In Part 2, we left off with Benjamin Rapoport and other co-founders walking away from Neuralink, citing vague concerns about safety. But what exactly did they see that made them leave? What were they so worried about that they’d abandon a company they’d helped build from the ground up?
The Scientists Who Left
Benjamin Rapoport left Neuralink in 2018, two years after co-founding the company with Musk and six other scientists. He went on to build Precision Neuroscience, launching it in 2021 with a very specific mission: to prove that brain-computer interfaces could be built safely, without the invasiveness that Neuralink was pursuing.
In a 2024 interview with the Wall Street Journal, Rapoport was careful but clear about why he left. “I’ve pretty much devoted my entire professional life to bringing neural interfaces from the world of science to the world of medicine,” he said. “But I felt that in order to move to the world of medicine and technology, safety is paramount.” The contrast was evident in his next statement. “For a medical device, safety often implies minimal invasiveness.”
The difference is simple: how they interact with brain tissue. Neuralink uses penetrating microelectrodes, thin threads that pierce into the brain itself. Precision uses surface microelectrodes that sit on top of the brain without penetrating it. Rapoport was frank about the risks of Neuralink’s approach: it causes “some amount of brain damage” during insertion. His company was founded on the belief that you could extract high-quality data from the brain without causing that damage.
Rapoport was one of many who left. By August 2020, only three of the eight founding scientists remained at Neuralink, according to Stat News. The exodus included Max Hodak, the company’s president and co-founder, who departed in 2021. Paul Merolla, Philip Sabes, Tim Gardner, Tim Hanson, and Vanessa Tolosa had all left by then as well. Stat reported that Neuralink had seen “years of internal conflict in which rushed timelines have clashed with the slow and incremental pace of science.”
The scientists who stayed behind or joined later found themselves under what multiple former employees described as relentless pressure. Tim Hanson, a scientist at the Janelia Research Campus who was part of Neuralink’s founding team, told Fortune there was a “mismatch” between the pace at which basic science can develop and what Musk and Neuralink wanted done. Another employee was more direct:
“Everyone in that whole empire is just driven by fear.”
So… what were they so afraid of?
The Animal Trials
The answer became clear in February 2022, when an animal rights group called the Physicians Committee for Responsible Medicine filed a complaint with the U.S. Department of Agriculture. The complaint was based on over 700 pages of veterinary records and necropsy reports obtained through a public records request from the University of California, Davis, where Neuralink had conducted its animal testing from 2017 to 2020. If you want the full details of what those records revealed, Business Insider and CNN did excellent reporting on this, and I’d recommend reading their coverage. What follows is a summary of what they found.
The documents revealed that out of 23 monkeys implanted with Neuralink’s brain chips at UC Davis, at least 15 had died. The details were disturbing. One monkey developed a bloody skin infection at the implant site and had to be euthanized. Another was found with missing fingers and toes, “possibly from self-mutilation or some other unspecified trauma,” according to the records, and was put down. A third monkey began vomiting uncontrollably shortly after surgery and days later “appeared to collapse from exhaustion/fatigue.” An autopsy revealed the animal had suffered a brain hemorrhage.
The PCRM’s complaint accused UC Davis and Neuralink of nine violations of the Animal Welfare Act.
“Many, if not all, of the monkeys experienced extreme suffering as a result of inadequate animal care and the highly invasive experimental head implants during the experiments,”
the complaint stated. Jeremy Beckham, the research advocacy coordinator for the Physicians Committee, said the documents
“reveal that monkeys had their brains mutilated in shoddy experiments and were left to suffer and die.”
Neuralink acknowledged eight deaths, but the PCRM’s review of records documented at least 15. The company attributed the deaths to ‘one surgical complication involving the use of the FDA-approved product (BioGlue), one device failure, and four suspected device-associated infections, a risk inherent with any percutaneous medical device.’ UC Davis stopped working with Neuralink in 2020, and the company opened its own in-house facility to continue testing.
The mortality rate was what stood out. Other companies working on brain-computer interfaces, including Synchron and Precision Neuroscience, had demonstrated better safety records in their animal testing. Fifteen deaths out of 23 test subjects suggested something was being pushed too fast, that proper protocols might not be followed, that the device itself might be fundamentally more dangerous than alternatives. The scientists who had left seemed to have seen this coming.
The FDA’s No & Yes
Those questions reached the FDA in early 2022, when Neuralink applied for permission to begin human trials. The agency said no. Reuters did critical reporting on this rejection, speaking with seven current and former Neuralink employees who described what happened. The FDA handed back a list of dozens of safety concerns that needed to be addressed. The major issues included the device’s lithium battery, which could potentially overheat inside the skull. There were concerns about the tiny wires migrating to other areas of the brain, causing inflammation or other damage. The FDA also wanted to know how the device could be safely removed without damaging brain tissue, a question Neuralink had acknowledged in its own presentations as “a challenging problem” it hadn’t solved yet.
