There’s a big difference between applied science and pure science.
Applied science is what engineers speed running our return to the moon are doing. They are re-discovering processes and construction techniques, improving some old ideas with better new ones. But, with very few exceptions, they are not coming up with any new big ideas, ready to be proven.
There’s an existing concept in the literature called Technology Readiness Level. It goes from 0 “a fervent wish” (my extension to the scale), through observing something that might be a big idea (1), reproducing that observation, coming up with an idea about how it why it works (2), writing an idea how to prove the big idea down (3), actually proving that idea in a lab (4), proving that the idea works in the less than lab perfect conditions (5), coming up with a new idea for how to use the big idea and building a working model (6), a first run copy of the final new technology that requires a lot of handholding from designers (7), a first run copy that can work without the designers on standby to fix problems for a short run (8) and a proven model (9)
You can speed run TRL 4 through 9 until you run out of ideas (0 through 3) (pure science).
However medical “science” and physical science are very different sciences. There is a stupendous gap in academic rigor. There are certain things we don’t know how to measure accurately, or would be too expensive to measure accurately. And there are errors in every instrument we use to measure the world, which make all of the conclusions we draw from them slightly flawed.
Mathematicians gave us statistics, which allows us to look at a group of measurements and infer from that group where the real answer is. However, by using statistics, we invite the possibility that we are lying to ourselves about what the measurement really is. We are applying arbitrary rules to a set that might not apply.
Thankfully, Mathematicians also gave us techniques to assess the quality of our use of statistics on a set of measurements, which quite nicely resolves into a percentage chance that we are completely lying to ourselves about what is being observed.
Back to academic rigor: physical sciences tolerate no more than a one-in-a-million chance that you are lying to yourself (also called six sigma) in order to publish an idea.
Medical “science”, motivated by a fervent desire to provide life improving products fast, tolerates a 30% chance of your idea being fiction. And there are some publications that are even more tolerant at 50%. Medical science would use better tools if they existed (and there’s work being done on computer simulations of medicine, but discouragingly the same Mathematians who have helped before believe that modeling medicine may be too complex a problem to ever solve).
So, Medicine doesn’t have the tools at the moment for good ideas at the TRL 1 and 2 levels. It also lacks the tools to really gang up on problems at the TRL 3 through 9 levels. This is one of the reasons the usual approval process for new medicines and new techniques requires years (used to be generations) of observing those who did and didn’t take a product, because medicine lacks the tools to honestly assess long term effects. It is just hopeful guesswork.
You can spend a lot of money trying to speed run hopeful guesswork. You might even be able to deliver more products to market than real science. But they will very likely be at best ineffective, and at worst harm people.
You might want to know “why can’t I speed run TRL 1, 2, and 3”? It’s a fair question. Most (maybe close to all) initial observations at TRL 1 are the result of a happy accident : someone with the right training is in the right place at the right time to notice something unexpected. That person has enough leisure time to mess around with what they observed (because no employer lets anyone “tinker randomly” for a living). And that by happy accident the person observing something weird figures it out it the limited time he or she has to devote to it. And, by a third happy accident someone with similar skills and training reads the first observation and decides to independently check it out.
The closest we possess to speed running this is : someone guesses that (1) exists and goes looking for it. They either find what they are looking for, or prove that (1) doesn’t exist under a certain set of conditions.