It wasn't a change to the circuit, it was a change to the theoretical model of the circuit. Originally, before we had built one, the theory said we'd need those resistors, After we built it and listened to it, it became apparent they were not actually necessary. After some experimentation and measurements, I realized the theory was too simplified - a more refined theory agreed with the real-world results (both subjective and objective) that the resistors are not needed.
As I just posted in another thread, if the SEX amp is not loud enough with low-impedance phones, the 120 ohm resistor at the headphone jack can be reduced or replaced with wire. In some cases, even sensitive phones will benefit from the better damping, unless they are so sensitive that the noise floor becomes a problem. A couple of clip-leads to short the resistor makes an easy experiment to determine if this is the case.