The rejection was significant. Kip Ludwig, the former program director for neural engineering at the U.S. National Institutes of Health, told reporters that Musk’s “well-known impatience” posed a challenge for the FDA approval process. “Neuralink doesn’t appear to have the mindset and experience that’s needed to get this to market anytime soon,” Ludwig said.
Fourteen months passed. Neuralink, according to employees, continued working on addressing the FDA’s concerns. They ran additional tests on pigs, trying to resolve some of the safety issues the agency had raised. And in May 2023, the FDA granted approval for human trials.
Here’s the thing, though. We have no idea what actually changed. And I mean that literally. We don’t know. Neuralink didn’t publicly release updated animal testing data showing improved safety outcomes. The company didn’t announce any major design changes to address the lithium battery concerns or the wire migration problem. There was no public explanation from the FDA about which specific concerns had been resolved or how they’d been addressed. The entire approval process happened behind closed doors, with no transparency about what data convinced the FDA that the problems had been adequately handled.
This matters because we’re talking about a device that gets surgically implanted into people’s brains. The FDA identified dozens of serious safety issues in early 2022. Fourteen months later, those issues were apparently resolved well enough to move forward with human testing.
The First Human Trial
The approval was announced, and Neuralink immediately began recruiting for its PRIME study (Precise Robotically Implanted Brain-Computer Interface). The first participant would be someone with quadriplegia, someone who had limited or no ability to use their hands due to a spinal cord injury. The goal was to help them control computers and smartphones using only their thoughts.
On January 29, 2024, Elon Musk announced on X that Neuralink had successfully implanted its device in the first human patient.
I’m sorry, let me say that again. The first human brain implant for this company was announced on X. The social media platform. The one Musk owns. Someone underwent experimental brain surgery to have electrodes threaded into their brain tissue, and we found out about it sandwiched between cryptocurrency scams and AI-generated slop. Not through a medical journal. Not through any official channel that involves peer review or regulatory oversight. Through a tweet.
The patient was recovering well, Musk said. Initial results showed promising neuron spike detection. A few weeks later, on February 20, he posted again: the patient could control a computer mouse using their thoughts. This cavalier, almost performative disregard for how medical trials are supposed to work would become the pattern for how we’d learn anything about Neuralink’s human experiments going forward.
The big reveal came on March 20, 2024, when Neuralink introduced the world to Noland Arbaugh through a livestream on X. Arbaugh was 29 years old, paralyzed from a diving accident eight years earlier. In the video, he demonstrated the implant’s capabilities by playing chess online and moving the cursor around the screen using only his mind. He was enthusiastic, grateful, clearly excited about what the technology might mean for his independence. The demonstration was impressive. Media coverage was widespread. It looked like a triumph.
What Neuralink didn’t mention during that March demonstration was that something had already gone wrong. Bloomberg broke this story in May 2024, reporting that in the weeks following Arbaugh’s January surgery, a number of threads had retracted from his brain tissue. The threads are the tiny electrode-laden wires that Neuralink’s surgical robot inserts into the brain to detect neural signals. According to Bloomberg’s reporting, approximately 85% of the threads had pulled away from the tissue they were supposed to be monitoring.
This was a major hardware complication. The threads retracting meant the implant was receiving significantly fewer signals than intended, which would make it far less effective. Neuralink’s solution, according to the company, was to modify the algorithm to make the device more sensitive to the remaining threads. In other words, they fixed a hardware problem with software. Let me be clear about what that means: 85% of the physical threads that were supposed to be in someone’s brain tissue had retracted, so they just turned up the sensitivity on the remaining ones.
This is a very ‘Musk approach’ to problem-solving. It’s the same logic that’s defined Tesla’s self-driving development for years. Why invest in LiDAR systems, the hardware that most autonomous vehicle companies consider essential for detecting obstacles, when you could just make the AI better? Better hardware is expensive. More AI, though? That makes his companies and his friends’ companies a lot of money. Why address a fundamental engineering limitation when you could compensate with code and call it innovation? It’s worked out great for Tesla, if you ignore all the crashes.
And now that same philosophy is being applied to devices implanted in people’s brains.
What made this more troubling was the lack of disclosure. Noland Arbaugh had been posting on X since March 2024, sharing updates about his experience with the implant. But sometime after the Bloomberg report came out, many of those early posts disappeared. Social media posts were deleted. The public record of what had happened during those early weeks and months after his surgery was quietly edited. Neuralink has said that the modified surgical approach used for subsequent patients, placing the implant closer to the brain surface and reducing brain motion during surgery, has prevented thread retraction in later cases. They’ve reported no similar issues with the second patient or those who came after. But we only know that because they’ve chosen to share it. There’s no independent verification, no FDA-mandated adverse event reporting that’s made public, no outside researchers monitoring these trials.
