Ready to have your mind blown? Dive into the incredible world of Neuralink, digital telepathy, and brain-computer interfaces! From mind-controlled speech to China's market-ready chips, this video unpacks the future of human enhancement and communication. Don't miss out—subscribe for more mind-bending science and comment your favorite part below! #technology #science #future #Neuralink #innovation
👉 This channel was created in collaboration with https://www.youtube.com/@realidadeimpressionante
🔊🎧Music Info: Provided by the channel "MokkaMusic" and by https://inaudio.org
0:01 - Neuralink’s Digital Telepathy Vision
1:08 - The Terry Case and Speech Synthesis
2:29 - How Brain-Computer Interfaces Work
7:01 - Technological Progress and Global Competition
11:26 - China’s Neo and Beinao Implants
15:42 - Toward Machine-to-Brain Communication
18:46 - Future Possibilities and Human Enhancement
19:55 - Musk’s Vision for Conceptual Telepathy
👉 This channel was created in collaboration with https://www.youtube.com/@realidadeimpressionante
🔊🎧Music Info: Provided by the channel "MokkaMusic" and by https://inaudio.org
0:01 - Neuralink’s Digital Telepathy Vision
1:08 - The Terry Case and Speech Synthesis
2:29 - How Brain-Computer Interfaces Work
7:01 - Technological Progress and Global Competition
11:26 - China’s Neo and Beinao Implants
15:42 - Toward Machine-to-Brain Communication
18:46 - Future Possibilities and Human Enhancement
19:55 - Musk’s Vision for Conceptual Telepathy
Category
🤖
TechTranscript
00:01Elon Musk promised that one day Neuralink could create a kind of digital telepathy between humans.
00:07It seemed like a pretty far-fetched idea, didn't it?
00:10But now Neuralink has just taken a real step in that exact direction.
00:15This man you're looking at is thinking of words and a machine is speaking for him using his own voice
00:21recorded earlier.
00:23And this might just be the beginning, okay?
00:26And scientists are already trying to do something even stranger, the reverse process.
00:31In other words, instead of just pulling information from our brains, now they want to send information directly into it.
00:38And if these technologies ever meet, we'll be entering the very future that Elon Musk has been describing for a
00:44few years now.
00:45Direct brain-to-brain communication, storing digital memories, and even senses that we as human beings have never had.
00:52Yeah, I know, all of this still sounds like science fiction, right?
00:56But how much of that future has already begun in our time?
01:00That's what we're going to figure out right now.
01:09Through Neuralink, Elon Musk had been promising things that seemed like real miracles.
01:13But until now, his most visible results had been focused on, for example, controlling machines with the mind.
01:19But now something different is happening, you see.
01:22Something that is really starting to bring technology closer to something he's been talking about for a long time.
01:29Directly turning what happens inside our brains into a machine-generated voice.
01:33The man seen in this demonstration is named Terry.
01:36In 2024, he was diagnosed with bulbar onset amyotrophic lateral sclerosis, a form of ALS that can early on affect
01:43the muscles involved in speech and swallowing.
01:46Fortunately, Terry can still move his body and his mouth.
01:49The problem is that his ability to speak has deteriorated over time.
01:53According to Neuralink, it got to the point where even the people closest to him began to have difficulty understanding
01:58what he was trying to say.
02:00But there was one thing the disease hadn't erased.
02:04Inside Terry's brain.
02:05The words were still there.
02:07And that's precisely what Neuralink is trying to access.
02:10Terry is participating in VOICE, a clinical study investigating whether the N1 brain implant can turn the signals associated with
02:16the intention to speak into text or directly into speech, as in this case.
02:20The study is still experimental, okay?
02:22And the device has not yet been approved or commercial use by the US FDA or other regulatory authorities.
02:29To understand what is happening, we need to go back a few milliseconds before a person speaks a word.
02:35When you and I decide to say something, regions of our brain involved in speech production generate the signals that
02:41typically end up controlling the muscles of the mouth, tongue and larynx.
02:45And in some neurological conditions, that chain is broken.
02:49The person knows what he wants to say, the brain produces the activity related to that, but the body can
02:55no longer turn those signals into intelligible speech.
02:58Neuralink's idea is to create an alternative path.
03:01Instead of relying on muscles, the implant tries to capture neural activity directly from the patient's brain.
03:07The N1 chip uses 1024 electrodes distributed across 64 extremely fine threads thinner than a human hair.
