I’m glad they’re having success, but this is not a use case for neuralink. It’s a use case for brain-computer interfaces in general. Research in BCIs is ongoing and a very relevant topic of research, but usually research institutes can’t indiscriminately kill monkeys for this. I am very excited about this tech (my PhD is tangentially related to it) but wish it was being developed by literally any research institute not owned by Musk
I might be completely stupid and wrong about this and everything, but I don’t think I’m too far off.
Essentially Neuralink is just very high tier EEG, electroencephalography, measuring brainwaves.
You can buy yourself a kit for 100 bucks or so. But… you’ll get like a two-channel device with a few sensors, which won’t really have any sort of accuracy. You could go for a professional model for a few thousand or something, and have good sensors and a few dozen channels.
They’ll still be outside the skull though.
I just looked this up the other day, and uh, guess the amount of channels on Neuralink?
“Neuralink has developed an application-specific integrated circuit to create a 1,536-channel recording system. This system consists of 256 amplifiers capable of being individually programmed, analog-to-digital converters within the chip and peripheral circuit control to serialize the digitized information obtained.”
So I don’t know what would be a good comparison to the scale difference here. Perhaps one is like listening to someone having sex in the next room while there’s also loud music playing (limited channels, poor-ish signal), and the other is actually being in that room, having that sex yourself (100x the channels, signal pretty much straight from the source.)
Someone smarter can correct me where I went wrong.
This technology is fascinating, I’m just so annoyed/apprehensive Musk owns the company doing the pioneering work.
Having a thousand, or a hundred, or a dozen channels means nothing if you can’t individually control them. You can use a hundred channels to sample a brain, but if you’re playing chess (As in the famous video) all you need is “up/down, left/Right, click”. That’s 3.
The brain is incredibly complicated and there is a lot of noise. More channels do make a huge difference. It’s not like each channel is associated to one direction or something and you only need 4. Also, to add to the other users point, the signal strength is much stronger inside than out since the skull acts like a barrier.
Your example is the equivalent of saying you don’t need a keyboard because you can use a computer with just the arrow keys and the spacebar.
Although BCIs are already a thing, the difference between what you can do with external compared to internal ones is probably quite vaste at our current tech level and will probably lead to better external ones at the same time.
You also don’t control the electrodes, you just receive the signal from them and they are obviously all being sampled.
Well you’ll need three states which are easily translated to the commands.
The more channels you have, the easier it is tor recognise any specific configuration of brainwaves.
It’s not like thoughts are as easy to pinpoint as spatial coordinates. You don’t really decide what your specific brainwave output will be when you focus on thinking “up”. It will be a combination of different brain waves, and the more accurately you can measure that, the easier it is for the computer to pick up when you’re thinking “up”.
Try playing a tune with just three tones available to you, it’d be hard.
The technology isn’t new, but the resolution sort of is.
Hell they used the tech in a House episode in like, 2009.
What do you mean “if you can’t control them”? The channels are how many bands you’re reading, and they are using quite fancy computers to do the reading. What’s there to control? BCI is just a fancier interface than a keyboard, it’s still only one-way, so I don’t understand what you mean by “control” in this context.
I’m glad they’re having success, but this is not a use case for neuralink. It’s a use case for brain-computer interfaces in general. Research in BCIs is ongoing and a very relevant topic of research, but usually research institutes can’t indiscriminately kill monkeys for this. I am very excited about this tech (my PhD is tangentially related to it) but wish it was being developed by literally any research institute not owned by Musk
I’m also pretty sure you can do this without implants…
Yeah, you’re not wrong, per se.
I might be completely stupid and wrong about this and everything, but I don’t think I’m too far off.
Essentially Neuralink is just very high tier EEG, electroencephalography, measuring brainwaves.
You can buy yourself a kit for 100 bucks or so. But… you’ll get like a two-channel device with a few sensors, which won’t really have any sort of accuracy. You could go for a professional model for a few thousand or something, and have good sensors and a few dozen channels.
They’ll still be outside the skull though.
I just looked this up the other day, and uh, guess the amount of channels on Neuralink?
“Neuralink has developed an application-specific integrated circuit to create a 1,536-channel recording system. This system consists of 256 amplifiers capable of being individually programmed, analog-to-digital converters within the chip and peripheral circuit control to serialize the digitized information obtained.”
So I don’t know what would be a good comparison to the scale difference here. Perhaps one is like listening to someone having sex in the next room while there’s also loud music playing (limited channels, poor-ish signal), and the other is actually being in that room, having that sex yourself (100x the channels, signal pretty much straight from the source.)
Someone smarter can correct me where I went wrong.
This technology is fascinating, I’m just so annoyed/apprehensive Musk owns the company doing the pioneering work.
Having a thousand, or a hundred, or a dozen channels means nothing if you can’t individually control them. You can use a hundred channels to sample a brain, but if you’re playing chess (As in the famous video) all you need is “up/down, left/Right, click”. That’s 3.
The brain is incredibly complicated and there is a lot of noise. More channels do make a huge difference. It’s not like each channel is associated to one direction or something and you only need 4. Also, to add to the other users point, the signal strength is much stronger inside than out since the skull acts like a barrier.
Your example is the equivalent of saying you don’t need a keyboard because you can use a computer with just the arrow keys and the spacebar.
Although BCIs are already a thing, the difference between what you can do with external compared to internal ones is probably quite vaste at our current tech level and will probably lead to better external ones at the same time.
You also don’t control the electrodes, you just receive the signal from them and they are obviously all being sampled.
Well you’ll need three states which are easily translated to the commands.
The more channels you have, the easier it is tor recognise any specific configuration of brainwaves.
It’s not like thoughts are as easy to pinpoint as spatial coordinates. You don’t really decide what your specific brainwave output will be when you focus on thinking “up”. It will be a combination of different brain waves, and the more accurately you can measure that, the easier it is for the computer to pick up when you’re thinking “up”.
Try playing a tune with just three tones available to you, it’d be hard.
The technology isn’t new, but the resolution sort of is.
Hell they used the tech in a House episode in like, 2009.
What do you mean “if you can’t control them”? The channels are how many bands you’re reading, and they are using quite fancy computers to do the reading. What’s there to control? BCI is just a fancier interface than a keyboard, it’s still only one-way, so I don’t understand what you mean by “control” in this context.