Elon Musk’s Neuralink has demonstrated in a recently released video how he used his sensor hardware and brain implant to record a baseline of a macaque’s activity while playing an on-screen game in which he had to move a token with a joystick to different places with his hand.
Based on the basic data, Neuralink was able to use machine learning to predict where the macaque named Pager would move the physical controller, and he was able to accurately predict it before the move was actually made. Researchers then removed the paddle completely and did the same with Pong, where it ended up in a place where Pager no longer even moved his hand in the air on the non-existent paddle. Instead, Pager controlled the game through his mind via the link hardware and embedded neural threads.
In an earlier blog, the company explained that the same technology could help manipulate a cursor on a computer and even extend it even further to touch controls on an iPhone and tap on a virtual keyboard.
Future iterations of the product include transmission between an in-brain node and neural pathways in the legs, allowing paraplegics to walk again.
Despite these wildly risky claims, the company has cited numerous research findings that underscore its existing demonstrations and short-term goals.
For more information, see the original story of Techcrunch.