03:15They are surgically implanted to record neural activity.
03:18From there, artificial intelligence algorithms then attempt to identify patterns in this activity and convert them into commands, or in
03:25the case of voice, the ideas of communication.
03:28And Terry's training makes this whole story particularly fascinating.
03:31You know, in the beginning, he tried to articulate the words as best as he could.
03:35The algorithm observed the patterns produced while Terry attempted to speak.
03:39Little by little, the system began to associate certain neural patterns with the words he actually intended to say.
03:46Until the turning point arrived, Terry no longer needed to try to pronounce them the same way.
03:52He could simply think of the words he wanted to speak without even moving his lips and hear them being
03:57spoken.
03:58And even more importantly, in his own voice, which was cloned using audio files from when he could still speak.
04:05That very same week, Neuralink posted another video, a shorter one, showing the participant in front of someone close to
04:10him.
04:11And he doesn't speak a sentence in the conventional way.
04:13The implant recorded the signals associated with the intent to speak, the system interpreted that.
04:19And then, a voice said the following, I love you.
04:22Okay, yes, this is a very short demo, but there's a huge difference between what was done and simply selecting
04:28letters on a tablet screen.
04:29For example, speech carries identity, rhythm, and intonation.
04:34And above all, the feeling that that person continues to be heard, this program isn't exclusive to Neuralink.
04:40Other research groups have already demonstrated brain-computer interfaces that are also capable of synthesizing speech from neural activity.
04:48A study published in Nature, for example, showed a neural prosthesis capable of producing real-time speech in a participant
04:53with ALS,
04:54and even preserving features like intonation.
04:56Therefore, Neuralink did not invent brain-to-voice communication.
05:00What it's actually doing is trying to turn this capability into an implantable interface that is completely wireless
05:05and practical enough to be used outside of isolated demonstrations.
05:09And this is precisely where the most interesting part begins, you see?
05:13Because Neuralink's stated goal isn't just to have someone produce an occasional sentence.
05:18The company has set a target to get spoken communication up to approximately 140 words per minute in ut.
05:24And that starts to approach the speed of normal human conversation.
05:27Normal, not fast like mine, blah, blah, blah, I talk way too fast, I know.
05:31Imagine someone who has completely lost the ability to speak.
05:34The person looks at you, there's no movement in their lips, they're not using a keyboard, not selecting letters with
05:39their eyes.
05:40What they do is simply formulate what they want to say, and you hear their voice coming from a smartphone
05:46or another device.
05:48That is precisely where the phrase used by Elon Musk came from, digital telepathy.
05:52But there is a very important difference, okay?
05:55This is still not telepathy in the traditional sense.
05:58The device has not shown the ability to get inside someone's brain and discover their random thoughts.
06:03There is also no direct transmission of thoughts between two brains in this demonstration.
06:09Neuralink's very first product was named telepathy.
06:12And the initial idea was simpler.
06:14Paralyzed people would imagine the movements.
06:16And the implant would convert that neural activity into digital commands.
06:19Moving a cursor, clicking, typing, gaming, and even controlling the computer.
06:24Today, participants are already using this technology for everyday activities.
06:29And Neuralink has also been extending these experiments to assistive robotic arms controlled solely by the user's thoughts.
06:35So, pause right there for a second, and look at the progression of it all.
06:39First, thought became cursor movement.
06:41Then, thought started controlling physical machines.
06:44And now, thought is being transformed into words and voice.
06:48And the next step is already on Neuralink's trial roadmap.
06:52It's no longer just about taking information out of the brain.
06:55The idea is to try going the opposite way, the reverse path.
06:59It's to send information directly into the brain.
07:01But we'll talk about that a bit later in this episode.
07:04Right now, we're going to understand how this chip's technology is actually working.
07:13If what happened to Terry seems like telepathy, there is an inevitable question.
07:19How can a machine know what a person is, quote unquote, thinking and what they are trying to say?
07:24The answer starts with something that happens inside the brain all the time.
07:28Even before you move your hands, say a word or click on something, groups of neurons are already producing electrical
07:34activity related to the intended action.
07:36Throughout practically all of human history, that information remained trapped inside our body, inside our brain, to be more precise.
07:45To turn an intention into action, the brain needs to use the biological pathway.
07:49Neurons, nerves, muscles.
07:51But a brain-computer interface, a technology like that, tries to create a second pathway.
