One small point I'd like to make regarding USB mics (and USB headsets): the #1 rule of digital audio is to only have a single sample clock for the entire system (the sample clock is the thing that defines when a voltage value gets converted to/generated from a digital sample).

Although I understand the plug-n-play appeal of these devices, they uniformly violate this rule because they cannot share whatever sample clock is being used for the other "direction" (i.e. the USB mic has its own clock, which is independent from the one used for playback).

Using these sorts of devices implicitly requires the audio software stack to do resampling to keep both input and output in sync with each other. For podcasting the change in quality that this causes is probably not an issue, but users should at least be aware this is happening under the hood.

You're much better off getting a cheap USB audio interface, an XLR mic and doing both recording and playback via the same device (with a single sample clock).