This impedance matching issue usually comes up when there's not enough silicon (or tubes) in a circuit. Maybe that is the whole point if this part of an academic assignment.
In the commercial world, IM is only critical for things like power transmission systems and RF amplifiers, where transformers / RF transistors are expensive, and costs need to be kept down, such as an amplifier for CB Radio (burner), the performance is highly dependent on IM, since it only has a single RF transistor. OTOH, an audio amplifier IC itself has dozens of junctions and is impedance full of mis-matches all over the place! Then the outside the IC, output ohmage of an amp driving an 8ohm speeker is milliohms.
I can see a problem on the circuit (if still there), delete the 215K and 10 uf. & connect Q6 base straght to Q1 coll.
It's desireble to have an impedance mismatch going on between the two stages. A emit-follower's input is often a lot higher than the sourceZ it's connected to. Just like the output ohmage of a good amplifier is much lower than the '8 ohm' speaker.
Finally, if Q6 really is going to be a class-A output stage driving speaker load worth just 1.6 Ohms, then you'd need to add another transistor to boost the current. Make it a darlington or alpha pair or whatever it's called. Then you'll need to reduce the value of R16, you could make it a light bulb and the class-A amplifier project can double as as light and heat too.
Unless the text of the original assignment is posted here in it's original context, it's likely newsgroup posters misunderstand and all go off on different tangets!