07:56And this is where devices like Neuralink come into the picture.
08:00I need electrodes to record the activity produced by populations of neurons.
08:04And just to be clear, the system doesn't find a written word inside the brain.
08:07There's also no specific neuron there saying right, left, or I want to speak.
08:11What exists are patterns of activity.
08:14And the algorithms are trained precisely to recognize those patterns.
08:17Imagine someone trying to move their hand, to the right for example.
08:21Even if an injury prevents their arm from responding, specific regions of the brain still produce activity associated with the
08:26intention of that movement.
08:28As when an AI comes in, repeatedly observing these signals, the person tries to move their hand, the system records
08:35what happened in the brain, and then they try again.
08:38That means more data and more patterns, until at some point the algorithm starts to establish a connection between specific
08:44neural activity and what the person intended to do.
08:47That's how an intention can begin to turn into a digital command.
08:51And this explains why the first results of brain-computer interfaces seem relatively quite simple.
08:56Moving a cursor, browsing the internet, controlling a video game, or moving a robotic arm.
09:00If we can even call that simple, right?
09:03The AI doesn't need to understand everything the person is actually thinking.
09:07It just needs to get better at recognizing a specific set of signals.
09:12But speech makes the problem much more complex.
09:15When Terry tries to say a word, there's an extremely fast sequence of neural activities that are related to the
09:22movements that would normally be needed to produce all of that.
09:25Tongue, lips, jaw, larynx, breathing.
09:28The goal of voice is to find enough information in that activity to reconstruct what Terry actually intended to say.
09:35And as the algorithms improve, the amount of information we can extract from these signals also increases, of course.
09:42And that's when the technology starts to take on a different character.
09:45Because controlling a cursor means turning brain activity into a few possibilities.
09:49Left, right, click, or don't click, if you want.
09:52But reconstructing a conversation means distinguishing a gigantic number of combinations in real-time.
09:58Words, phrases, rhythm, and intonation.
10:01And eventually even elements of the patient's own voice.
10:04And this explanation helps explain why what we are seeing now is so important.
10:09We are not learning to read our brains completely.
10:12What we are doing is getting better and better at decoding certain types of information produced by it.
10:18And there are different ways to do this, okay?
10:21Some interfaces use sensors outside the head, others place electrodes directly on the membrane that surrounds our brain.
10:28Meanwhile, more invasive systems like Neuralink, for example, place electrodes very close to or inside the brain tissue.
10:34And of course, there is a reason for that.
10:36Closer to the neurons, the more detailed the signal can be.
10:40But it also increases the complexity of the medical procedure.
10:43This balance between the amount of information and invasiveness is producing several different strategies.
10:49This is important because while the United States through Neuralink is trying to place thousands of contact points extremely close
10:55to neurons,
10:56another country is taking a slightly different path or different paths, right?
11:01And one of them has already achieved something that not even Neuralink has managed to do so far, okay?
11:06Turning one of these chips into an approved product to be sold to its citizens, maybe even on a large
11:11scale, right?
11:13Want to know which country this is? How will the population receive these chips? Will they accept them?
11:17Seek with me and I'll show you in a bit.
11:26As promised before the break, I want to say that this country is China.
11:30Yes, you guessed it right, if you thought of China.
11:32And no, they didn't decide to force the chip implant into the population's brains, okay?
11:37Their device is called Neo, and for the first time, this kind of technology is no longer restricted to clinical
11:42trial participants.
11:43It has now received authorization to reach patients who need established medical criteria.
11:48The approval happened in March 2026 and was granted by the Chinese regulatory authority to Neuralco Medical Technology, a Shanghai
11:55-based company.
11:56According to Nature Biotechnology, this is the world's first commercial approval for this kind of interface, an invasive brain-computer
12:02interface.
12:03But there's a very important difference compared to Neuralink, China decided not to try to get as close to the
12:09neurons.
12:10The Neo is roughly the size of a coin, and instead of inserting dozens of threads directly into the brain
12:14tissue, the device is placed over the dura mater, one of the membranes that protects the brain.
12:19It sits above the region of the cortex responsible for motor control.
12:22And this means that the captured signal, yes, is less detailed than the one obtained by systems that directly penetrate
12:28the brain tissue.
12:29But there is an advantage to all of this. The brain doesn't need to be punctured by the electrodes.
