False assumption. Planes fly in response to aggregate passenger demand. Changing commuting habits for the sake of the environment would lead to fewer flights.

This is not true. Planes fly for many reasons. Load Factor is just one of them. Maintaining slots at airports is another. Moving cargo is a third. Staging aircraft for other, more profitable routes as well.

I know it's counter intuitive, but once a route is established, flying on it reduces the co2 per passenger.

Don't take it from me, take it from the carbon offset people:

http://www.co2offsetresearch.org/aviation/Occupancy.html

Nope. You are simply ignoring the role of passenger demand in creating more flights. Additionally, CO2 per passenger is not the only consideration for sustainability; the goal is minimizing total CO2 output, i.e. total number of passengers.

Once you've told this to each of the 250 guys who do the same, er...

If we're going to be insufferably pedantic here, the plane burned more fuel because he was on it, because it had to fly at a marginally higher angle of attack because of the extra weight.

Nope.

http://www.co2offsetresearch.org/aviation/Occupancy.html

"uses the following example to illustrate this: with an occupancy rate of 51%, the fuel burned per passenger-km is 0.176 lb. In comparison, a load factor of 100% corresponds to fuel burned per passenger-km of only 0.112 lb (RCEP, 2003)."

.176 is not double .112, it's only 57% more, so it seems that doubling occupancy results in a 27% increase in total fuel usage ( (2 * .112) / .176).

That's much better than a linear increase, but it's hardly zero. Am I doing the math correctly? (Ignoring the 50% vs. 51% difference).

"less fuel per passenger", the claim made in your link, is not the same as "less fuel", the claim you're refuting.