> Which isn't to say it is 100% safe for other reasons. I'd say auto-immune reactions are the biggest risk (e.g. the body using the signature encoded to attack the body itself).
Out of curiousity, why couldn't potential auto-immune response be simulated on a computer up to a degree of confidence, if nothing? Would quantum computers drastically rid of current computing related handicaps wrt to simulating a large search space typical of such simulations? What role might yet the recent resurgence in deep-learning play for immediate future?
"Out of curiousity, why couldn't potential auto-immune response be simulated on a computer up to a degree of confidence, if nothing?"
Basically every answer you could imagine to that question is true. We don't understand the system well enough to simulate it at that level of detail. We don't have the computational power to do it if we did. If we did have aliens from space deliver to us both the power and the simulation, we'd have a difficult time verifying the validity of the simulation!
Quantum computing is still not well-understood enough to know if it's helpful; there's a certain barrier of "what can it actually do" that we won't really understand until we have them, and we start raising up quantum programmers. No slam against the theorists intended, it's just the difference between "the halting problem is unsolvable" and "here's a program for calculating payroll" is small in theory but large in practice. We won't really know what they can do until we work with them for a while.
I work in computational drug discovery. With current hardware, we are basically limited to simulating single proteins for microseconds of time, and even that requires a lot of simplifying assumptions. So we can only simulate a small part of any given process. We can't develop high confidence about a complex cascading process this way.
Quantum computing would definitely help if we had it. And researchers are using deep learning in many parts of the process now.
Out of curiousity, why couldn't potential auto-immune response be simulated on a computer up to a degree of confidence, if nothing? Would quantum computers drastically rid of current computing related handicaps wrt to simulating a large search space typical of such simulations? What role might yet the recent resurgence in deep-learning play for immediate future?