Neuralink, Elon Musk’s brain-implant company, has set out the two products it intends to bring to market: Telepathy and Blindsight. Both are wireless brain-computer interface chips, and the company has already implanted its devices in several people with serious neurological conditions.
Telepathy
This is the first planned product. It would let a user control a computer or another device by thought alone, at a distance, and communicate with specific people without speaking or typing.
Musk’s claim for it is a speed claim: that within a few years, exchanging information through Telepathy will be faster than human speech or fast typing. That is the part to treat carefully. Thought-to-cursor control has been demonstrated by several research groups; thought-to-language at conversational speed, reliably and outside a lab, has not.
Blindsight
The second product aims to restore sight to people who have lost it for any reason — and, Musk says, to people who were born blind. He has said the engineering work is finished and the chip is now waiting on approval from regulators.
The congenital case is the ambitious one, and it is worth understanding why. In someone who lost vision later in life, the visual cortex has already learned to interpret signals from the eyes; a device that feeds it signals is working with an established system. In someone blind from birth, that interpretive machinery never developed. Restoring useful sight there is not only an engineering problem but a neurological one, and it is the claim that will attract the most scrutiny from regulators.
What "engineering finished" does and does not mean
A completed device is the beginning of the regulatory process rather than the end of it. Implants that sit in the brain are reviewed on a long timeline, and the questions are about what happens over years: how the tissue responds, whether the electrodes stay useful, what happens when the hardware needs replacing, and what the surgical risk is relative to the benefit.
Nor does a product name confirm a shipping date. Neuralink has named both chips and described their function, but the company has not published the clinical evidence that would let anyone outside it judge how well either performs.
Why this matters beyond the US
Devices like these reach markets like Bangladesh late and expensively, and the first wave is usually in a handful of specialist hospitals abroad. But the category is worth following for a different reason. The same interfaces that restore function to people who have lost it are the ones that raise the hardest questions about neural data — what gets recorded, where it is stored, and who may read it. Those rules tend to be written in the countries that approve the devices first, and then inherited everywhere else.
What would count as evidence
For a field that produces striking demonstrations, the useful measures are unglamorous. For Telepathy: how many words per minute, sustained over how long, by how many implanted users, and how much the accuracy degrades between the first month and the twelfth. For Blindsight: not whether a recipient detects light, but whether they can resolve shapes well enough to perform a task they could not perform before.
Those numbers are what regulators will be given and what independent researchers will want published. Until they exist, a product name and a completed engineering phase describe intent rather than capability — which is the distinction worth holding on to when the next announcement arrives.
