Arctop

A person puts on a pair of earbuds. Somewhere, software starts listening.

Not to their music. To them. To the millions of faint electrical signals their brain throws off every second - the ones that quietly answer questions no survey ever could. Are they focused or drifting? Calm or stressed? Learning this, or just nodding along?

That listening software is Arctop. It is a small company - around fifteen people working out of Los Angeles - with an unusually large idea: that the brain has been talking this whole time, and the only thing missing was something good enough to hear it. Arctop doesn't sell the earbuds. It builds the part that understands what the earbuds pick up.

Fig. 1 - Somewhere right now, a brain is mid-sentence. Most products can't hear a word of it.

Everyone wanted into the brain. They were all digging in the wrong place.

For years, the glamorous version of brain-computer interfaces meant going inside: implants, surgery, electrodes threaded into tissue. Impressive. Also, for the other eight billion of us, completely impractical. You do not get a craniotomy to skip a boring training video.

Arctop's founders made a contrarian read. The barrier, they argued, was never really access to the brain - cheap skin sensors in everyday wearables can already pick up its electrical chatter. The barrier was meaning. Raw brain signal is a firehose of noise. Turning that noise into "this person is focused" or "this person is confused" is a software problem, not a surgical one.

It is a tidy bit of irony: an entire industry racing to drill into skulls, while the actual bottleneck sat in the code. Arctop decided to fix the code.

A Harvard neuroscientist and a cybersecurity engineer walk into a brain.

In 2016, neuroscientist Dan Furman teamed up with software engineer Eitan Kay. Furman's path to the problem was not casual: at sixteen he watched an open-brain Parkinson's surgery, then went on to earn a neurobiology degree at Harvard and a PhD at the Technion. There he became the first person to show that scalp-measured brain signals - the supposedly hopeless, noisy kind - could decode imagined movements of individual fingers. People had said that couldn't be done.

Their bet was simple to state and hard to win: build brain-decoding software so accurate it works on the cheap sensors people already wear, and license it to everyone else who wants to build on top. No hardware empire. No surgery. Just the intelligence layer.

Fig. 2 - Two résumés that should not fit on the same business card, and somehow do.

Meet Neuos - the part that does the understanding.

Arctop's flagship platform, Neuos, sits between brain-sensing hardware and the apps that want to use it. Feed it the electrical signals from a headband, earbud, or VR headset, and it returns real-time, high-resolution maps of feelings, reactions and intentions. Focus. Stress. Enjoyment. Memory. Intent. The states that used to require a questionnaire and a guess.

One detail Arctop is loud about: it does not share your raw brain data. Permissions live in a dashboard you control. In a field where "we read your mind" is the pitch, "and we don't pass it around" is a surprisingly radical footnote.

Money, medicine, and a refusal to hand over your data

Belief is cheap in neurotech. Arctop has receipts. In October 2023 it closed a $10 million Series A co-led by Fifth Growth Fund and Supermoon Capital, bringing total funding to roughly $14.17 million. That capital is pointed at one thing: scaling the decoding software so more developers can build on it.

The proof is also in the partners. Arctop works with Stanford Medicine on surgical training, alongside Wellcome Leap's surgical-coaching program, and is available to U.S. government agencies through Carahsoft. Its founding lineage even traces back to work on Stephen Hawking's communication technology - a reminder that this team has been near the hard end of this problem for a long time.

Fig. 3 - Bars are illustrative. The point underneath them is not: reach has always been the prize.

An interface for the part of you that never had one

Strip away the jargon and Arctop's goal is human-scale: to expand what people can do by making the brain an input that everyday technology can actually understand. Training that adapts to whether you're truly learning. Communication tools for people who've lost their voice. Experiences that respond to how you feel, not how you click.

It's a phrase that should sound like a stretch and somehow doesn't. Every other layer of the digital world got an interface decades ago. The person using it never did.

Back to the earbuds

Return to that person slipping in their earbuds. Today, the gesture means one thing: sound goes in. Arctop is betting that soon it means two - sound goes in, and understanding comes out. The device finally knows whether the lesson landed, whether the work is draining them, whether they're present or gone.

That future is not guaranteed. Decoding the brain accurately, on cheap sensors, at scale, while keeping the data in the user's hands, is genuinely hard - which is exactly why most companies went looking for shortcuts inside the skull instead. Arctop took the harder road on purpose.

The brain was always talking. The only open question was who would build something patient enough to listen. For now, a fifteen-person company in Los Angeles is raising its hand.

• Furman's curiosity started at sixteen, watching an open-brain Parkinson's surgery.
• Arctop is hardware-agnostic - it doesn't build the headband, it builds the brain inside the software.
• Its founders bridge two worlds: a Harvard-trained neuroscientist and a Tel Aviv cybersecurity engineer.
• The company insists it never shares your raw brain data - you hold the permissions.