Here's a possible way to reframe the problem. There is a range of smart persons with different personality types, and there is a range of careers for them to pi

Here's a possible way to reframe the problem. There is a range of smart persons with different personality types, and there is a range of careers for them to pick from. To grossly simplify, let's say you have the following factors that define a career and a person's interests: Money (M) Social impact (S) Intrinsic Motivation (I) and Difficulty (D). These factors are defined in relative terms based on the opportunity cost for the smart person. Ideally, we want to maximize the number of such persons working in high S careers. The rest is just incidental from a societal perspective. Our problem is that high S careers also tend to be lower in M and higher in D, because they involve painstaking long-term research where profit occurs much later. On the other hand, they also tend to be higher in I but it's not guaranteed. Let's assume that everyone wants low D, since it will logically be the case for high M and for high S and high I the D will not be seen as D in a sense. If we start from the basis that we should increase governmental investment in the sciences anyway, we don't have to worry about people who seek high S and/or high I. Either we'll be able to meet their correspondingly low-to-medium M requirement or they will naturally seek out the 'right' career without any input. We can forget about high I low S because we can't really offer them something that will motivate a career change towards the topics we want. The problem with the middle ground is that in a sophisticated capitalistic economy, the opportunity cost for M can rise to incredible heights. The alternatives open to a smart person in our world are enormously high-M, and the definition of a decent living has to take this into account. For the government to convert a medium or high M, low S, low I person into a high S high D career, you'd better add enormous amounts of M over the long period of time needed for research, which is not something that the public sector can justify even if they do technically have the resources for it somewhere. The I can help balance this but working in low-S sectors can also have plenty of competing I on its own, and I is not guaranteed in the high-S world to begin with. We can start thinking about hypothetical ad-less FAANG-less financebro-less worlds to try to get rid of that M opportunity cost, but it is a fairly contrived exercise and doesn't guarantee that other very high M careers might not pop up in their absence.