Do people think this is new? We have been able to do this for decades. I’m a lowly PhD student and even I get to work with humans whose brains we are actively recording from (although I don’t put the electrodes in there myself).
Just another instance of Muskrat talking about things he doesn’t know. I used to think he was a genius when he was talking about rockets, then he started talking about things I know (neuro & AI)…
Elon never invents anything new. He finds a complex concept, scuffs off at it’s complexity and announces it’s actually really simple. Creates company that over-simplifies things.
Sometimes his project fails enough times that it starts working (space x).
He has so much money that he can keep doing it. And hire the best in the field - there’s no money in academia so of course they’ll go. And then he’ll take credit for their hard work eventually of course.
So what are actually useful applications that might be feasible soon for this kind of stuff? I could google it but I’d mostly get a bunch of sensationalist BS that is meant to generate clicks.
BCI - Brain Computer Interface. This can allow people with disabilities to control prosthetics using their brains. For example, this one from 20+ yrs ago. They are in clinical trial stages now - lot of data over 20yrs showing it’s pretty safe. There are some differences like BrainGate uses “Utah” electrodes which sit on the brain rather than go inside the brain.
Medical diagnosis - Some patients (with things like epilepsy) get their brains recorded like this to find the region of the brain that is malfunctioning. Then sometimes this region is removed and believe it or not it actually helps! Edit: DBS is another option sometimes like the other commenter said but that needs “stimulation” also, not just passive recording.
Understanding the brain - these recording data can help make sense of the brain. We still don’t understand much of how the brain works so this data can help and maybe help with treatments in the future.
For all of these currently we only have patients (because “healthy” people wouldn’t want metal electrodes in their brain). But neuralink’s promise is to make these electrodes so thin and dense (so that you can record more) while keeping SNR high that it might be possible to put it in healthy people without brain damage. I wouldn’t hold my breath for that, though.
I’m hoping so hard for a brain/computer interface. I have a chronic condition that makes me a walking repetitive stress injury generator. Being able to control a computer with my noggin would be a game changer. I currently use an eye tracker combined with a camera head tracker, plus speech recognition, but it’s not the best. It certainly killed my (non-existent) computer programming career.
Both of them can be done shitty-ly now. But to do it with quality that even healthy people will voluntarily get it? That would need several breakthroughs.
We can stimulate some neurons now; to be able to stimulate enough neurons to do either of those in good quality will be hard. Cutting edge stuff can stimulate ~1000 neurons (only monkeys not even humans) but the human optical nerve is more than a million fibers. So we probably will need 3 orders of magnitude improvement and somehow do it in humans safely.
Didn’t the team he put together at least come up with better miniaturization of a BCI, with denser/more numerous electricodes, and a more advanced implantation process to minimize scarring?
It’s a current limitation with all BCIs from what I understand, so one hurdle for them is to try to eliminate it from ever forming at all. Not sure where his thing is at, but I guess no one here knows anything about it.
Elon musk is a fucking moron, but he is paying a team of actual neuroscientists and surgeons to develop this.
Yes you are right, they are trying to improve on what exists. My response was more for the “OMG musk is doing a sci-fi” - recording spikes is not really new or hard.
Do people think this is new? We have been able to do this for decades. I’m a lowly PhD student and even I get to work with humans whose brains we are actively recording from (although I don’t put the electrodes in there myself).
Just another instance of Muskrat talking about things he doesn’t know. I used to think he was a genius when he was talking about rockets, then he started talking about things I know (neuro & AI)…
Elon never invents anything new. He finds a complex concept, scuffs off at it’s complexity and announces it’s actually really simple. Creates company that over-simplifies things.
Sometimes his project fails enough times that it starts working (space x).
He has so much money that he can keep doing it. And hire the best in the field - there’s no money in academia so of course they’ll go. And then he’ll take credit for their hard work eventually of course.
So what are actually useful applications that might be feasible soon for this kind of stuff? I could google it but I’d mostly get a bunch of sensationalist BS that is meant to generate clicks.
Basically three things -
BCI - Brain Computer Interface. This can allow people with disabilities to control prosthetics using their brains. For example, this one from 20+ yrs ago. They are in clinical trial stages now - lot of data over 20yrs showing it’s pretty safe. There are some differences like BrainGate uses “Utah” electrodes which sit on the brain rather than go inside the brain.
Medical diagnosis - Some patients (with things like epilepsy) get their brains recorded like this to find the region of the brain that is malfunctioning. Then sometimes this region is removed and believe it or not it actually helps! Edit: DBS is another option sometimes like the other commenter said but that needs “stimulation” also, not just passive recording.
Understanding the brain - these recording data can help make sense of the brain. We still don’t understand much of how the brain works so this data can help and maybe help with treatments in the future.
For all of these currently we only have patients (because “healthy” people wouldn’t want metal electrodes in their brain). But neuralink’s promise is to make these electrodes so thin and dense (so that you can record more) while keeping SNR high that it might be possible to put it in healthy people without brain damage. I wouldn’t hold my breath for that, though.
I’m hoping so hard for a brain/computer interface. I have a chronic condition that makes me a walking repetitive stress injury generator. Being able to control a computer with my noggin would be a game changer. I currently use an eye tracker combined with a camera head tracker, plus speech recognition, but it’s not the best. It certainly killed my (non-existent) computer programming career.
Thanks! So how far away are we from something like this:
Both of them can be done shitty-ly now. But to do it with quality that even healthy people will voluntarily get it? That would need several breakthroughs.
We can stimulate some neurons now; to be able to stimulate enough neurons to do either of those in good quality will be hard. Cutting edge stuff can stimulate ~1000 neurons (only monkeys not even humans) but the human optical nerve is more than a million fibers. So we probably will need 3 orders of magnitude improvement and somehow do it in humans safely.
Didn’t the team he put together at least come up with better miniaturization of a BCI, with denser/more numerous electricodes, and a more advanced implantation process to minimize scarring?
You probably won’t be worried about scarring after you die of a brain bleed so that might not be the best selling point.
It’s a current limitation with all BCIs from what I understand, so one hurdle for them is to try to eliminate it from ever forming at all. Not sure where his thing is at, but I guess no one here knows anything about it.
Elon musk is a fucking moron, but he is paying a team of actual neuroscientists and surgeons to develop this.
Yes you are right, they are trying to improve on what exists. My response was more for the “OMG musk is doing a sci-fi” - recording spikes is not really new or hard.