Neuralink looks to the public to solve a seemingly impossible problem
The company is issuing a public challenge to find a new way to compress data
Elon Musk's quest to wirelessly connect human brains with machines has run into a seemingly impossible obstacle, experts say. The company is now asking the public for help finding a solution.
Musk's startup Neuralink, which is in the early stages of testing in human subjects, is pitched as a brain implant that will let people control computers and other devices using their thoughts. Some of Musk's predictions for the technology include letting paralyzed people "walk again and use their arms normally."
Turning brain signals into computer inputs means transmitting a lot of data very quickly. A problem for Neuralink is that the implant generates about 200 times more brain data per second than it can currently wirelessly transmit. Now, the company is seeking a new algorithm that can transmit this data in a smaller package — a process called compression — through a public challenge.
As a barebones web page announcing the Neuralink Compression Challenge posted on Thursday explains, "[greater than] 200x compression is needed." The winning solution must also run in real time, and at low power.
Crucially, it specifies that the compression must be "lossless." A "lossy" compression would be like a low-quality MP3 file, compared to pristine vinyl.
The reward for developing this miraculous leap forward in technology? A job interview, according to Neuralink employee Bliss Chapman. There is no mention of monetary compensation on the web page.
'Seeking a miracle'
To kickstart the challenge, Neuralink released one hour of raw brain recordings from a monkey playing a simple video game. Participants are asked to compress this data. Controlling a video game — dubbed MindPong — was one of the earliest demos of the implant, revealed in 2021.
Observers on social media immediately branded the task "impossible," even speculating that Neuralink staff launched the challenge as a way of convincing the infamously incalcitrant Musk that it couldn't be done.
Neuralink did not respond to a request for comment.
The skepticism is well-founded, said Karl Martin, chief technology officer of data science company Integrate.ai. Martin's PhD thesis at the University of Toronto focused on data compression and security.
Neuralink's brainwave signals are compressible at ratios of around 2 to 1 and up to 7 to 1, he said in an email. But 200 to 1 "is far beyond what we expect to be the fundamental limit of possibility."
Key to Neuralink's pitch is wireless data transfer. Earlier versions of brain implants required bulky wires to protrude from the patient's head to communicate with an external device that read the neural signals and translated them into instructions.
That won't do in the real world, so companies working on wireless brain implants — there are several besides Neuralink — are seeking technology that can transfer large quantities of data very quickly. For Neuralink, that means compression.
Martin explained that, rather than a software engineering problem, the challenge primarily concerns information theory, constrained by the laws of physics. Succeeding would require a completely "atypical" advance in the field that redefines how we understand neural signals and their compression.
"It's essentially seeking a miracle."
The problem, said Roy van Rijn, director of software consultancy OpenValue Rotterdam, is that the files appear very noisy. In other words, they contain many unique data points without common patterns. "If there aren't enough 'patterns' in the data, it is mathematically impossible to compress something further," he said in an email.
Van Rijn wrote a simple algorithm that compressed the Neuralink files at a ratio of 3.37 to 1. He speculated that participants using a similar approach might be able to compress the neural signals further, but that the "general consensus is that [200 to 1] is just outlandish."
Neuralink controversies
Neuralink's technology has progressed slowly, especially since the idea behind the implant isn't new. In 2004, Massachusetts man Matthew Nagle became the first person with paralysis to be given a brain implant by a company called Cyberkinetics. The implant allowed him to control a computer and even play Pong.
That was 20 years ago. Cyberkinetics, along with several other companies, is still pursuing brain implant technology. In 2021, it even succeeded in transmitting wireless instructions from a human brain to a computer.
Neuralink seeks to differentiate itself with an implant that is "cosmetically invisible," and installed by a robot rather than a human surgeon.
Thirty-year-old Noland Arbaugh is the first and so far only person to receive a Neuralink N1 implant, which was installed in January. The implant primarily allows him to move a computer cursor and play video games such as the popular strategy game Civilization 6.
Last week, the Wall Street Journal reported that Neuralink received clearance from the U.S. Food and Drug Administration (FDA) to implant a second person in June. The company plans to implant a total of 10 people this year, according to the report.
Neuralink has been dogged by controversies. Employees told Reuters that pressure from Musk to speed up testing led to human error and botched experiments, causing more animals to be killed than necessary. Nearly 300 of those were monkeys, sheep and pigs. Reportedly, the internal panel overseeing Neuralink's experiments was stuffed with company employees who stood to benefit from the implant reaching the market.
The failed experiments have led to federal probes. Though a 2023 probe from the U.S. Department of Agriculture found no animal welfare violations, the FDA identified several quality control lapses later that year. Those included not calibrating equipment such as a "vital signs monitor" during experiments.
Neuralink's sole human trial to date hasn't seen smooth sailing, either. On May 8, Neuralink revealed in a blog post that weeks after Arbaugh's surgery, "a number of threads retracted from the brain," making the implant less effective. Arbaugh told Wired that he hadn't been informed that this was possible, and while the company reassured him that the threads had stabilized, and it's unlikely to happen again, he "[has] fears about that."
Despite the setback, Arbaugh is still excited about his Neuralink implant. "It's helped me reconnect with the world, and to become more independent," he previously told CBC News.
Far beyond reasonable
The compression challenge is worded vaguely enough that it's possible a winning entry might not exactly satisfy the stated requirements. Submissions will be "scored" based on how much the algorithms can compress the neural signals, with "bonus points" for speed and power efficiency.
It's worth noting that for other types of signals — such as videos on social media — "lossy" compression is often acceptable, making such significant compression ratios potentially possible. "If you are allowed to throw data away, large compression ratios are possible," Van Rijn said. A ratio of 200 to 1 would be "easy."
But Neuralink requires completely lossless compression.
"It's hard for me to speculate, but I think when any company encounters a fundamentally hard technical challenge, seeking input from a broader community of experts is a reasonable approach," said Martin.
"What was unique in this case is that the framing of the challenge and the goal is far beyond what the expert community would consider reasonable."
ABOUT THE AUTHOR
