Skip to content
  1. Home
  2. AI & Machines
AI & Machines

The first consumer neural interfaces will be boring, and that is the point

Forget mind reading. The shipping devices detect intent to click, and that narrow promise is what makes them viable.

The first consumer neural interfaces will be boring, and that is the point
The first consumer neural interfaces will be boring, and that is the point — XENVORA Photograph: EEG Recording Cap, Chris Hope / Wikimedia (CC BY 2.0) · Licence

The demos that go viral show cursors moving by thought alone. The products heading to shelves do something far smaller: they read a handful of muscle and nerve signals at the wrist and turn them into two or three reliable gestures.

Narrow beats magical

A device that recognises a pinch 99.4% of the time is more useful than one that infers intent 80% of the time. Accessibility teams figured this out years ago, and consumer hardware is finally following their lead.

The interesting constraint is latency, not bandwidth. Anything above roughly 80 milliseconds stops feeling like your own body and starts feeling like a remote control with a bad connection.

Expect the first wave to be unglamorous: silent input for headsets, one-handed control while carrying something, and a genuinely better answer for people who cannot use a touchscreen.

Omar Adel
Machine Learning Correspondent
Omar Adel

Omar reports on applied machine learning, with a stubborn habit of asking vendors for reproducible benchmarks.

Reader comments 0 comments

Sign in to comment.

No comments yet. Be the first.