The Dismantling
For a moment in December 2024, it looked like accountability might actually be coming. The Securities and Exchange Commission reopened an investigation into Neuralink, focusing on safety claims related to its brain chip technology. The FDA sent a letter to Congress citing Neuralink’s California animal testing facility for “objectionable conditions or practices” based on inspections conducted in June 2023. The USDA inspector general, Phyllis Fong, was actively investigating the company for alleged animal welfare violations. After years of complaints from animal rights groups, after the PCRM’s documentation of 15 dead monkeys, after all the reports of rushed timelines and shoddy experiments, multiple federal agencies were finally scrutinizing Neuralink’s practices. Maybe, just maybe, the system was working the way it was supposed to.
And then came January 20, 2025. Donald Trump was inaugurated for his second term as president. Within his first week in office, Trump appointed Elon Musk to lead the Department of Government Efficiency, or DOGE, an initiative aimed at cutting what the administration characterized as bureaucratic waste across federal agencies. Musk, whose companies are regulated by multiple federal agencies including the FDA, was now in a position to influence the very regulators tasked with overseeing his work.
Seven days after the inauguration, on January 27, Neuralink announced that the University of Miami’s Miller School of Medicine had been selected as the second U.S. site for the PRIME study. The Miami Project to Cure Paralysis would begin implanting patients in Florida.
And then, systematically, those investigations that had just been opened began to disappear. During Trump’s first week in office, he fired 17 inspectors general across federal agencies, including Phyllis Fong, the USDA inspector general who had been leading the investigation into Neuralink’s animal testing practices. She had to be removed from her office by security after refusing to comply with her termination, which she believed was illegal.
The FDA purge followed. On February 15, 2025, the Trump administration started firing probationary employees across federal health agencies, with the FDA hit particularly hard. According to Reuters, the cuts included about 20 people in the FDA’s Office of Neurological and Physical Medicine Devices. That’s the division that oversees clinical trial applications for brain-computer interface devices. Several of the fired employees had specifically worked on Neuralink’s applications.
The firings continued through February, March, and into April. Some employees were fired, then hastily rehired when it became clear their roles were critical, then fired again. The chaos was widespread. On March 28, 2025, the Department of Health and Human Services announced it would eliminate 10,000 jobs across its agencies. The FDA alone would lose 3,500 positions. The NIH would cut 1,200 jobs.
The conflict of interest was impossible to miss. The company under FDA review was being run by the same person who was now firing the FDA employees reviewing that company, including some of the specific staff who had worked on Neuralink’s application. The inspector general investigating the company’s animal welfare violations was fired. The SEC investigation that had just been reopened in December went quiet. And the FDA citation for “objectionable conditions” at the company’s animal lab? No follow-up, no enforcement, no consequences.
And by April 2025, the first patient had been implanted at the Miami site. His name was RJ, and he was a paralyzed U.S. military veteran (more on him in part 4). The procedure was reported as successful. By August and September 2025, Neuralink had expanded to Canada, implanting two patients in Toronto. In October 2025, the company announced its first patient in the United Kingdom, a man named Paul. As of November 2025, the company has claimed to have implanted somewhere between 12 and 13 people, though the exact number remains unclear. The expansion was rapid. The oversight was gone.
Where We Are Now
Here’s what we’re left with. The scientists who co-founded Neuralink left over safety concerns. The company’s animal testing killed at least 15 of 23 monkeys. The first human patient’s device failed, with 85% of the threads retracting from his brain, and the company’s solution was software, not fixing the actual hardware problem. And then, just as federal investigations were finally opening in December 2024, the new administration took office and systematically shut them down. The inspector general was fired, the FDA employees reviewing Neuralink were fired, and the oversight just… disappeared.
Which brings us to the patients. As of November 2025, Neuralink has claimed to have implanted somewhere between 12 and 13 people. Some are very public. Others were announced once and vanished. And when I started trying to verify their stories, trying to find evidence of their lives before Neuralink, what I found genuinely alarmed me.
That’s Part 4.
Sources Cited
News Reports and Investigations
Physicians Committee - Original Records of Each Monkey Used by Neuralink at UC Davis
Business Insider - Animal-rights group says monkeys used in experiments for Elon Musk’s Neuralink were subjected to ‘extreme suffering’
Reuters - Animal testing records investigation
Quartz - Benjamin Rapoport interview (2024)
CNBC - Neuralink co-founder Max Hodak leaves Elon Musk’s brain implant company
Fortune - Neuralink former employees say Elon Musk applies relentless pressure and instills a culture of blame
CNN - Thread retraction in first patient (May 2024)
Bloomberg - Neuralink Says Implant Had Issues After First Human Surgery
The Guardian - USDA inspector general escorted out of office after defying Trump order
Fortune - DOGE FDA employee firings (February 2025)
Elon Musk X Posts
First Human Implant Announcement - January 29, 2024
“The first human received an implant from @Neuralink yesterday and is recovering well. Initial results show promising neuron spike detection.”
Telepathy Product Announcement - January 29, 2024
“The first @Neuralink product is called Telepathy”
Company Websites
Neuralink Official Website
Precision Neuroscience - Alternative BCI approach
Synchron - Competing BCI company with alternative safety record
