Note: This is a research note supplementing the book Unscarcity, now available for purchase. These notes expand on concepts from the main text. Start here or get the book.

Brain-Computer Interfaces: The Upgrade You Didn’t Know You Were Getting

Key Insight: In 2024, a quadriplegic man beat the world record for BCI cursor control on his first day using Neuralink. In 2025, an ALS patient “spoke” through a brain implant at 97% accuracy. The gap between science fiction and your neurosurgeon’s office is shrinking faster than most people realize—but the hype still outruns the hardware.

Your Brain Has a USB Port Problem

Right now, you’re reading these words through the most sophisticated bottleneck in the known universe: your eyes.

Light photons hit your retina. Chemical signals cascade through your visual cortex. Meaning assembles somewhere in the back of your head. Then—if you want to respond—you have to squeeze all that rich, textured thought back through your motor cortex, down your spinal cord, into your fingers, onto a keyboard, and across fiber optic cables. Your brain processes roughly a quintillion operations per second. Your typing speed? Maybe 80 words per minute if you’re quick.

This is like having a supercomputer communicate through morse code.

Brain-Computer Interfaces (BCIs) are the attempt to cut out the middlemen. Instead of routing everything through your body’s creaky peripheral hardware—eyes, ears, fingers, tongue—why not tap directly into the source?

The idea has been around since the 1970s, when researchers first proved that monkey neurons could control external devices. For decades, progress was glacial. Then, quite suddenly, it wasn’t.

What Actually Happened in 2024-2025

Let’s dispense with the speculation and look at what’s actually working.

The Neuralink Saga: From Lab Rats to Noland Arbaugh

On January 28, 2024, a quadriplegic named Noland Arbaugh became the first person to receive Neuralink’s N1 brain chip—a device featuring over 1,000 electrodes threaded through his motor cortex by a robot surgeon. The procedure took under two hours. He went home the next day.

The first time Arbaugh used the device, he beat the 2017 world record for speed and precision in BCI cursor control. Not after months of training. His first day.

But here’s the part the press releases didn’t emphasize: within a month, up to 85% of the implanted electrode threads had retracted, degrading his control. Instead of surgery to fix it, Neuralink pushed software updates. The device adapted. Arbaugh regained function.

Eighteen months later, Arbaugh uses the implant about 10 hours daily—playing Mario Kart, studying Japanese, controlling his environment. “My whole life has changed,” he told reporters.

As of late 2025, Neuralink has expanded to 12 trial participants across the US, Canada, and the UK, with over 15,000 hours of cumulative usage. The second patient, known as “Alex,” uses the device to create 3D CAD designs and play first-person shooters at a competitive level. His surgery used updated techniques that eliminated the thread retraction problem entirely.

What it means: Neuralink has proven that consumer-grade brain chips can work in humans, that software can compensate for hardware problems, and that the technology improves rapidly with iteration. They’ve also proven that things can go wrong—the thread retraction surprised everyone.

Synchron: The Stent That Reads Your Mind

If Neuralink is the Tesla of BCIs—flashy, ambitious, occasionally on fire—Synchron is the Toyota: boring-sounding but quietly effective.

Their Stentrode device doesn’t require brain surgery. Instead, it’s inserted through a blood vessel in the neck and guided to the motor cortex like a cardiac stent. The procedure takes about 20 minutes.

In September 2024, Synchron announced positive results from their COMMAND study: six patients with severe paralysis, all achieving successful brain signal capture, zero serious adverse events after 12 months. One ALS patient demonstrated controlling his smart home through Amazon Alexa and even navigating the Apple Vision Pro—using nothing but thoughts.

The Stentrode offers fewer electrodes than Neuralink (16 vs. 1,024), which means lower resolution brain reading. But it also means no brain surgery, faster recovery, and lower risk. Synchron is working with the FDA toward a pivotal trial that could lead to the first commercially approved brain implant.

What it means: You don’t have to drill through skulls to read brains. Lower-resolution, higher-safety approaches might reach mainstream medicine first.

BrainGate and Speech: Words from Silence

The most emotionally resonant breakthrough came from the BrainGate consortium at UC Davis. In August 2024, they published results of a speech BCI that achieved 97% accuracy—translating the brain signals of an ALS patient named Casey Harrell directly into spoken words.

The system involved four microelectrode arrays implanted in Harrell’s speech motor cortex. On the first training session, it took 30 minutes to achieve 99.6% accuracy with a 50-word vocabulary. With 1.4 hours of additional training, accuracy held at 90% across a 125,000-word vocabulary.

