Kenneth Shock had been watching. When Neuralink opened its patient registry for people living with ALS, he enrolled the same day. In January 2026, he became the second person to receive the company's N1 brain implant, designed to decode the neural signals that once would have traveled from his motor cortex to his mouth and emerged as speech. Eight weeks later, in a video posted to X, he looked at the camera and said: "I'm talking to you with my mind."
The video is real. The milestone is real. And the gap between them is also real.
Neuralink's VOICE trial — an early feasibility study registered as NCT07224256 on ClinicalTrials.gov, conducted at the University of Texas Southwestern Medical Center — is testing whether the N1 implant can restore communication to people who have lost the ability to speak. The implant uses 1,024 electrodes distributed across 64 threads, recording activity from the speech motor cortex. That signal is then mapped to phonemes, assembled into words, and run through a text-to-speech converter. The output is synthetic but chosen to sound like the patient's own voice, based on recordings made before their decline.
The system does not read free thought. It maps phoneme-level neural signals to words through a three-phase training sequence: first spoken sentences, then silent mouth movements mimicking those sentences, and finally imagined speech. Kenneth reached stage three. That is genuinely new. The question is what "reaching stage three" actually means in practice.
Neuralink frames this as progress, and it is — by a narrow and specific measure. Multi-second latency separates the decoded word from the patient's intended word. The system requires supervised training sessions to calibrate to each individual's neural signature. And the clinical framing deserves scrutiny: VOICE is an early feasibility study, not a pivotal trial. FDA Breakthrough Device Designation, received in May 2025, is a regulatory shortcut for serious unmet needs, not a seal of efficacy. The 21 Neuralnauts enrolled across Neuralink's trials worldwide are generating safety data that will take years to contextualize.
There is also a competitive picture that Neuralink's public communications tend to compress. In August 2024, a study published in the New England Journal of Medicine described a fast and accurate brain-computer interface capable of restoring communication in an ALS patient. Separately, UC Berkeley and UC San Francisco presented a neuroprosthesis in 2025 capable of restoring more natural voice in real time — a benchmark Neuralink has not yet reached. Paradromics, a neurotechnology company working on similar speech restoration implants, is another competitor in the space. The gap between "implanted and decoding" and "implanted and decoding in real time with natural voice" is not a small one.
Kenneth's story is not a tech-bro fantasy of telepathy. It is a man with ALS, living in a body that has taken most of what he used to be able to do, who enrolled in a trial because the alternative was silence. That is urgent and real and deserves to be told without the mythology layered on top of it. The mythology makes it harder to see what the technology actually does, who it actually serves, and how far it still has to go.
The trial will generate more data. The system will improve. The question for the field — and for the patients watching from the other side of a screen — is whether the timeline for meaningful, scalable speech restoration runs faster than the disease it is designed to outrun.
