Don't get me wrong--I'm self-taught too, I'd been writing code for a decade-ish before I went to college, but this stuff changed my life profoundly, and I think minimizing it out of hand is downright tragic. Maybe if there was a better way to get people to acquire this a few years on (calling to mind the idea that people should have a few years of experience before getting an MBA), but there isn't right now.
As far as paying a premium goes--I graduated with about $20K of debt, mostly because it was effectively free money (my total interest payments before I paid them off were less than $2K), and was getting paid through school (did Google Summer of Code twice, ran my own web dev shop, etc.). Bad choices can be made with regards to college, but that's a criticism of overly expensive (private) colleges.
I'm a high school dropout and have started/ran/sold a startup, managed a division for a well-regarded consulting firm, and worked on data taking for a detector at the LHC. Perhaps the problem is that I minimize the amount of effort that goes into being autodidactic (determine the problem, determine what information and skills are needed to solve the problem, acquire said information/skills, execute, repeat).
Let me correct/clarify my above post: Theory isn't a waste, per se. Is it best to spend a significant amount of time front loading knowledge one may not need? And paying a substantial premium for that experience? That's my problem with the college experience for tech professionals.
I'm not a huge fan of bootcamps, but I do believe there is much to be gained by Google's efforts here. There is value we've lost in the old apprenticeship system, and I hope to see it revived over time (I owe my skill and career progression to the luck of finding quality mentors along my way).
There are hard problems I want to solve that I want to have the resources (mostly time) to solve. College would not teach my those skills, but me trying to solve those problems are lessons in themselves.