Because – and we can’t say this too many times – biology is not mathematics or electronics.
There are more than 20,000 protein-producing genes in the human body, many of which produce multiple proteins. There are several times more whose function is to regulate those proteins. Those aren’t just necessarily A affects B interactions, either. Protein A could affect dozens, or hundreds, or thousands of operations in the body.
Out of all those millions of actions, reactions and interactions, scientists must find a) which ones are bad, b) something that actually turns the bad off, c) without damaging all the good actions, reactions, and interactions, d) that works consistently enough to be considered “safe and effective,” and e) can actually be produced in the real world.
And then, you have to test and retest the compound to make sure it’s actually having an effect, not just a coincidence. You have to test it over time to make sure the body doesn’t rebel against it. And you have to test it in different populations to make sure its effects are predictable in men or women, old or young people, and so on.
Certainly computers and models can help. Computers are very good at going through huge amounts of data and uncovering trends that would take humans forever (if they noticed at all.)
But the process is completely different from developing a faster computer chip, so thinking that having a computer that’s 10 times faster will lead to discoveries coming 10 times faster is simply wrong.