First you grow talent the same way it grew in the bay area, by letting people work on challenging software products decade over decade.
Second, population growth leads to strains on civic infrastructure which in the bay area is way under invested in. Additionally, there is a very limited amount of infil available to build new housing on, so it would require that other housing get torn down in order to build said new housing. In areas with small lots, this is harder to do because you need to buy up a bunch of surrounding properties to build a larger building.
This often contributes to increased displacement as developers are buying up properties, using laws such as the ellis act to evict and redevelop properties. Oddly enough, the restricted zoning which keeps prices high, also reduces the amount of redevelopment that happens which keeps land values lower and maintains rent control that would not be allowed for new buildings under costa hawkins.
Additionally, the political infeasibility of San Francisco should not be overlooked. This is a city with a long history of anti-capitalist activity and capitalists should recognize this fact and go to places that will be more accepting of their investments.
I don't know how to reliably "grow talent", and I'm not sure most companies would either. Building housing and attracting talent are solved problems, so why not try that first?
I'm not saying companies shouldn't be in Austin or Huston, far from it. However, SF and NYC are titanic economic engines that could be even more dynamic if they had more housing. There is an obvious solution to the not enough housing problem, difficult though it may be, it's probably worth solving.