[1] https://www.reddit.com/r/science/comments/5c2ljz/brain_impla...
They could record the way my brain sends walk signals and if I end up paralysed they have the data to train the model.
There are multiple spinal conditions that result in paralysis after not be been paralysed, not everyone who ends up in a wheelchair gets their by an accident.
Then you would have an entire generation thinking millions of times deeper, broader, and faster than the previous one, and communication mostly broken. I sort of fear this scenario because I would be on the wrong side of it.
Do you use it?
Bio is like that. We really don't know what is going on with a body most of the time. When we try to leverage it, it gives totally nutto results. Like, yeah, water is good for you! Like, maybe we should make it inhale-able! Or yeah, caffeine is alright, makes you wake-up! You should just take NoDoz and never sleep again!
Those are absurd examples, but I hope they illustrate that in bio, 'leveraging' something typically means you are gonna die either real soon or you are gonna get cancer. Fen-phen is a great example of 'leveraging' something in bio [0]. Yeah, you loose weight, but then you blow your heart valves out and end up as worm food.
So when you say that we can just throw computers at the problem, you make a huge mistake with that idea. Your input data is everything. You have to run a LOT of experiments in the real world to get the right data, not just noise you think is data. Living things are really good at staying alive until they aren't. They are not just noisy, they are also altering the experiment along with the experimenter. Homeostasis is a real thing and bodies and living things try to stay happy, healthy, and alive. Like, here is the thing, your experimental test-bed is alive and thinking and is, in some cases, trying to experiment-on/kill-you right back. Try feeding that into a computer! You have 300k+ dependent variables (that you know of) and 10 equations (if you trust those other authors). You are trying to find the mean of N interacting Poisson/Gaussian data sets (you hope they are normal) in N dimensional space and you know that you don't know all the variables and you can only do this on 5 rats.
But maybe it is. You can also put people on a VR environment and have them "walk" there (see experiments by Nicolelis et. al.) and try to capture signals like that
Let's guess that each person's unique representation for walking does not change, or changes slowly. That seems reasonable since it is at least partially based on the physical arrangement of neurons.
We could bank it on a regular basis--recording via a device we wear around for a month every 5 years, for example. Then if we are paralyzed, the most recent recording is used to map an implant.
Personally, I will not surprised if we find that each person's nervous system is so different from the next that nothing significant is generalizable. Rather than each person running an "OS on standard hardware" we might find that a more accurate analogy is that each person has constructed their own custom logic circuit and then wrote custom software on top of that.