OK, so Adam has switched to using 32 kiloHertz for his Daily Source Code. Hope this pushes more podcasters to do so because in my opinion 32 kHz is an ideal choice; It gives us a range better than that of FM-stereo (the EBU standard anyway).
First, just know about this. Basically it says that if you want to record a frequency range including 20 Hz to 16000 Hz tones, you’ll need a samplerate of 32000 Hz. It is actually quite simple if you think about how a waveform looks, it has a top, and a bottom side, and you’ll need two points of reference to interpret them, so double the highest frequency you want to hear.

Just a little advice to those who want to use 32 kHz for their podcasts; It is best to also record at the samplerate you want to deliver, so you won’t have to resample. And if you really need to, be sure to use a good resampling algorithm, with anti-aliasing. The one in the LAME mp3 encoder is actually really high quality, so it’s safe to sample down from 44.1 kHz, for example, using that encoder.

By the way; Always wanted to know how MP3 really works? Look here