Let's do some figgerin'...
The average healthy but sedentary, college-aged male has a VO2max of approximately 45 mL/min/kg. However, I have seen it argued based on studies of, e.g., aboriginal tribes (and there are population data from Europe as well as military inductees here in the US to suppor the conclusion) that the "default" VO2max of the average human male is closer to 50 mL/min/kg, and the only way to get below this is to assume a couch-potato lifestyle, gain excess weight, etc. (and/or grow old, of course). So, I'll go with that latter number.
With short-term training, VO2max increases by 15-25% on average, with another perhaps 5-10% possible (on average, anyway) with more prolonged and/or intense training. That gives a total increase of 20-35%, so I'll go with 30% just for argument's sake.
So, if VO2max is, on average, 50 mL/min/kg and increases by, on average, 30%, that means that the average Joe ought to be able to raise their VO2max to about 65 mL/min/kg with training. Indeed, there are many, many, many, MANY amateur endurance athletes with VO2max values of around that number (not to mention the fact that athletes in team sports with an endurance component - e.g., soccer - often have a VO2max of around 60 mL/min/kg, something that is also true in other sports that you don't normally consider to be of an endurance nature, e.g., downhill skiing or motocross - i.e., motorcycle - racing).
The question then becomes, how high might functional threshold power fall as a percentage of VO2max (again, on average), and what does this translate to in terms of a power output? The answer to the former is about 80% (LT, on average, being about 75% of VO2max in trained cyclists), which means that in terms of O2 consumption, a functional threshold power corresponding to a VO2 of 65 mL/min/kg * 0.80 = 52 mL/min/kg could be considered average. If you then assume an average cycling economy of 0.075 W/min/kg per mL/min/kg, this equates to...
3.9 W/kg
