the brain is not obvious, the brain is highly differentiated, you can almost exactly know where the brain controls your right hand, left hand, legs, and those kinds of things or parts of vision. It is actually very precise in its location. So some people might think, look, the brain is like, it could be anywhere. But in reality, our brain is highly differentiated, even if it doesn't look like it.

Application Supervisor:

So, yes, we can put a patient considering implanting this implant into a functional magnetic resonance imaging machine, have them imagine hand movements, you know, due to spinal cord injury, these movements won't happen, but just imagining these hand movements will cause these brain areas to light up in the fMRI scanner. So we have a good idea, in fact, based on each participant, which part of their brain will respond to the imagined actions of the hand. Therefore, we can map these imagined actions to control a cursor on the screen, just like we do when moving a mouse.

Elon Musk:

Yes, but anyway, I think it's an important point that the part of your brain that controls your hand is not likely to be anywhere in the cortex, it happens in a very specific area and is very common among people.

Daniel Zhang: Precision is key. Left-handed, right-handed, it's the same in my mind. For example, if you are right-handed, you want the device on the left. Yes, just like the latter.

Application Supervisor:

Another risk mitigation measure we are considering for the future. The implant has a certain size, and the depth at the bottom of the implant is actually thinner than the normal human skull. So what we want to do is control the size of the gap below the implant, allowing the threads entering the brain from the implant to be as relaxed as possible. We did not do this in the first participant because we did not want to manipulate any of their tissues, we absolutely did not need to.

Application Supervisor:

In the upcoming implant, our plan is to deliberately carve the surface of the skull to minimize the gap below the implant, so that the bottom of the implant is completely flush with the normal contour inside the skull. This will bring the implant closer to the brain, eliminate some tension on the gap, and we believe this will reduce the tendency for thread contraction.

Musk:

We actually built a tool, which is a very important detail. You really want the internal contour to be flush with the skull. The brain doesn't like to frown in gaps, which is really important. This is like minimizing the conflict between airbags and implants with the internal contour of the skull, these are two very important improvements for us.

DJ:

The additional benefit here is that you do see some traces, we call them incisions. So your head will have a slight discomfort, but this actually further eliminates it.

Musk:

Yes, what I mean is, our goal is that if you put your hand on the top of the skull, you won't feel any device. Even if someone is bald, you won't really notice it. From the perspective of the internal contour of the skull, from a physical point of view, the brain doesn't really notice that there is a depression in the skull because there is no depression.

Matthew:

Another aspect of the human brain, you know, that is significantly different from any animal we have tested is that the human brain is much larger. So you may not realize that this means that the movement of the human brain is much greater than any smaller brain of any other creature. Therefore, when we open the skull, we see that when the heartbeat and breathing occur, the brain as a whole moves in and out of the robot by about 3 millimeters. So this movement, you know, it adds a small challenge to the robot in accurately selecting the depth of insertion for each thread. It's not a huge challenge, we have upgraded the robot's ability to more precisely target depth, even with very fast movements of the brain.

Application Supervisor:

You might think that the most obvious relief for threads pulled out of the brain is to insert them into deeper areas. We think so too. Therefore, we will expand the depth range for inserting threads. For the first participant, we have a lot of animal research data. We have highly optimized the insertion depth to maximize the intersection of the layers of interest in the cortex with the electrodes we record Now that we know that retraction is possible, we will insert at different depths, even in different scenarios with different numbers of retraction threads. We will set the electrodes at the appropriate depth and use the deepest threads to track how much retraction has occurred on the brain surface. So we will have more threads in the correct layer and better data to illustrate how much impact retraction has.

Matthew:

If you are a BCI, you may know that being able to control the individual depth of each thread is not something most neural interface devices provide. Most neural interface devices are a static, fixed, rigid array where you push it in, and all electrodes are at depth. Being able to do this is actually a very novel part of robotics.

Elon Musk:

It sounds pretty crazy. Yes, just with pneumatic cameras. It sounds a bit barbaric. This is not what we do, but it has been done before, especially with hammering nails into the brain to look better, and it can actually work. Surprisingly, it actually works well.

