But this actually matters! While your example makes it sound indeed silly, full "life-cycle assessments" and counting externalities matters. Permit me to take your example to an even more farcical end: imagine there was a correlation between bikers eating beef and bus commuter being vegetarians - in such a silly world the bus commuters would probably be much greener even though the bus is a gas guzzler. The problem is that with nuclear these analyses are really complicated, and should not be dismissed out of hand. I do believe that nuclear energy is very important in the fight against global warming, but such dismissals do not help convince people of it.
A more real world comparison is an electric bike commute. If the extra power for a cycle commute comes from human food then it's entirely possible that the extra carbon cost of making and charging the battery are a worthwhile investment to help save carbon footprint even though the bike is now heavier.
As you say, vegan vs carnivore (and red vs other meats) would be a factor just as whether the electric power comes from renewable or fossil power.
And to get super complicated you could then weigh the cost of healthcare if the rider is using less physical exertion.
You're not wrong, but mind the point from manfredo about double-counting, too. I.e., the bicycle is not _itself_ a source of CO2 as the bus is, so if you cownt the beefeaters' full carbon hoofprint, you have to subtract that elsewhere in the overall assessment.
You say that would be silly, but there are undoubtedly correlations between diet and overall athleticism. Whether beef-eating is positively or negatively correlated with bicycle commuting is difficult to guess, but it would be surprising if there were no correlation.
