Basically, you become a pattern matching database of non-obvious cost/benefit trade-offs.
What some might call "experience." What makes problem solving skills valuable is the ability for the wielder to pattern match effectively. The more problem patterns you can positively match to solution patterns, the more valuable your experience. To a lesser extent, simply being able to identify problem patterns is also very valuable, as if you can avoid the creation of a problem it doesn't matter if you know a solution to it (I don't know how to escape a black hole, so I generally avoid them when traveling in space).
Additionally, if you're young and naive, you'll be likely taken advantage of. Why else naivete might be desirable in a business setting?
I see this type of thinking applied to an issue in a way that results in people becoming increasingly irritated. Silicon Valley wants to hear "yes yes yes", not reason why you think XYZ solution won't work.
The crucial transition is turning yourself into a senior engineer - someone who acts as a force multiplier in your organization (a possibly big organization).
The valley problem with that transition is that a lot of value is left to the junior developer due to the startup ecosystem - the engineer who writes the first prototype, shoddily for a demo without worrying about possible future problems.
The senior developer is left out in a plateau, if they can't switch back to core developer mode (build demo, get funding, hustle) and back, which is a not a skill issue, but more economical.
There's the risk that the switch to a lower pay (for equity, hopefully) affects your next job offer in a BigCo if the idea does not pan out (& most of them don't).
The reality of being hired in a Startup is that founders don't like people who aren't "all in".
Once you take out "being employee #5" at the startup from the equation, the best part of Bay Area is gone.
There lies only one path back in and that's building your own, except there the VCs get to actually pick on your health as a possible risk factor before handing you some $$$.
Software tech is the only area that I know of that doesn't care what you've done in the past or have experienced. All that matters is what stupid algorithm you can figure out on a white board in the language they use.
As I grow older, I notice these changes, and while I do regret not being able to remember IP addresses after switching to a different window (get a larger monitor, or just copy&paste), I am very happy with the overall shift.