Does the invention of Neuralink indicate the brain is not as encrypted as first thought to mind reading?

 

Good question — and the honest answer is more interesting than a simple yes or no. Let me search for what Neuralink has actually demonstrated, because the gap between marketing and demonstrated capability matters a lot here.Now I can answer properly. The short answer is no, not really — and the longer answer actually strengthens the original essay's argument rather than undermining it.

Here is what Neuralink has actually demonstrated. As of early 2026, 21 participants have received implants, and the headline capability — marketed as "Telepathy" — is that paralysed users can move a cursor, play video games, browse the internet, and post on social media just by thinking. A separate speech-restoration system received FDA Breakthrough Device Designation in 2025 and aims to decode attempted speech movements from motor and language areas for people with ALS, stroke, or cerebral palsy.

That sounds like mind-reading. It isn't, in the sense the essay was worried about. What the device actually does is read motor intent from the motor cortex — the same neural signals that would have moved a hand or formed a word, intercepted before they reach the muscles. The user thinks "move cursor right" and a small patch of motor neurons fires in a pattern the system has been trained, for that specific user, to associate with rightward movement. It's a high-bandwidth replacement for the keyboard and mouse, not a window into thought.

Three things make this clear:

The signals being decoded are deliberately produced and motor-coded. The user is consciously trying to move something. The chip cannot tell what someone is daydreaming about, who they secretly love, whether they are lying, or what they think of their boss. Spontaneous cognition leaves no comparable signal at the motor surface.

The decoder is bespoke per person. Every implant has to be calibrated to its host through training sessions. There is no universal "thought codebook" being built — each brain encodes even simple motor intentions slightly differently, and the calibration drifts and has to be retrained. This is exactly what the essay's neurology argument predicts: cognition is spatiotemporal patterns across shifting networks, not fixed cells with fixed meanings.

The electrode count is small relative to the brain. Even Neuralink's high-density arrays sample on the order of a thousand neurons in one patch of cortex, out of ~86 billion. They get a useful signal because motor cortex is unusually well-organised and the task is narrow. Reading abstract thought would require sampling — and understanding the dynamic code of — vastly more of the brain at once, and nothing on the horizon does that.

So Neuralink updates the picture in a specific way: the brain is not as totally encrypted as the essay implies when it says "no MRI… will ever fully read your actual mind." Certain narrow channels — particularly motor and attempted-speech signals — turn out to be decodable with direct invasive access and per-user training. That's a real shift. It means a person who chooses to have electrodes implanted and trains with them can have specific intended outputs read.

But it leaves the essay's core claim intact. Involuntary, ambient, unexpressed thought remains unreadable. There is no scanner that walks past you and pulls out what you're thinking. The "few cubic centimetres" are still yours unless you open the skull and spend months training a decoder on cooperative output. And — crucially for the essay's deeper point — the threat model it actually warned about (manipulation, suggestion, forced inhibition) is not what Neuralink does. Neuralink reads outgoing motor intent in volunteers. The Orwellian fear was something quite different: predicting and shaping the unspoken interior.

Where it does sharpen the worry: once a BCI is implanted, the stream of decoded intent is data, and data has custodians. The risk shifts from "can the mind be read?" to "who owns the cable, and what else is being inferred from the timing and pattern of motor signals?" That's closer to the essay's real concern — the social and political vector — than to a clairvoyance breakthrough.

So: not encryption broken, more like a side-channel opened by consent, for a narrow class of signal, in cooperating users. The Flash Mind Reader joke at the end of the essay still lands.