Brain-Computer Interfaces

Brain computer interfaces (BCI) connect the central nervous system to external technology directly. Imagine your computer receiving directions from your thoughts rather than your keyboard and mouse. AE is developing cutting-edge machine learning software to understand patterns in neurological activity like electrical signals and blood flow. As longtermists, we believe BCI will ultimately become the primary means of human interaction with each other and machines. As optimists, we believe the medical, and economic implications are unimaginable. But unless we want ads interrupting our thoughts, every line of code must be guided by neuroethical principles that increase human agency!

callout image

AE Takes First Place in the NLB Challenge!

AE’s data science team offers the world’s best models for extracting patterns from neurological data, as proven by our victory in the neural latents benchmark challenge (NLB) in every category. Among its entrants were the finest neurological research labs on the planet. This is the first step on a long journey, and every participant will play a part. We will play ours. To support the great work of those pushing the boundaries of knowledge and the future of Brain Computer Interfaces, AE will be reinvesting any prize money back into the neuroscience community. How so? Stay tuned, we’ll have an exciting announcement in a few days.

0. What Are Brain-Computer Interfaces?

Brain Computer Interfaces (BCI) describe any direct interface between the central nervous system and an external technology.

BCIs are the hardware and software that allow information to flow directly between brains and computing devices (computers, smartphones, robots, electrical simulators of limbs, etc.)

Current technology is indirect: From your brain to hands to keyboards/mice to computing device to screens back to eyes and back to your brain.

BCI technology shortens the path:  From your brain to the computing device to a brain (yours or someone else’s!).

The hardware reads brain activity from electrical signals, blood flow, or other features that reflect the underlying neural activity.

The software reveals patterns in those signals, interprets those patterns, and converts them into commands a computer can understand.


1. What are we doing?

Developing the software component of BCI. In short, we develop machine learning algorithms to interpret brain activity (neural decoding).

Real-time decoding and BCI UX design

Collaborating with research groups to advance the state of the art (and participating in machine learning competitions).

Amplifying the impact of the world’s best research labs by providing better algorithms, better software, and better standards.

Reading, writing, discussing, and planning the implications of BCI within our optimistic, longtermist view of society.

Hiring the best machine learning engineers and neuroscientists on planet earth to increase human agency in ways we have not yet imagined.


2. Why Does This Matter?

Human agency is increased through our relationship with computers, and BCI will likely magnify this exponentially if developed with that intent (or diminish agency if its creators are narrowly-focused on short-term financial gain).

Currently, BCI technology is being deployed to restore hearing via cochlear implants, restore use of paralyzed limbs, and to treat Parkinson’s.  And yet, it remains a neglected topic.

Research is currently limited by availability of data. This is driven by the hardware problem of extracting signals from the brain and our ability to standardize, share, analyze, and act on those data.

Transcription fluency is improved when computers respond directly to thoughts, rather than buttons and clicks and humans can communicate seamlessly.

These interactions will become deeply personal when the response enters the brain directly.

BCI should enhance the agency of its users rather than interrupting us mid-thought with ads and selling our brain data to the highest bidder.

Dark patterns are bad, but interruption of thought without consent is worse (and we can expect both in BCI if it is designed like most modern digital technology).


Want to talk?

Want to discuss BCI or simply learn more about AE?

3. What Outcome Do We Want?

The creation of BCI software that maximizes human agency.

The creation of BCI software that incorporates (and advances) neuroethics at its core and throughout the development process.

The creation of BCI that respects human consciousness (no unwanted neurofeedback, “neuromarketing”, hacking, “brain-jacking”, or other invasions of privacy) and enables seamless natural transitions between thoughts and improves understanding of self.

Users controlling who uses their data and how.

Users controlling any extension of their thoughts and self (and must, with agency and understanding, opt-in and consent to any access thereof).

Better mental health, better transcription fluency, restoration of agency for the paralyzed, and most importantly, inspiration for the future tech giants to place human agency at the center of their objectives (it’s good for people and good long-term business).


4. How Can We Get There?

Maintaining independence and avoiding shareholders, venture capital, or private equity

(AE is already a fully-bootstrapped business that has grown from 0 to ~100 with agency-increasing BCI as its ambitious Big, Hairy, Audacious Goal).

Growing a profitable, longtermist software company, training the best developers, designers, and engineers on the planet (or any other), building agency-increasing products for those clients and ourselves, and investing in altruistic, agency-increasing BCI initiatives

(We love funding internal skunkworks projects, and VCs help develop incredible products and incredible profits, but decisions surrounding the development of BCI must be motivated by increasing human agency exclusively)

(Google’s founders never wanted ads, but their investors did).

Following the core principles of increasing human agency, rather than short-term financial incentives (like making sure that Google Nest doesn’t send you lightbulb ads if you think the room is dark or that MetaTap doesn’t send you an ad for beer if you’re thirsty).

Following our core values (A/B testing baby steps, and realizing that we are only 1% as good as we could be).


5. What Can You Do?

If you are an academic researcher looking to improve your neural decoding algorithms, come share your experience with us!

If you are a hardware company interested in joining a select list of early adopters licensing our technology for decoding brain signals using state-of-the art techniques, or want to understand how the data from your hardware can help improve human agency, come talk to us!

If you are a developer or data scientist, excited, curious, or concerned about the future of BCI, come work with us!