Yes I agree it is useful to consider the rest of the world. However that was not the cause for the discrepancy and the numbers are still out by a factor of 100. If you think that 1% of global surplus attention could annually create 10000 wikipedias instead of 100 then you are likely to reach some very different conclusions. Maybe he switched context from 1% to 100% - that would make sense.
You're completely right. To convince myself:
trillion = 10**12
wikipedia = 10**8
surplus = trillion
applied_surplus = surplus * 0.01
print applied_surplus / wikipedia
which leaves us with only 100 wikipedias, as your numbers claimed.
He switched context from discussing the US-only consumption to total global television consumption. It seems worth thinking about the implications of this change even if the numbers are inaccurate.