In most engineering fields we don't give the monopoly to people until they have actually demonstrated success beyond a reasonable doubt. There will always be groups of people claiming that the math/methods they happen to know are the best at explaining some behavior (even now there is that learning mechanics paper on top page)

The takeaway is to stop pretending that we can do good science when the ambiguity is so high, the majority of funding should go to people working on more concrete problems. We never locked in on vacuum tubes because the downsides were so obvious and the upsides of silicon transistors (if they could be made to work) were also obvious even to people outside the field, where your talent comes from. At the very least funders can't allow shifting goalposts, make them up front answer questions about the drugs. That will give you something to estimate the value of the drug and then when they come back with study after study outside the ranges they gave, you lower their funding. E.g. This is supposed to work on someone who was stage 2 and stop progression and then 5 years later it only "works" for stage 1 patients.

Strange breakthrough ideas can't even exist in the current system structurally, so going this route is the only logical choice. Which begs the question, why aren't clinical trials a private venture already? Governments are burning billions of taxpayer dollars for either nothing or cynically to keep the boomers alive and voting even longer, while 1/5 children are obese. For the rest of us we've socialized the risk and privatized the profits.