I'm a Neuralink skeptic, at least in the short term, but this article is ridiculous.
Apparently the first three things "Neuralink is missing" are the highly technical problems that are core to their business: how to implant a device without the brain rejecting it and covering it in scar tissue, how to digitize and transmit brain signals, and finally the machine learning to decode those signals and convert them into the actions the patient really wishes to perform.
And the fourth problem is that Medicare might not pay for the devices.
Who could possibly think Neuralink don't have teams dedicated to solving and improving the first three? Everything out of the company says they're working hard on all of these, they don't consider any of them solved problems or things they can handwave away.
As for the fourth, if the product is good enough, someone will pay for it. Perhaps it will be healthy wealthy private individuals first. I'd be more worried if Neuralink was devoting all its efforts to lobbying the FDA or wining and dining insurance executives, rather than focusing on the technological issues that need to be solved first.
Whatever Neuralink is missing, it's none of these things. They're certainly missing a lot, as is their owner, but it's absolutely none of these things. Like, we've been using brain electrodes to control devices for a very long time now. They know the issues. As they are originally a neuroprosthetics company whose whole goal was to develop means for paralyzed people to control limbs and devices, they're perfectly in their zone. That's not their issue.
As what their owner wants them to be, with data transmission into the brain and individually addressable neurons, they're leagues off from a functional product. Extracting and interpreting brain signals is, while not simple, well-known and well-practiced. The other direction is practically unexplored and the most we normally do is poke around with a bit of electricity and hope it makes the brain work better. To move data from a computer into the brain would be a total revolution. Needless to say, their current product does not do this, and they've shown zero development to this end. Which kind of makes it feel like they're hoodwinking Musk to develop actually useful medical devices instead of playing into his naive technocratic futurism. That just means they know what they're doing even more strongly than what their other actions have already shown.
> Extracting and interpreting brain signals is, while not simple, well-known and well-practiced. Their current product does not do this, and they've shown zero development to this end.
I was following you until this. Can you clarify? Isn’t their most recent work precisely an example of inferring and applying a person’s intent by deciphering their brain signals?
Ah, sorry, one of those situations where you think about what you'll write and don't end up writing it down. I am of course referring to the inverse in the second sentence, moving data into a brain from a computer. Corrected in the original comment now, thanks!
> if the product is good enough, someone will pay for it.
Many people are critical of US health care, with its costs and inefficiencies, However, it does let you pay money and have a procedure done. captitalism.
In some other countries, you can't do that, sometimes at the cost of your life.
Seeing the demo did make me wonder why they chose cursor movement as a first achievement even though that's also possible using eye tracking. Still, seems like something that could greatly benefit people so I hope it will.
10 comments
[ 4.0 ms ] story [ 32.6 ms ] threadApparently the first three things "Neuralink is missing" are the highly technical problems that are core to their business: how to implant a device without the brain rejecting it and covering it in scar tissue, how to digitize and transmit brain signals, and finally the machine learning to decode those signals and convert them into the actions the patient really wishes to perform.
And the fourth problem is that Medicare might not pay for the devices.
Who could possibly think Neuralink don't have teams dedicated to solving and improving the first three? Everything out of the company says they're working hard on all of these, they don't consider any of them solved problems or things they can handwave away.
As for the fourth, if the product is good enough, someone will pay for it. Perhaps it will be healthy wealthy private individuals first. I'd be more worried if Neuralink was devoting all its efforts to lobbying the FDA or wining and dining insurance executives, rather than focusing on the technological issues that need to be solved first.
As what their owner wants them to be, with data transmission into the brain and individually addressable neurons, they're leagues off from a functional product. Extracting and interpreting brain signals is, while not simple, well-known and well-practiced. The other direction is practically unexplored and the most we normally do is poke around with a bit of electricity and hope it makes the brain work better. To move data from a computer into the brain would be a total revolution. Needless to say, their current product does not do this, and they've shown zero development to this end. Which kind of makes it feel like they're hoodwinking Musk to develop actually useful medical devices instead of playing into his naive technocratic futurism. That just means they know what they're doing even more strongly than what their other actions have already shown.
I was following you until this. Can you clarify? Isn’t their most recent work precisely an example of inferring and applying a person’s intent by deciphering their brain signals?
Many people are critical of US health care, with its costs and inefficiencies, However, it does let you pay money and have a procedure done. captitalism.
In some other countries, you can't do that, sometimes at the cost of your life.
https://en.m.wikipedia.org/wiki/Matt_Nagle