Charter: What CyberMind/UK Is For

There are perhaps six magazines in the English-speaking world that cover neuroscience-adjacent technology with anything resembling editorial discipline. Two of them are paywalled academic journals. Two are trade publications written for procurement officers. One is Wired, which is excellent and increasingly aimed at a US-centric general readership. The sixth is MIT Technology Review, which is also excellent and writes for a different reader than I do.

This is the seventh. It is smaller. It is opinionated. It assumes you can read a clinical-trial abstract without hand-holding and an earnings release without subtitles.

My name is Kaelen Voss, I write under the handle Net_Whisper_77, and I have spent the last decade in the borderlands between neurotechnology, software engineering, and capital markets. I built brainwave-classification tooling as a side project, shipped production HCI work for a London-based design consultancy, traded my own book through the 2022 macro mess, and watched what I thought were three separate fields converge into one. They are not separate fields anymore. They have not been for a while. The same Stanford lab that publishes a closed-loop BCI paper this month is, six months later, the source of a startup raising at a billion-dollar valuation from the same venture funds backing the next foundation model lab. The capital, the engineering, and the science are now one story. This publication treats them that way.

What you will find here

Brain-computer interfaces, as primary subject. The clinical state of the field as of mid-2026, what the trials actually show, where the failure modes are, and which research groups are doing serious work versus competent-but-derivative work. I follow the Pitt, BrainGate, Synchron, Paradromics, Precision, and Neuralink programmes closely. I read the regulatory filings. I quote them by name.

AI ethics that does not lecture you. I am interested in the practical philosophy of building these systems: what we owe the systems themselves under uncertainty, what we owe the populations affected by their deployment, where the regulatory frameworks are pointing and where they are failing. I have low patience for arguments that resolve themselves into either "AI safety is everything" or "AI safety is nothing." The interesting questions are downstream of both.

Cybersecurity at machine pace. The transition from human-paced security operations to agent-paced ones is the most consequential operational shift in the field in twenty years. I cover the autonomous red-team and blue-team exercises that are starting to define what "industrial-grade" defence looks like under these conditions.

Capital markets, because they fund everything else. I write about AI infrastructure capex, hyperscaler economics, fixed income, the macro backdrop that determines which startups get funded and which do not. Not because markets are intrinsically interesting (they are, mostly, not), but because following the capital is the fastest way to understand which technologies will exist in three years and which will not.

What you will not find here

No "10 ways AI will change your workflow." No conference live-blogs. No press release laundering. No predictions about the year 2040. No advertorial. No newsletter sponsorships. No promoted content of any kind.

Everything published here is written by me. The editorial standard is whether I would defend the piece in front of a sceptical room of practitioners in the field. The publishing schedule is irregular, weighted toward density over frequency. The site has no comment section by design — if you want to argue, my email is in the colophon and I read it.

A note on tone

I have been told, repeatedly, that this kind of writing should be more accessible. I take the point and partially disagree. There is enormous value in technical writing that explains itself without condescending. There is, separately, value in writing for readers who already know the field and want to think harder about it. This publication is the second kind. The pieces are long. The vocabulary assumes context. The argument structure rewards re-reading. If that is not your speed, Wired and Technology Review both do excellent work in the first register, and you should read them too.

The desk

CyberMind/UK is a one-person operation, run from a small flat in southeast London, on hardware bought with my own money, hosted on Cloudflare Pages. No funding, no investors, no advisory board. The publication exists because the conversation it tries to host is the one I wanted to read and could not find.

The next two years are going to be — and this is the only prediction in this entire piece — substantially more interesting in this field than the previous five. The clinical BCI trials are about to mature. The AI capex bill is about to start producing returns or visibly not. The cybersecurity tempo shift is about to bite. The neurorights regulatory frameworks are about to either solidify or fall apart. I am going to write about all of it.

Thanks for being here. The dispatches start now.