12:34With the Neo, the idea is relatively simple. A person with paralysis, for example, imagines moving their own hand.
12:40Even though the spinal cord injury prevents the command from reaching their muscles, the brain continues to generate the activity
12:45related to the intention of that movement.
12:47The eight electrodes capture these signals, and they are then sent to a computer where an AI is running.
12:52There, it identifies the intention, and the command is then sent to a pneumatic robotic glove placed on the patient's
12:58hand.
12:59The person then thinks about moving their hand, and the glove performs that movement for them.
13:04It's important to mention that to gain approval, the system underwent clinical validation involving 32 patients, which was conducted in
13:10collaboration with 11 Chinese medical institutions.
13:13And there's another very interesting detail in all this. The implant doesn't have a conventional battery that needs to be
13:19surgically removed in order to be replaced.
13:21A coil positioned on the outside of the head transmits energy via a magnetic field to another coil inside the
13:27implant.
13:27And that same connection allows brain signals to be transmitted outward.
13:31In practice, the team tried to eliminate both an implanted battery and cables permanently passing through the person's skin.
13:38And those are indeed two elements that could create problems in a device meant to stay inside someone's body for
13:44several years.
13:45In the brain, to be more exact.
13:47And it was precisely this approach that allowed China to reach a milestone first that Neuralink is still trying to
13:53achieve.
13:53Moving from experiments to entering the market.
13:56In other words, these chips can now be sold.
13:58But that doesn't mean any Chinese citizen or anyone who goes there can just walk into a hospital and ask
14:03for their own brain chip.
14:05Okay?
14:05The indication must be approved and quite specific.
14:08The new is intended for people between 18 and 60 years old with quadriplegia caused by certain cervical spinal cord
14:13injuries.
14:15Except that the new is only part of what's going on over there.
14:17Because China is also already developing another system called BENA-1.
14:22And this is where the scale starts to ramp up.
14:24Alright?
14:24In July, Chinese authorities reported that their system had already reached 16 implants.
14:30And the participants with the longest implantation time had already surpassed a year of use.
14:35Combined, the systems had accumulated more than 55,000 hours of operation.
14:39And the participants had already used the interface to control robotic arms using only their thoughts and the technology.
14:45And the goal has always been to work on reconstructing motor functions.
14:48The announced plan is to reach 36 implants later this year.
14:52And soon after, begin trials in qualified top-tier hospitals in 2027.
14:57And this reveals that the story is bigger than the race between Neuralink and some Chinese company.
15:02The government there has already placed brain-computer interfaces among the so-called industries of the future.
15:07Meanwhile, its companies, hospitals and research centers are advancing simultaneously across different technologies.
15:13So now, we are beginning to watch two strategies evolving in parallel.
15:17In the United States, Neuralink is trying to build a high-bandwidth interface capable of turning brain activity into increasingly
15:23complex commands.
15:25And now, even into speech, as we've seen, right?
15:28Meanwhile, in China, different groups are exploring approaches that seek to balance this capability, safety and clinical implementation.
15:35But both are essentially trying to solve the same problem to create a direct channel between the human brain and
15:41the machine.
15:42And in my opinion, this is exactly where this race starts getting really interesting to watch, you know?
15:47Because so far, we've mostly talked about people controlling something that's outside the body, a cursor, a glove, a robotic
15:53arm, an artificial voice.
15:55But know that researchers are already working in the opposite direction.
15:59Instead of just asking, how do we get information out of the brain?
16:02They are now beginning to ask, how can we send information directly to the brain?
16:06And when that door begins to open, my friends, the brain-computer interface stops being just a way to control
16:11machines.
16:13It can start turning into a way to alter what a person can truly perceive.
16:18Pretty strange, huh?
16:26And Muralink is already trying to do exactly what we just saw.
16:30The project is called Blindsight, and its goal is to use a brain implant to try to restore some level
16:36of visual perception to blind people.
16:37But there's a detail that makes this technology especially interesting, right?
16:41The idea is to create vision without depending directly on the eyes.
16:45When we see something, our eyes are just the beginning of the process.
16:49The retina turns light into electrical signals.
16:51That information then travels through the optic nerve to the areas of the brain responsible for visual processing.
16:57And it's only there that those signals begin to transform into what we perceive as an image.
17:02And Blindsight tries to explore precisely this other end of that pathway.