But something like deep brain simulation is indeed effective, it actually helps people align. That's a great product. Yes, but what I mean is, I think we can make a more refined version in the future. Like neural connection devices, it can really reduce brain damage and reduce the burden on patients. The goal is to enable someone to live a completely normal life.

You wouldn't even notice that someone has this device. So as I said, regaining the ability to control your computer and phone is our goal. And the next device will allow people to see things they couldn't see before. In fact, you can show people things like the Dorado furnace in "Star Trek". Infrared, yes, infrared, ultraviolet radar. So I think another way to put it is that we want to give people superpowers. So not only do we restore your previous brain function, but you actually have greater abilities than normal people, which is a huge thing.

DJ:

I also think, you know, we are often asked why you have to really get into the brain? But if you put it on the surface or outside the school, in short, how it works, you really need to let the sensors, these sensors facing the sources in the brain, which neurons, get as close to it as possible. Otherwise, you cannot achieve the level of control we think of.

Elon Musk:

Yes, maybe a good analogy would be like this, if you want to understand what's happening in a factory, you need to walk into the factory. You can't just put a stethoscope on the wall and try to figure out what's happening. Trying to read things from the outside is like a person putting a stethoscope on the wall of a factory, trying to understand what's happening inside the factory, which is ineffective. You have to be inside, so the threads have to be inside. But I want to emphasize again, the goal is to give people superpowers, not just to restore previous functions. I think this should bring hope to many people in the world, and the future will be exciting and inspiring, as technology will give them superpowers. That's amazing Matthew:

Yes, one of the most exciting parts of this story is that we can do a lot with 15% of the threads, while you have more channels. This actually gives you not just faster mouse control, because in the motor cortex, neurons don't represent the same thing. If you try to understand what a finger is trying to do, you may or may not have electronic text. The more channels in your brain, the higher the representational or decodable quality of all your fingers on your hand.

So, if you want to output text quickly, it's very important for those who are completely locked in, who can't speak at all, who are trying to say "I love you" to loved ones in their family, or ask for a glass of water or something else, to be able to type faster. The more fingers you can access, the more likely you are to be able to do this effectively. So, I'm very excited about how high the ceiling we can reach.

马斯克 (Elon Musk):

We are currently at about 10 bits per second, but ultimately we hope to reach megabits, I think eventually achieving a whole brain interface is possible. I think gigabits are possible in the long term. So, it's quite amazing.

Now, this is still about version 1 of the device, as we mentioned, it's version 1, with only 15% of the threads working. The current device has 64 threads, with 16 electrodes on each thread. Our next device has 128 threads, with 8 electrodes on each thread. Because as we gain more confidence in how and where to place the electrodes, threads require fewer electrodes. So basically, without substantial changes, if we place the threads accurately, we potentially double the bandwidth. Then the next generation device may have more threads, aiming for 3000 electrodes. So it will get better and better, I think rising by an order of magnitude. It won't be long before a person with a Neuralink device can communicate faster than a person with a fully functional body, faster than the fastest typist or auctioneer.

Neuralink Executive:

This may be a very interesting part of it. Basically, we currently connect standard inputs to computers and keyboards. Soon we will have a higher bandwidth interface, and we need to consider new ways to actually build the interface.

马斯克 (Elon Musk):

That's a good point. Because the current input devices are hand-centric, we have these mobile devices, like mice and keyboards, and Xbox controllers. But you actually no longer need these traditional control mechanisms. If you don't want to use your hands, you don't need these controllers, so ultimately I think you will be able to convey the entire uncompressed concept to others through neural links.

Neuralink Executive:

Even today, we still have some issues, like, if you don't feel the click of the mouse under your finger, how do you know it really happened? Because, you know, you see it on the screen, but you're not actually there. And you don't have the proper individual feedback, you know, the keys or touchpad on your fingertips. So there are all sorts of interesting user experience challenges to really understand what their decoder is actually doing, what early errors are actually doing Matthew:

Yes, it's Bluetooth, just a Bluetooth connection, just like your regular Apple mouse or Apple Magic Keyboard connecting to your computer. Exactly the same thing. In fact, yes, if we want, we can basically expose it as a hidden interface. HID is just a name for the protocol used to send mouse actions to the computer, I can basically input anything.