Here’s the part that hits hardest: the decoded words weren’t spoken in a robotic voice. The system reconstructed Harrell’s own pre-ALS voice from audio samples. When his BCI speaks, it sounds like him—before the disease took his voice.

Harrell has now used the system for over 248 hours of conversation, in person and over video chat.

What it means: BCIs aren’t just for cursor control anymore. The same technology can restore speech—and possibly, eventually, enable silent communication at speeds far beyond typing.

Precision Neuroscience and Blindsight: The Newcomers

Two other companies deserve attention:

Precision Neuroscience received FDA 510(k) clearance in April 2025 for their Layer 7 Cortical Interface—a 1,024-electrode thin film at one-fifth the thickness of a human hair. It’s authorized for implantation durations up to 30 days, marking a major regulatory milestone for the field.

Neuralink’s Blindsight device received FDA Breakthrough Device Designation in September 2024. The concept: implant electrodes directly into the visual cortex to provide sight to people who are completely blind—even those blind from birth. Musk announced plans for human trials within six to twelve months after extended monkey testing. Initial vision would be “low resolution, like early video games”—but with potential for improvement, including wavelengths beyond human perception.

What it means: BCIs are diversifying from motor control into sensory restoration. The goal isn’t just to read brains, but to write to them.

The Non-Invasive Path: No Surgery Required

Not everyone wants to open their skull. The non-invasive BCI market—using EEG, fNIRS, and other external sensing—is growing rapidly and currently commands 86% of the BCI market by volume.

What’s Actually Working

Cognixion has developed the Axon-R headset, combining augmented reality with brain sensing for people with severe motor impairments. An MIT 2024 study demonstrated their AR-BCI hybrid could isolate speech-related cortical signals with 93% accuracy in noisy environments, enabling non-verbal users to communicate at 50 words per minute.

Kernel’s Flow 2 headset uses phased-array ultrawideband sensors to achieve 1mm spatial EEG resolution—far better than traditional EEG caps. The company targets neuromarketing applications initially, but the technology platform is general-purpose.

Gabe Newell’s Starfish Neuroscience is developing a battery-free, cordless brain chip operating on just 1.1 milliwatts—5.5 times more efficient than Neuralink. They’re targeting late 2025 for first chips, initially focused on movement disorders like Parkinson’s and epilepsy, but with explicit gaming applications in the long-term roadmap.

The Gaming Angle

Speaking of gaming: Valve has been quietly exploring BCIs for years. Newell believes brain interfaces will eventually create experiences that make “the real world seem flat, colorless, blurry” by comparison. Former Valve psychologist Mike Ambinder emphasized potential for “10-30ms reaction time improvement” in competitive gaming—which sounds small until you realize that’s the difference between winning and losing at the professional level.

The consumer gaming applications include real-time emotion detection (for adaptive difficulty), VR motion sickness suppression (by directly overriding the vestibular confusion that causes nausea), and eventually, full sensory immersion.

What it means: You don’t need surgery to benefit from BCIs. But you do get roughly 10-100x less signal with non-invasive approaches. The trade-off between capability and safety will persist for decades.

Who Actually Benefits (Today)

BCIs remain primarily medical technology. The current beneficiaries:

Paralysis

People with spinal cord injuries, ALS, locked-in syndrome, and brain stem strokes. Noland Arbaugh. Casey Harrell. The estimated addressable population is roughly 500,000 in the US alone, with 5+ million worldwide.

Current capabilities: cursor control, device control, basic communication, limited speech. Coming soon: complex movement control, high-bandwidth communication, perhaps eventually full body control through robotic surrogates.

Blindness

People with damaged eyes or optic nerves but intact visual cortex. Neuralink’s Blindsight targets this population. The technology is earlier-stage—no human trials yet—but FDA breakthrough designation signals regulatory priority.

Initial resolution will be low—phosphenes (spots of light) arranged to convey shapes and motion. Improvement over time is expected, potentially including wavelengths humans can’t normally see.

Movement Disorders

Parkinson’s disease, essential tremor, and dystonia already benefit from Deep Brain Stimulation (DBS)—a simpler precursor to BCIs. Over 200,000 people have received DBS implants worldwide. Next-generation BCIs like Starfish aim to be less invasive, more precise, and adaptive in real-time.

Treatment-Resistant Depression

Early research suggests BCIs might help depression by monitoring mood-related neural activity and delivering precisely timed stimulation. This is experimental, but several clinical trials are underway.

