In contrast, Einstein on (special) relativity is excellent. I haven't revisited them since my undergrad days, but I remember being surprised by just how clear and direct the original papers were. Much easier to understand than many modern attempts! The terminology is a bit outdated in places (such as, if I'm remembering right, what you call the mass term γ m₀), but that's not hard to deal with.

Yea, Einstein’s papers and a lot of those early 20th century physics papers are surprisingly quite readable, even for a moderate layman such as myself. Because of this, I must say I find Carroll’s claim that he’s never read one of Einstein’s relativity papers a bit dubious.

Dirac’s paper on magnetic monopoles contains a rather beautiful and general introduction on progress in physics. I highly recommend it.

http://users.physik.fu-berlin.de/~kleinert/files/dirac1931.p...

This makes me wonder how important or useful rigorous / axiomatic arguments for intuitively true “facts” are. As far as applications are concerned, it seems like the physicist’s math (what Feynman calls Babylonian math as opposed to the axiomatic Greek math) works well enough, while rigorization often trails by decades or centuries (as in the case of calculus).