DJ:

What I mean is, I think we chose that interface because it's ubiquitous. Yes, basically any device has Bluetooth functionality. Our long-term goal is to have our own protocol, you know, secure and reliable. But for now, you know, choosing it is for interoperability.

Matthew:

So the question is, can Neuralink provide long-term treatment for paralysis? We can't do that now. We have done preliminary work, implanting a second neural link in the spinal cord, and we can restore natural appearance and leg movement in animal models. But this is not something, don't hold your breath waiting for it, it will take some time. We have a lot of work to do, but theoretically we have no reason not to fix paralysis.

Elon Musk:

Basically, there are no physical barriers to completely solving the paralysis problem. This may be a way to put it, your motor cortex sends signals, and if they are transferred to the damaged nerves, it's basically a communication bridge. So, you can bridge the communication from the motor cortex to the damaged point in the neck or spine, and from a physics perspective, restoring full body function is possible. It's a very difficult technical problem, but from a physics perspective, it doesn't prevent it from happening.

In terms of the next stage of expansion, we really want to make sure that we make as much progress as possible between each neural link, between each patient. So, we are now only moving to our second patient. But we hope that, if all goes well, we can have high single-digit numbers this year. This depends to some extent on regulatory approval and the technological progress we make, but we hope to have thousands in a few years.

DJ:

I think one thing that needs to be emphasized is, you know, we didn't just make one device and one surgery. We did a series of surgeries. We made thousands of devices, even just to explore the ability to dig into any low-frequency fault patterns. So we have invested a lot of money in infrastructure to be able to expand in terms of device manufacturing and surgery. We hope to help as many people as possible as soon as possible, through the appropriate barriers, namely regulatory challenges.

Elon Musk:

Device implantation really needs to become almost fully automated, like LASIK eye surgery, without an ophthalmologist holding a laser cutting machine. That's too crazy, but the Alpha model just supervises the basic machine and ensures the settings are correct, then the machine does everything and restores your vision. Many people have regained their vision through LASIK, which is really amazing.

I think there's also a surgery called a smile, they keep making it better. We need something like neural implants, so you can sit down and do any upgrades or brain repairs needed. This is reviewed by medical experts We obviously want to ensure that the review is correct, but it does need to be automatic. You sit down, and within 10 minutes you will install a neural device, very quickly. I mean, it's very much like a cyberpunk experience, like Deus Ex, if you play those games.

Neuralink Executive:

It's a good question to start interacting with other devices like wheelchairs. We are currently focusing on electronic computing and unlocking independence in the virtual world. Our plan is, as we mentioned before, to unlock independence with mechanical arms and wheelchairs in the physical world. Of course, if you make a computer, there will be additional risks, but we are working with the FDA to request approval.

Elon Musk:

If a wheelchair has an application, then the wheelchair only needs to have an interface. So if the wheelchair has a wheelchair interface, you can connect to the wheelchair via a Bluetooth interface, which is something we should do soon.

We can limit the speed so as not to create real disasters. We need to take it slow at the beginning, so that it can really be sorted out. Neural devices should generally be applicable to anything with a Bluetooth interface.

We can also talk to Optimus Prime, you can send directly to it, or if someone loses their ability to speak, they can still communicate with Optimus Prime. They can have a telepathic connection with Optimus Prime via Bluetooth, so even if someone completely loses their ability to speak, they can still control Optimus Prime or their computer or phone.

Similar things, like eating, you know, if you need someone to feed you, it's hard to eat with friends, it's a normal social experience. So if you can feed yourself, pick up a fork, eat pizza, chicken by yourself, yes, that's a big deal. It can prevent and save a lot of interaction with caregivers and others who depend on care in life.

I think in the long run, an exciting possibility is if you combine a part of Optimus Prime humanoid robot with neural link, assuming someone loses an arm or leg. Well, we can actually connect an Optimus Prime arm, Optimus Prime leg and perform neural link implants, so that large, movement commands from your brain will go to your biological arm, now to your robotic arm or robotic leg. Again, you basically have telekinesis.