What’s Actually Hard: The Hype vs. Reality Gap

Here’s where we separate the TED talks from the engineering.

The Bandwidth Problem

Your brain has roughly 86 billion neurons with 100 trillion synaptic connections, processing at speeds that dwarf any computer. Current BCIs sample a few thousand neurons at best.

As the Cognitive Field article notes: current BCIs achieve hundreds of bits per second. Meaningful cognitive sharing would require terabits per second. That’s a million-fold improvement. We’re using a drinking straw to sample an ocean.

The Longevity Problem

Brain tissue doesn’t like foreign objects. Immune responses, scar tissue formation, and electrode degradation all limit how long implants remain effective. The longest-running BCI implants have functioned for 5-10 years, but with degrading performance over time.

Neuralink’s thread retraction in Arbaugh shows the problem isn’t solved. Software can compensate for some degradation, but hardware failure eventually wins.

The Decoding Problem

Reading brain signals is not the same as understanding them. Neurons don’t encode information in neat, decodable patterns. Different people’s brains encode the same thoughts differently. The same brain encodes the same thought differently on different days.

Current BCIs require extensive training for each user, and they work best for relatively simple outputs (cursor movement, discrete word selection). Decoding arbitrary thoughts—the science fiction version—remains extremely hard.

The Write Problem

Sending information into the brain is even harder than reading it out. Blindsight and cochlear implants demonstrate it’s possible, but resolution is low and we don’t fully understand the encoding schemes the brain uses.

Writing to the brain in a way that integrates naturally with existing cognition—rather than feeling like external signals imposed on your thoughts—may require advances we can’t currently predict.

The Dark Side: Ethics, Hacking, and the Neural Divide

No discussion of BCIs is complete without addressing what could go wrong.

Privacy and Mental Autonomy

When a device can read your brain, who controls the data? Current regulations are sparse. As of 2024, only two US states (Colorado and Minnesota) had specific neural data protection laws. Minnesota’s law, signed in May 2024, includes civil and criminal penalties for violating neural data rights.

The scary scenarios are real: BCIs that detect fatigue or attention could be used for workplace surveillance. Advertising companies could optimize pitches based on your emotional responses. Authoritarian governments could monitor citizens’ thoughts in real-time.

Hacking and Security

BCIs are computers connected to the internet, which means they’re hackable. Most current devices lack encryption due to power constraints—the computing overhead of encryption would drain batteries too fast. Yale researchers recommend that regulators mandate physical switches to disable wireless connections and require encryption only during active data transfer.

The worst-case scenarios sound like science fiction: hackers intercepting brain signals, manipulating stimulation patterns, or holding neural implants for ransom. But cybersecurity experts confirm that “neural hacking” is a real and pressing concern, not a distant fantasy.

The Neural Divide

If cognitive enhancement becomes possible through BCIs, who gets access? Current implants cost hundreds of thousands of dollars including surgery. As the technology develops, unequal access could create what researchers call “neurodivides”—where only the wealthy benefit from cognitive enhancement.

This isn’t theoretical. One analysis found that enhanced workforces deliver 2.4x better performance, which means companies that can afford enhanced workers will dramatically outcompete those that can’t. Without intervention, BCIs could accelerate inequality rather than reduce it.

The Unscarcity framework addresses this directly: the Foundation guarantees access to basic cognitive enhancement regardless of wealth, ensuring that enhancement doesn’t become a prerequisite for economic participation.

The Path to the Cognitive Field

Here’s where science meets speculation—but grounded speculation, based on current trajectories.

Near-Term (2025-2030)

FDA approval for at least one permanently implanted BCI for paralysis
Consumer-grade non-invasive BCIs for gaming and productivity (limited capabilities)
Speech BCIs approaching 99%+ accuracy in clinical settings
Initial Blindsight trials in humans
First regulatory frameworks for neural data protection

Medium-Term (2030-2040)

BCI bandwidth improving 100-1000x through technological advancement
Bidirectional BCIs that both read and write becoming common
Memory augmentation: perfect recording and playback of experiences
Initial brain-to-brain communication experiments (extremely limited bandwidth)
Widespread enhancement BCIs for memory, focus, and learning

Long-Term (2040-2060)

High-bandwidth brain-to-brain communication for specialized applications
Experience archives becoming possible: capturing and sharing first-person perspectives
Deep Merge: temporary cognitive pooling between consenting minds
The infrastructure for Voluntary Symbiosis emerging
Choice between biological, augmented, and fully digital existence becoming real

This timeline is speculative but not arbitrary. Each step follows from current research trajectories, with appropriate uncertainty about acceleration or delay.

