I'm still thinking about it, but two things stand out to me:<p>1. Self-directed learning requires intrinsic motivation. Seems obvious, but I think it's really easy to lose that once you have extrinsic motivation (like a salary, or kids, making rent, a mortgage, etc). In these cases it may still be there, but redirected into things like hobbies, parenting, health, financial management, etc.<p>2. Self-directed learning is far more likely in a system where there is no (accessible) "existing pipeline" for externally-directed learning.<p>This was kind of how it was for us 20 years ago - first year university was the natural age you'd be exposed to programming, unless you did it yourself earlier than that because you wanted to, and if you did, there was no easy playbook, you had to RTFM, if there was one, experiment, fail, etc. And "computing" was niche enough that librarians or other non-academic figures couldn't really even tell you what you might want to consider.<p>These days, coding schools accessible to anyone are not even a dime a dozen, they're free. While I think this is a Good Thing™, it also falls into the trap described by the author where you're no longer creating an environment where students have to learn for themselves, you're creating an environment where they're taught.<p>Again, I think more teaching is a Good Thing™, but the kind of problem solving the author seems to be describing is not actually related to any of the learned content, whatever that content may be.<p>It's not about learning how to consume a resource, it's learning how to be resourceful.