Neuralink Executive:

The latency from neural link to your hand may be slightly faster than it is now. So you can imagine, if you are a pianist or anything that requires extremely fast hand movements, you may actually have a fairly unbalanced right mobile arm control with left physical arm control, that's one of the functions.

Elon Musk:

Yes, as future cyberpunk games say, you will have cyber upgrades better than your biology. As we expand to a large number of customers or patients for neural linking, the understanding of the brain will greatly improve. Because today we don't have a very fine-grained understanding of the brain, it's just that the sensors are not good enough. fMRI is quite good, but it's still not as good as having high-bandwidth electrophysiology in the brain Matthew:

Yes, I think this is underestimated as a research tool, which can drive the entire effort to truly understand the physical essence of human thought. We have not yet reached the required level of understanding. So Neuralink is actually a very powerful research tool.

马斯克 (Musk):

Yes, I think we can eventually understand and repair quite serious mental illnesses, such as schizophrenia.

Matthew:

Just like the illusion that they have chips in their breath.

马斯克 (Musk):

Yes. I just want to make it clear that currently only one person has a Neuralink chip in their brain. So, for those who think we have already implanted chips in their brains, I want to assure you that we have not done so. If in the future you wish us to implant a chip in your brain, which may help address the issues you think you have in your brain, then we can do that.

马斯克 (Musk):

Some people suffer from severe schizophrenia, their brains are malfunctioning to some extent. This is actually due to physical circuit problems, you can think of the brain as a biological computer. If some circuits are crossed, it will crash, or there will be other problems.

With neural devices, we can solve these problems and help those with brain diseases or mental illnesses lead a normal life. I think this is one of the things that may happen in the future. You can completely imagine, for example, some people have Alzheimer's disease, their memory is not as good as before, sometimes they forget who they are, who their grandchildren are, or what the date is, these are the problems that Neuralink can help solve.

DJ:

Yes, I mean, it's actually for personal reasons, like you're losing a part of your identity.

马斯克 (Musk):

Yes, this is just a small glitch in the biological computer that can be fixed, but essentially it's a short circuit.

Neuralink Executive:

Regarding how the device is charged, how long does the charge last? The current version takes about 45 minutes to charge and can last for 4 to 5 hours when fully charged. What we learned from Nolan (the first implanted person) is that this is actually one of the main limitations when he uses it more. In fact, it is quite difficult to use the product for more than 70 hours a week.

For example, sleeping for about eight hours every night is more than the time spent in bed. It takes 56 hours a week. So we are confused about 70 hours.

马斯克 (Musk):

But it is worth emphasizing that, for example, our first implanted person Nolan used the device for 70 hours in a week.

Neuralink Executive:

You may not like me sharing his computer usage publicly, but I am sure it is only used for productive things. One thing we learned is that in the next version of the device, we need to double or increase the battery capacity. Charging time can extend the battery life by twice, meaning you should be able to use it for about 8 hours.

DJ:

The goal of charging devices is to actually use everything they use, so you can charge in your sleeve or while sleeping Elon Musk:

Indeed, once you use it for 16 hours, you basically have 24 hours of usage time because it can charge while you sleep.

Executive:

There is another very important thing that I think needs to be pointed out here, if you are paralyzed, you cannot put the charger on your head by yourself. So it is important to consider the convenience, it is not just about better usage duration, can you charge independently? We spent a lot of time thinking about how to make it feasible, so you can charge the device without needing anyone's help.

DJ:

There is a larger coil charge, you know, that big. We actually put it in a packaged sleeve. Yes, or a knitted hat. Then he puts it on and commands it with his voice.

Elon Musk:

Regarding the arduous writing work, Nolan's current device is reading his worst actions. It is also pointed out that in the future, we have developed a second implant that allows the other hand to be used with higher precision. You can play games, it's actually electric pulses, not like reading neuron electric pulses, which is crucial for vision. So vision and writing, this is just a trade-off of the brain's visual part and electric pulses. We have this in monkeys too. We have been working with it and monkeys for a while, where we can flash a pixel and observe the monkey's reaction. The monkey's eyes will fly to that position.

You just need to verify the better triggering pixels on the right side of the brain, the initial resolution of vision will be relatively low, a bit like Atari graphics. But over time, it is possible to have better vision than normal.

