I am not so sure. Since the bandwidth of the amp is very low.
The bandwidth of LM318 isn't low. It's about 10 MHz at low supply voltage.
Biasing of the discussed circuit (from post #4) is basically good in my view. The scheme can be often found in single supply operation.
The voltage divider sets the +ve input to Vb/2, 100k is sufficient low resistance for the about 150 nA input current of LM318. The DC feedback factor is unity, in other words the circuit works as a voltage follower for DC, so within it's voltage limits, the OP will "copy" Vb/2 to the output, achieving maximum AC voltage swing.
For AC, the circuit uses the negative supply as it's ground node. Source and load should be also referenced to this node. Except for very low frequencies, where the capacitor values get unhandy, individual decoupling of each amplifier isn't bad. A common "virtual ground" brings up a risk of unwanted coupling between amplifier stages.
If you use e.g. 100 nF instead of 100 pF, you get a corner frequency of about 1.6 kHz for the feedback network, which is O.K. in many cases, for an audio amplifier, you would put in 10 uF.