If you cant beat it , join it....says Elon Musk

[Crispy]

This is where it enters science fiction. Two things, really:

1) using fMRI to find individual thoughts is like using a fold-out map of the world to find a footpath in your local woods. The resolution is crazily low compared to what you would need to get even close. The best fMRI can't really do better than a 1mm resolution over a time scale of about 1 second. That's a space that would contain literally billions of neurons firing thousands of times.

Yet despite that staggering crudeness, fMRI combined with some simple analysis can guess the general meaning of your thought with reasonable accuracy: Beyond Bananas: CMU Scientists Harness “Mind Reading” Technology to Decode Complex Thoughts - Dietrich College of Humanities and Social Sciences - Carnegie Mellon University
With orders of magnitude improvement in resolution, and more sophisticated analysis, this can be imporved a lot.

2) that's not how the brain works anyway. If you think about 1 and then the next day you think about 1, it will almost certainly not involve exactly the same brain pathways. You always exist in a context and your thinking responds to that context, making new connections and using different ones to form the same thoughts. That's one of the ways your brain is fundamentally different to a computer. Yes, we can say that there is a broad region responsible for dealing with numbers. But we can't say that thinking about "1" fires this specific neuron, even within a single brain.

You don't need to fire a particular neuron, just fire large groups of neurons in particular ways. What makes this easier is that the brain can do a lot of the hard work for you. If you get instant feedback on your attempt to think "1" then it doesn't take much practice for the "correct" way of thinking "1" to become second nature.

We will train these systems on our brains, and our brains will train themselves on these systems.

 

[kabbes]

Well, I guess we’ll see. But all this complexity of the different ways the things we are closest to computers in understanding (numbers) pales into insignificance compared to the way the brain interprets other contexts. If it is to be seen whether there is anything that can be usefully created out of a direct interface to thinking “1”, that’s nothing compared with the incompatibility of systems like thought speed or memory. As I say, that’s where it become nonsense.

 

[Crispy]

thought speed or memory. As I say, that’s where it become nonsense.

Agreed. Although from the user's point of view, the line between "what I'm thinking" and "what I'm asking the computer and what it's telling me in return" might get rather fuzzy.

 

[weltweit]

We are communicating with technology in various ways at the moment, we use our fingers and a keyboard for PCs and our eyes as inputs, etc, and we use the spoken word and hearing for the likes of Alexa. These use existing human input / output channels fingers, eyes, voices and ears.

Effectively could a first, more integrated, approach be to target the nerve endings that generate voice and hearing for an example?

 

[Crispy]

We are communicating with technology in various ways at the moment, we use our fingers and a keyboard for PCs and our eyes as inputs, etc, and we use the spoken word and hearing for the likes of Alexa. These use existing human input / output channels fingers, eyes, voices and ears.

Effectively could a first, more integrated, approach be to target the nerve endings that generate voice and hearing for an example?

Yes. Although every user would need to train their own device with their own brain. The learning process might be quite lengthy.

 

[weltweit]

Yes. Although every user would need to train their own device with their own brain. The learning process might be quite lengthy.

Pretty incredible though the existing learning that permits our brains to see through our eyes typed words in a particular language on a screen, decode them into coherent messages which our brains can then act further on etc..

 

[Crispy]

Pretty incredible though the existing learning that permits our brains to see through our eyes typed words in a particular language on a screen, decode them into coherent messages which our brains can then act further on etc..

Yep. We see with our brains more than our eyes.

 

[weltweit]

..
Effectively could a first, more integrated, approach be to target the nerve endings that generate voice and hearing for an example?

Perhaps a head up display on a pair of glasses and an ear piece.
Or a mike and an earpiece ..
for stage 1.