Hard problems are hard. Easy problems are easy. In a world where autonomy can only be purchased by wealth, any rational actor will spend his/her time optimizing the acquisition of wealth by solving easy problems as quickly as possible. This will then allow him/her the necessary autonomy to tackle hard problems.
This applies to all capital intensive endeavors of creation: science, film, software, hardware, etc. Even the least capital intensive endeavors such as pure mathematics require autonomy in the form of a guarantee of food, shelter, and healthcare to be viable in the long term.
As such, the solution is simple: Increase autonomy. But do we increase autonomy for everyone, or pick and choose winners?
The first strategy is basic income, and the second strategy is venture capital. Those involved in venture capital are now seeing the weaknesses in their own system, wherein picking and choosing winners will skew heavily towards the best signalers (i.e. marketers), rather than the most qualified/talented.
If VCs want to hedge their bets, they can put their lobbying money where their mouths are, and support basic income. This allows savants who are unable to market themselves the required freedom to work on their ideas with a tiny but guaranteed degree of autonomy. It is the equivalent of buying a Vangaurd index fund instead of picking and choosing stocks for a portfolio.
The winning strategy is to do both simultaneously.