I'm in neuro, that scenario is centuries away. You'd peobably need a clear skull first with exotic glasses we can't even begin to understand physics-wise today. We don't even know what your brain cells mostly are. Is it 10% neurons and 90% glia, or 50% neurons and 50% glia? What are glia even doing? The only data we really get on them is via stroke studies. Neuro is exciting precisely because there is so little known about the brain still.
But we don't really need to understand something to create technology leveraging it. And we are just getting started with deep learning. And our computers are extremely slow and have tiny storage compared to what they will be.
Ok, I have this pill/wire-probe/exercise-machine, I don't really know what it does, but in some folks it seems to make them happier/healthier, that is all the info you are going to get about said thingy.
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.
