I get that, but hear me out. I think what's fascinating is not necessarily the rounding error on your overall energy expenditure, but the likely sheer overall cost of that on the whole. I think similar thoughts about bitcoin mining -- the truth us, if you add up a lot of this computation in aggregate, how much of it is wasteful as a whole to the organization or network paying for it? That number, while maybe not huge enough for many orgs to do anything about it, is still pretty large, and certainly interesting for me to think about.

The problem of web bloat isn't that you waste a little more electricity on a particular site once a week. It's that all the million users waste all that energy each - which multiplied by millions of users, and by most popular sites being bloated, adds up to possibly relevant amounts of electricity (and carbon emissions).