That’s no mistake, folks are suggesting that on a per transaction basis - today - Bitcoin is cheaper or more efficient if you take into account hand wavey externalities not captured in the classical transaction cost model of a payment processor like visa.
What I’ve done is invalidate that argument not by scaling visa down to a single transaction (as that doesn’t capture externalities, per the argument posited) but scale BTC up as the argument is BTC captures all externalities. By then showing that it’s physically impossible I’ve disproved the thesis.
You’re making a different argument about avenues for scaling and that deserves its own conversation.
Good point, you're right, I was forgetting this.
> since the value of each unit would have to increase to satisfy the demand.
I didn't follow this though, what do you mean?
> In a world where Bitcoin replaces gold or visa
It doesn't seem very meaningful to me to talk about energy usage in this world because Bitcoin can't replace visa. To replace visa we would have to make enough changes to it that I don't have much of an idea of what the energy usage of the modified system will look like.
Though let's consider Nervos' NC-Max, which gets a lot closer to replacing visa than an unmodified Bitcoin can. It works by, effectively, massively increasing the block size limit. In a NC-Max world energy usage would still be proportional to the coinbase but the coinbase might stay small because there's so much room in each block that transaction fees stay low.
I should have worded it better, I just mean that if Bitcoin were to replace gold as the primary store of value, the price of Bitcoin would have to rise significantly to absorb the demand, since supply is inelastic. This in turn would increase the environmental externalities.
Since transaction inclusion is a marketplace and since most products we care about will continue to be denominated in local currencies I don't think transaction fees will change much as the price of bitcoin changes.
As the dollar-denominated value of a 0.01btc fee increases people will become less willing to spend that 0.01btc. There might be some stickiness but I expect the kwh-denominated fees will grow much more slowly than the price of bitcoin.
The first is that scaling Bitcoin up to the size of visa is not particularly possible. You might be able to match visa's throughput by greatly increasing the block interval (e.x. one block every week) but the protocol cannot reasonably scale past a couple hundred transactions per second without taking a massive hit to transaction latency. So, in this sense the question of how much energy Bitcoin would consume if you scaled it up is meaningless, it cannot be scaled up.
The second mistake is that Bitcoin's energy consumption has no relationship with its scale. Say you built a hypothetical MiniBitcoin which was exactly the same except each block was so small they could fit just one transaction, MiniBitcoin would consume almost exactly the same amount of energy as Bitcoin does now. The energy usage goes up as competition between miners increases. The energy usage does not go up as more transactions are processed.