Anyone else just watch the Neuralink 2022 presentation that just ended?
I’m really excited for this stuff and what it means as a consumer/gamer. The Neuralink and Elon are very keenly aware of this aspect of the technology. One of the team members who talks with a focus group of actual handicapped people on what kinds of things they want to do most in life (they all have some kind of neurological or motor function issue)
and one patient said he most wanted to play shooter games again. So Neuralink is very aware of these kinds of wishes, and said it’s very much in the realm of possibility one day
Would you play OW with a brain chip? Youd essentially be staring at your monitor and think what you want to do, and it streams your “output” thoughts to a nearby computer or phone, which interprets them into whatever you need, in this case would be controller/keyboard/mouse input for the game.
It could probably be set up to have your eyes movement control the actual aiming direction/cursor on-screen.
One example they showed with their test monkeys had an infrared camera tracking the monkey’s eye movements while it looked at a screen with dots that appeared. If the monkey looked at a dot, it would get a small food treat.
Then they would send a signal into it’s brain chip to simulate the dot directly in the monkey’s mind, but the dot wouldn’t actually appear on the monitor, and the monkey would still look at it and get the treat
From what my armchair scientist mind could understand, there’s currently 1024 channels in the current model that can send out or receive bits. So they can essentially give you a 1024 pixel image implanted directly into your mind, bypassing your eyes
These 1,024 channels can also be used as bits, or 1 and 0, on and off. Pressing a key on a keyboard is 1 bit of information, so the average human types maybe 5-10 bits per second, which is magnitudes below what the current neuralink chip can send out/receive. So you could type or send movement commands, up to 1024 per second which is way above what a human could think at
Their next model will be 16,000 channels, and they plan on letting it double up and install 2 at once, one on each side of the visual processing area of brain which wraps around back of head and is both on left and right
So that’s 32,000 total channels, or a 32,000 pixel image. They also talked about compression and algorithms to not have to take a 1 to 1 approach, but maybe 1 to 10 or 1 to 100
Idk I just am really excited and wanted to talk about this stuff. An audience member asked about the long term plans and Elon mentioned memory being very important future goal of theirs to work towards, being able to save memories (or even write them from outside program, the example joke being that they could teach you kung fu with a download)
I am really excited, esp seeing their progress and upgrades from 2 years ago, and where they plan to go in next iteration. They even improved from the previous surgery model and no longer remove skull to put chip on brain.
The new model sits below the skin but on top of the dura, a hard “shell” between skull and brain, and the threads are inserted through the dura itself, so there’s no longer chip-brain surface contact (where healing tissue would form in previous versions)
The main bottleneck right now, in the grand picture, is thread length. They are so small (like 10x less thick than a hair) and so short, and the needle insertion is so tiny that they can’t go very deep into the brain, and avoiding veins is a major issue as well, you can only go straight and will run into them all over
So they can’t reach memory or emotion sections right now, basically nowhere but areas close to the outer edge, which is why they’re focused on visual and muscular sections, visual and the spine which gives movement info to rest of body