"Bitcoin creates a global energy arbitrage market for the first time" -- There are existing energy arbitrage technologies, such as aluminum smelting. The requirement is low capital cost with high energy cost, so you can have plants that are idle most of the time, waiting for low marginal-cost energy. Bitcoin mining isn't actually very good for this: the capital costs of rapidly depreciating mining chips are quite high.
"Incentivizing cheap green energy" -- every consumer of energy does this. Consumers that run mainly during the day are much better at incentivizing wind and solar.
"Mining for ambient heating" -- also suffers from the problem that mining chips are too valuable to leave idle, so you can't place them in homes and only run them when it's cold.
In this case, the Moses Saunders dam  used to get a nice steady stream of draw (necessary to keep the equipment operational) from the Alcoa plant, but when it shut down that draw went away. Throwing a bitcoin mining operation into the old facility is a great way of keeping things running smoothly.
This makes sense only in two cases: you're using energy from power sources which can't regulate output, or you're adjusting usage immediately based on feedback from the source. Otherwise the generation side just raises, accounting for your usage. Does that actually happen anywhere?
> Energy can't be transported long distance
Of course it can be. I expect you're quite far from a power plant right now. On the extreme we've got high voltage lines going over 2000km.
I'm pretty sure miners don't pick and choose when they mine, to coincide when there's going to be 'wasted surplus energy' otherwise. They just add on to the demand on the grid, around the clock.
They almost sound like Buddhist monks re: meat eating, from the way you phrase things!