17:07According to them, a camera will be able to capture the environment,
17:10an external device will turn that into data and the neural link in the person's brain would directly stimulate the
17:16visual cortex.
17:17In other words, the machine would attempt to deliver an image directly to the brain.
17:21In 2024, the FDA granted Blindsight the breakthrough device designation.
17:25And this facilitates the interaction between the company and the agency during development.
17:30This doesn't mean that the technology has been approved for clinical use.
17:34Okay?
17:34There's another detail that's important for me to mention.
17:37If or when this technology actually works, we shouldn't imagine someone immediately regaining vision similar to ours.
17:43What happens is that one of the foundations studied in this field are phosphines,
17:47perceptions of small points or flashes of light produced when certain regions of the visual system are stimulated by something.
17:54The big challenge is being able to control many of these stimuli and turn them into patterns that the brain
17:59can actually interpret as useful information.
18:02Perhaps, initially, this could allow distinguishing shapes, outlines, objects or even help someone navigate through their surroundings.
18:10But the principle behind it is much greater.
18:12Because with Terry, we saw a machine trying to learn how to interpret the signals that are produced by our
18:16brain.
18:16With Blindsight technology, the problem is reversed.
18:20It is the machine that needs to figure out how to produce the signals that the brain can actually interpret.
18:25And if humanity truly manages one day to master both of these paths, brain to machine and machine to brain,
18:31we will have built something very different from the interfaces we use today, alright?
18:36We will no longer be talking merely about controlling computers with our thoughts, using our mind.
18:41We will be talking about a technology capable of exchanging information directly with another human brain.
18:46In other words, a chip installed here could talk to a chip installed in your brain.
18:51That raises one final question, right?
18:53Today, all technologies are primarily developed to restore what a person has lost.
18:57Movement, speech, vision, for instance.
19:01But what if or when we can do that well enough to take it a step further?
19:07Will we continue using brain interfaces only to restore lost human capabilities?
19:12Or will we start using them to create capabilities we've never had before?
19:23That possibility mentioned at the end already has a name, human enhancement.
19:27And it's precisely here that we need to separate what these technologies can already do from what they might one
19:33day achieve.
19:36Right now, there is no brain implant capable of turning a healthy person into someone more intelligent,
19:41instantly downloading knowledge, or even storing everything we experience.
19:44Current implants are still trying to solve very fundamental, much more specific problems,
19:50recognizing an intention, moving a cursor, reconstructing a word.
19:55But Elon Musk believes this story could end up very far from this beginning.
20:00One of the ideas he has championed for several years and openly talks about is what he calls conceptual telepathy.
20:06The logic is interesting.
20:08When we think of something complex, we need to compress that idea into a few words, don't we?
20:12Then we need to pronounce those words slowly.
20:14Another person will listen and try to reconstruct in their own brain what we were thinking.
20:20For Musk, this is an extraordinary form of communication, extremely slow and limited.
20:25During a Neuralink demonstration, he speculated that these brain interfaces,
20:29when they are advanced enough one day, could transmit concepts or even thoughts much more directly
20:34without relying on conventional language, according to him.
20:37It would be quite different from the telepathic technology that exists today.
20:41Not just brain to computer, but eventually brain to computer and then to another brain.
20:46And Musk goes even further, you know?
20:48He has already spoken publicly about a future where brain-computer interfaces could record
20:54and also replay memories as well as store them as a kind of backup.
20:58And in an even more distant scenario, even transfer all that into another body or even into a robotic body.
21:04For example, in 2024, he talked about this possibility of storing memories once again.
21:09And he suggested that if this technology became extremely safe and offered superhuman abilities,
21:16hundreds of millions of people around the world could eventually choose to have one of those brain implants.
21:21You wouldn't have to be someone who's sick to get a chip put in.
21:24And if you like thinking about the future this way,
21:26do you know where you can find other futuristic ideas based on current technologies?
21:30Right here, in this book right here.
21:32This book was written by me and it took me a few years to write it.
21:35Okay?
21:36If you like fiction, take a look at an excerpt.
21:38The link is down here in the description and also in the pinned comment.
21:41Okay?
21:42The book is available on Amazon for purchase, both digital and physical versions.
21:46And this physical version looks gorgeous.
21:48Huh?
21:48I'd buy it.
21:49Check it out.
21:49If you like it, buy it.
21:50It helps us out too.
21:51Thanks.
21:52Bye.
21:52And see you next time.
21:54.