> Bitcoin is fairly unique in its energy use. Almost every other use of resources will try to be more efficient if it can be.

It's not like energy use in the world has gone down. If it's cheaper or more efficient it will go up or at best stay the same. Bitcoin has made the transition to more efficient ASICs because that increases profits, it just didn't result in less energy use because now you make more profit.

If a car can be more efficiently produced, the energy spent producing cars will go up because either more cars will be produced or more features will be added to the car. This is a win for the car consumer. This is very different from Bitcoin where any improvements to the efficiency of mining it (such as transitioning to ASICs) does not lead to a tangible benefit for the Bitcoin user. All major non-ASIC miners will eventually be out-competed out, but the amount of energy wasted will always be correlated to the price of Bitcoin.

While not a tangible benefit for them, they do get a more secure network as a result.

I don't think they do if you're just referring to how difficult it is to pull off a 51% attack. Assume the status quo is that it costs X to secure 51% of the hash power. If there is a new ASIC that can hash twice as fast with the same energy use, then suddenly it will cost 0.5X to secure 51% of the hash power. But if the price of Bitcoin remains the same, then miners will have an incentive to double their hash power as that was their breakeven point before. And so now you've gone back to the previous status quo. Assuming the attackers have access to the new ASICs, it still costs X to 51% attack the network.

I'm not sure I understand your reasoning regarding the price of bitcoin staying the same and the miners having an incentive to double their hashpower. In the situation you described, you'd have the same hashing power but half the energy cost? Doubling it would double their hashing power and keep the energy cost the same. I don't see how this doesn't lead to a more secure network?

I don't think you can neglect the price of a new ASIC in a real situation though, those things are very expensive. Where it was first possible to attack the network using regular hardware, it now requires a much bigger capital investment. When ASICs were first released they had a bigger advantage than they have now, it seems unlikely a 2x improvement over the previous generation is still possible (assuming it was possible before, I didn't check).

It doesn't lead to a more secure network because the costs to attack the network are the same. Attackers will have access to the same technology and have similar costs.

Right, I'm not saying there will be a new ASIC; just that there will be no innovation in the Bitcoin space (unless they switch out of PoW) that leads to less energy being used.

I agree with that, it seems very unlikely we'll go to a less energy situation unless a price decline happens. At least until all bitcoins have been dug out.

And you'd end up building more trucks because of Jevon's paradox: https://en.wikipedia.org/wiki/Jevons_paradox.

Bitcoin mining does respond to market pressure, if the price goes down, or you wait until the subsidy halves, there will be less incentive to mine.