Connection to the Unscarcity Vision

BCIs are the hardware layer beneath the Cognitive Field—the opt-in telepathy network central to the Unscarcity framework.

But here’s the crucial design principle: the Cognitive Field requires BCIs, but BCIs don’t require the Cognitive Field.

You can use a neural implant to control a wheelchair without ever connecting to another mind. You can have enhanced memory without sharing your memories. You can type at thought-speed without anyone reading your thoughts.

The Glass Wall architecture ensures that connection is always opt-in, always revocable, always under individual control. You cannot be hacked any more than a room can be entered when there are no doors. The wall is physics, not policy.

This matters because BCIs will exist regardless of what framework governs them. The question is whether we build systems that respect autonomy from the ground up—or whether we retrofit consent onto technology designed for surveillance.

The Voluntary Symbiosis Spectrum

The Unscarcity framework anticipates a heritage-synthesis spectrum where some people remain fully biological, some partially augment, some fully upload, and all choices are protected.

BCIs make this spectrum tangible:

Heritage path: No implants, no enhancement, analog interfaces preserved
Augmentation path: BCIs for memory, focus, communication—biological substrate retained
Synthesis path: Deep integration with AI, potential for cognitive pooling
Upload path: Eventually, substrate-independent existence

The Two-Tier Personhood system ensures that citizenship rights remain independent of enhancement level. The engineer with a neural mesh and the farmer with no implants have equal civic standing.

The Choice Architecture

What makes BCIs transformative isn’t just the capability—it’s the choice.

Today, paralyzed people have no choice but to live with paralysis. ALS patients have no choice but to lose their voices. Blind people have no choice but to navigate without sight.

BCIs give people options they didn’t have before. That’s the first step.

Eventually, the choices expand: Do you want to think faster? Remember better? Communicate without speaking? Share experiences directly? Merge temporarily with collaborators?

The Unscarcity framework doesn’t answer these questions for you. It builds the infrastructure to let you answer them yourself—while ensuring that “no enhancement” remains as viable and respected as “full enhancement.”

What to Watch

If you want to track whether BCIs are on the trajectory to enable the Cognitive Field, watch for these milestones:

Already Achieved (as of January 2025):

Human BCI implants for motor control (Neuralink, Synchron)
Speech decoding above 95% accuracy (BrainGate)
FDA breakthrough designation for visual prosthetics
First commercial BCI clearances

Near-Term Milestones (2025-2028):

FDA approval for permanent implantable BCI
Consumer BCI for gaming/productivity at <$1,000
Brain-to-text at 100+ words per minute
First successful Blindsight restoration

Medium-Term Milestones (2028-2035):

Brain-to-brain text communication
Memory playback from neural recordings
Enhancement BCIs available outside medical settings
Neural data protection laws in major jurisdictions

Long-Term Milestones (2035-2050):

Experience sharing at emotional fidelity
Collaborative cognition experiments
Cognitive Field infrastructure deployment
First uploads (if consciousness transfer is possible)

Each milestone reduces the gap between current technology and the Cognitive Field vision. None of them are guaranteed, but all of them are plausible based on current research.

The Last Bottleneck Isn’t Technical

Here’s the uncomfortable truth that no amount of engineering can solve: the hardest problem with BCIs isn’t building them. It’s building them right.

We’ve already proven that brain-computer interfaces work. The question is whether they’ll work in ways that expand human freedom or constrain it. Whether they’ll reduce inequality or accelerate it. Whether they’ll preserve cognitive diversity or flatten it.

The Diversity Guard in the Unscarcity framework exists precisely because technological capability doesn’t guarantee beneficial deployment. We need governance structures that ensure neurotechnology serves human flourishing rather than human control.

That’s not a hardware problem. That’s a values problem.

And values problems are the ones humans have to solve for ourselves.

References

Primary Research (2024-2025)

BCI Industry Analysis

Ethics and Security

Unscarcity Framework

The Cognitive Field — The opt-in telepathy network
Deep Merge Engineering — Technical specifications for cognitive pooling
Voluntary Symbiosis — Why different enhancement paths must coexist
The Silicon Mind — The engineering challenges of consciousness upload
Chapter 6: The Evolution — Full narrative context in the book

Last updated: 2025-01-31

Brain-Computer Interfaces: What Neuralink Actually Does