TMTPOST -- Today is my last day at CES 2025, and one thing has impressed me. Elon Musk appeared online for an interview. Besides discussing AI and the colonization of Mars, another topic Musk continuously brought up was the significant breakthroughs in brain-machine interface (BMI) technology.

He emphasized that his brain-machine interface company, Neuralink, has already implanted BMI chips in three human patients. Last year, Neuralink also showcased an experiment where a monkey played a game using its device, demonstrating that the technology could decode brain signals and control external devices. This marked a key breakthrough in the "neural language decoding" part of BMI technology.

Many might not fully understand "language decoding." In simple terms, it is a technology that decodes language information by reading and interpreting brain activity. To put it plainly, it is about transforming brainwaves into communicable language, allowing for "silent communication" simply through the fluctuations of brainwaves. Without accurately achieving "language decoding," any human-related goal that BMI technology aims to accomplish would be nearly impossible.

However, while Musk’s Neuralink is still experimenting with monkeys and has just implanted BMI chips into three patients, a remarkable event took place in January 2025 in Shanghai, China. In a hospital room at Huashan Hospital, a female patient’s brain signals were successfully decoded by a Chinese-made brain-machine interface technology, producing the Chinese characters "HappyNew Year 2025" and simultaneously issuing a command that made a robotic hand perform a "heart-shaped" gesture.

This marked the world's first instance of using brainpower to send New Year’s greetings. The technology was developed by NeuroXess,, in collaboration with the Neurosurgery Department at Huashan Hospital. Through their original invasive flexible brain-machine interface technology, they were able to not only perform motor decoding in real-time but also decode language in real-time.

Just five days after receiving the flexible brain-machine interface surgery, the patient achieved a 71% decoding accuracy rate for 142 common Chinese syllables. The decoding speed reached 40 Chinese characters per minute, with a latency of less than 100 milliseconds for single-character decoding. This breakthrough also directly connected brain signals to a large language model, reaching the highest level of Chinese language decoding.While people were excited about Musk’s progress, the breakthrough that occurred in Shanghai has been significantly underestimated and overlooked. In fact, the Shanghai team has surpassed Musk’s Neuralink’s monkey experiments by a huge margin, achieving true global leadership in this field.

As early as 2020, TMTPOST published a cover story titled "Brain-Machine Revolution: Musk Turns Left, Chen Tianqiao Turns Right." At that time, Shanda’s founder Chen Tianqiao told TMTPOST that “the new brain electrode developed by Tao Hu from the Chinese Academy of Sciences’ Microsystem Institute, which we support, is actually progressing faster than Neuralink.” By that time, Chen had already shifted most of his focus toward supporting research in brain science.

Five years later, in the crucial "language decoding" technology, Tao’s company, NeuroXess, has far outpaced Neuralink and created a miracle. The Tiaianqiao & Chrissy Chen Institute, co-founded by Chen, is the largest and most important investor behind NeuroXess.

Hu recalls, "When we first sought funding for NeuroXess, we spoke to several venture capitalists who showed great interest, but when they heard that commercialization would take over ten years, they all backed out. Just when we were about to give up, Mr. Chen and I had a one-hour video meeting. He decided on the spot to invest 30 million yuan (about US$4.2 million) and told me not to rush. He was willing to wait for 20 or 30 years, and if we failed, it would be considered a charitable investment in scientific research."

Tao went on to say that during subsequent discussions, Chen's global vision and forward-thinking left a deep impression on him. Chen had a keen interest in language decoding, and he also funded research into the world’s most successful brain-machine interface technology for decoding English. In NeuroXess’s later rounds of financing, Chen not only continued to invest but also introduced top-tier venture capital firms like Sequoia Capital to invest as well.

So why do I say that the significance of what Chen has accomplished cannot be overstated? We can gauge its importance by understanding just how critical "language decoding" in neuroscience is to medicine, technology, and society. It has truly disruptive implications. Not only can it help restore the language abilities of patients with communication disorders, but it can also drive forward the development of neuroscience, artificial intelligence, and brain-machine interface technologies. Additionally, it leads to profound ideas on human thoughts and  freedom.

From a medical perspective, "neural language decoding" allows patients with language disorders (such as those suffering from stroke, ALS, cerebral palsy, or other neurological conditions) to directly read brain activity and convert their thoughts into text or speech, helping them reconnect with the outside world and replacing traditional assistive technologies. It could also help restore language functions lost due to brain injuries (such as traumatic brain injuries or strokes). For research into neurological diseases, analyzing the brain's language-related regions could provide deeper insights into the impacts of Alzheimer's, Parkinson’s disease, and other neurodegenerative diseases on language, ultimately improving diagnosis and treatment methods.

More importantly, this wave of the AI revolution has shown us the power of large language models. Language is a crucial part of human cognition, and only through neural language decoding can we truly understand the mechanisms of how language is generated, encoded, and processed in the brain. This helps scientists explore how language interacts with other cognitive functions, such as memory and emotion.

As AI research deepens, neural decoding technology can also provide real data on how the brain processes language, allowing for improvements in generative models. This makes the models more aligned with the way humans process language, enabling AI to better and more naturally understand and respond to human intentions.

Musk’s Neuralink is also planning to integrate with the humanoid robot Optimus, aiming to control the robot or prosthetic limbs directly through brain signal decoding.

In the future, when humans can communicate through brainwaves for "silent communication," the way we interact with the world may undergo a complete transformation.