> What's needed is a fast-living animal, that ages, and gets cancer, in the same way humans do. Then you can just experiment without understanding.
A simulated human?
A simulated human?
And after listening to one sleep-inducing waterfall video for a single long night of sleeping the next day about 2/3 of my Youtube homepage's suggestions are about soothing, meditative, nature videos... what a laugh.
(I know that's A Lot of computing power, but maybe an algorithm like HashLife could make this possible by, say, 2100?)
For the chemical simulation side of things, it would take many centuries of continued Moore's Law progress to make today's most accurate algorithms for fully quantum chemistry run fast enough. And the wild ride of Moore's Law already appears to be closer to its end than its beginning. We'd need fundamental algorithmic breakthroughs to make it feasible without relying on better simulation hardware solving the problem. That might be possible, but I don't when/how/if such algorithms may appear.
On the problem of getting a ground-truth molecular map of a cell in the first place, we know that there are many gaps in current models. We're not sure where all of them are. Even if you had the fast, efficient, vast molecular simulation capacity I sketched above, it would still take a lot of experiment/simulation comparisons to be sure that all the biologically relevant features had been captured in the original model.
Even now I'm probably badly understating all the problems on the biology side. I worked at the interface of chemistry, biology, and chemical physics, but I was more familiar with the software/physics side of things. I learned just about enough biology to be useful in my role and to be grateful that even my most complicated debugging sessions didn't involve 3 billion years' worth of undocumented legacy code running on wet nanotechnology.