summaryrefslogtreecommitdiffstats
path: root/ang/readings/12/dokument.md
blob: ba61772c67cd1681559704381c3030de8a490f05 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
# Neuralink surprised the public with an implant-controlled joystick

We were surprised by the publication of a video in which Neuarlink shows a macaque, which wirelessly directs the cursor on the screen and plays pong via a brain implant. It is a confirmation of Elon Musk's past claims, but it is not entirely clear why it happened right now, as it is not accompanied by a scientific publication.

When Neuralink, Elon Musk's company for the development of biocybernetic nerve implants, presented the first rats with implants two years ago, Musk, by the way, accedentally revealed that they are already experimenting on doing something similar with monkeys. But at last year's infamous presentation of scaled-down wireless devices, most of the demonstrations were done on pigs, and no monkeys were mentioned. Since primate testing is practically necessary in order to proceed with human experiments, it was only a matter of time before we saw this kind of laboratory animal playing with a computer at a distance. That role now belonges to Pager, a nine-year-old macaque.

Six weeks before the video, Pager was implanted with two implants with a thousand electrodes, one on each side of the head above the lobe. The training took place with the help of a prize in the form of a banana frappe, when Pager brought the cursor to the desired location with a joystick. After some time, Neuralink's control algorithm learned to decipher Pager's brain signals and interpret the monkey's intention to move his hands from them. As shown in the video, the joystick was then disconnected and the macaque cursor on the screen was controlled only through implant signals. In this way, he then played the whole game of Pong, with a relatively fast movement of the ball, which is a significant achievement and something so advanced in Neuralink has not been shown so far.

It is an interesting and extremely inspiring confirmation of Elon Musk's past claims, but it should be read carefully. This time, in addition, they did not deliver a publication in any of the scientific publications that would provide a more detailed insight into the technical details of the venture, and the presentation was also not public. Therefore, it is not entirely clear why such a way of publishing, as Neuralink has shown the novelties so far at carefully prepared major events. At the end of the video, there is again a call for recruiting new collaborators, and cynical observers might point out that the lab has been seriously grappling with the departure of burned-out researchers for some time. The only accompanying activity was a series of Musk’s tweets, in which he reiterated that their main goal is to help people with serious nerve damage. Whether we can also hope for any announcements of experiments on humans in the near future remains unknown.

Another important detail is in the very nature of technological achievement. To be honest, this kind of performance has been shown in the past - but with one important difference: at the time, it was still wired implants. The Neuralink implant is the first to boast of the displayed performance of wireless signal transmission and such successful miniaturization that the device is invisible from the outside. This is quite a big achievement because the signals from the brain electrodes require a notoriously large bandwidth. The electrochemical activity of neurons is, according to the peasant, quite chaotic, and usually when excited they do not trigger a single discrete shock, but a series of such different strengths, which is not easy to translate into meaningful electrical commands. Neuralink engineers seem to have managed to compress the data in some way, most likely also by cutting a good chunk of the signal. This is a real advancement in this publication and, last but not least, the direction in which the vast majority of research will have to go, as our understanding of the exact meaning of brain signals still falters and is a much bigger obstacle to commonly used implants than electrode design itself.