DJ:

And I guess some additional applications, writing to the brain could be useful, like order applications, with feedback, appropriate feedback sets, and a prompt feedback.

Especially on the mechanical arm, if you want to grasp, you need to understand it.

It's a very delicate balance, not only to initiate movement but also to get feedback and control it accordingly. So there is a sensory cortex, near the cortex for movement.

Elon Musk:

So after inserting the device, any changes in nerve growth, we have not seen any signs of nerve damage, but, I think we have seen some rebound on the electrodes, right?

Matthew:

Actually, when you are about 25 years old, the plasticity of the whole brain decreases. But each insertion causes a little damage, but compared to anything else, it's just negligible. It's an easily recoverable amount of damage, actually only detectable after the animals are no longer alive and observed under a microscope, you can't really judge if there is any damage in your life.

Executive:

Another way to explain this issue, any changes in nerve growth after inserting the device, is like users learning how to use the device. I think there has been tremendous progress in this regard. Nolan spent hundreds of hours trying to figure out the best way to use this device because he really believed that if he could figure it out, he could help share this knowledge. He started a meeting on Friday night at 8 o'clock to figure out how to elevate his performance to a new level. This is truly a unique learning process because there are not many people in the world who have had the experience of moving something Yes, he really turned it into one.

DJ:

He's only talking about the working time in the past six months, right? Yes, in many ways. Like, I mean, he uses it while traveling and on planes, right? In fact, BCI has left a lot behind.

Musk:

I mean, one of the questions is how close we are to converting thoughts into text. Right now it's more about moving the curse from the screen to a virtual keyboard. But in the long run, you should be able to transmit the entire idea faster than anyone else.

Matthew:

Is the brain trying to naturally reject the threads? I mean, it's a common feature of any implant in the body, trying to reject it. The goal of surgeons and technical teams is to fight against it. So, through screws on artificial hip joints and spines, we have found biocompatible materials and technologies to fix these implants in the body after a certain age. It's hard to find someone without implants, you know, yes. They have some kind of screws on their spine. So we have solved this problem very well. So to answer your question, yes, the body tries to get rid of any implants, but we can make sure. Basically can't.

Is the brain trying to naturally reject these devices? This is a common feature of any implant in the body, trying to reject it. The goal of surgeons and technical teams is to fight against it. So, through screws on artificial hip joints and spines, we have found biocompatible materials and technologies to fix these implants in the body. It's hard to find someone without implants, they have some kind of screws on their spine. So we have solved this problem very well. Yes, the body tries to get rid of any implants, but we can make sure.

DJ:

It is worth noting that these threads have not actually moved in the past 5 months. Some may be pushed in, some pushed out, but more or less very stable. Once you have brain surgery, the tissue needs some time to enter, and then our organization or needle film can really enter and secure the threads in place. Once this happens, everyone is stable and sees a lot of changes.

Musk:

Yes, the threads are very small and important. The smaller they are, the less likely the brain is to react to them. So that's why we want the threads to be very small, while minimizing any damage to the neurons.

It's quite challenging because you're trying to read and write electrical signals, the threads themselves need electrical isolation and are not affected by body corrosion. So, actually making this electrode work while not eroding its performance over time is very difficult.

DJ:

The human body's environment is very harsh, like a bag of seawater. It's a well-regulated high temperature, and sensors are prone to corrosion. But I believe people have the experience of instantly throwing their electronic devices into seawater.

Musk:

We're about to end the live broadcast, if there are a few final questions, in terms of upgrades, we do think it's important to be able to upgrade the equipment. You don't want iPhone 1 to stay in your brain forever, you might want iPhone 15 instead of iPhone 1. So we will take out the neural link device and put in a new one, we have done this experiment on some animals, in one case, we upgraded a device three times, Then an upgrade was made to a pig.

We do our best to take care of animals. When we had an inspector from the US Department of Agriculture come over, she said it was the best animal facility she had ever seen.

Neuralink executive:

I mean, having breakfast on the app.

Elon Musk:

And a monkey ordered room service.

Elon Musk:

Monkey room service, that's rare. We are the only ones who provide monkey room attendants, so we really do our best to maximize animal welfare. So, well, thank you all for watching