I've complained a lot about various obstacles in my slog through the academic career -- lots of "service work" not directly related to publishing papers, unsupportive advisors, generally being demoralized. I haven't discussed my own general ignorance very much, but increasingly I realize what a huge issue this is in my career at this point. While I learned plenty about Feynman diagrams and Einstein's equations in graduate school classwork, and did my share of solving really tough differential equations and worrying about factors of pi, I learned virtually nothing about useful physics for astronomy-related papers: atomic physics, fluid dynamics, general astrophysical situations.
It's been hit home repeatedly to me recently what an ENORMOUS literature is out there, and how it's been FOREVER (or never) since I learned the physics necessary to make sense of it. Thus I find myself reinventing wheels all the time, only later to find out how what I just derived or suspected was referenced in some paper written 10 years ago. Had I done my graduate degree in astronomy, or worked closely with an astronomer for the last 5 years or so, I might know more of this stuff than I do currently. I would estimate my knowledge of the physics involved to be somewhere around that of a second-year graduate student... and meanwhile I'm applying for faculty jobs.
For example, in some work I've been doing lately on clumping of gas in circumstellar interactions in type Ia supernovae, I introduced a parameter called "chi" to describe the density contrast of the clumps with the surrounding wind. It just seemed like the thing to do. Just now I found a paper by
Wang & Chevalier (2001) which not only introduces the same parameter, not only with the same physics impact, but even uses the same bloody Greek letter.
It's annoying to spend a week working on something and then discovering in 5 minutes how what I was doing was totally subsumed in an investigation already out there. But it's equally annoying to spend weeks and weeks just combing through enormous piles of papers without getting the traction necessary to move forward. On the last paper I worked on, the phase transition from "I don't know anything" to "I know as much as anyone else publishing in this field" was remarkably sharp, and I still wasn't convinced I was there... it's almost impossible to know without being completely up to date and trawling the preprint archives and rumor mills daily.