> Let's just say there is consensus over what the Bitcoin blockchain is, versus whatever the "Bitcoin Cash" or "Bitcoin ABC" blockchain is.
The consensus is the chain which has the most accumulated work. And also, the one which is backward compatible with the one which previously had the most accumulated work. A backward incompatible fork is a shitcoin.
> Users can't ignore the interest of miners, even if miners are a tiny minority. Miners may not even have any holdings.
You have this backwards. It is miners who cannot ignore the interest of the users. Miners can only profit by selling the bitcoin that people want to buy, and the bitcoin people want to buy is the one which is inflation-free. The market decides the correct chain and the miners follow it.
Miners can't orchestrate a fork because they need consent from the users, and they won't get it. If a majority attempted to mine a forked chain, the minority chain would just become unfairly cheap to mine (due to difficulty reduction), and cause more people to mine it again. In the mean time, the miners who forked away would be making nothing, as they have no market to sell their shitcoins into.
> A core value proposition of Bitcoin is that it's the "most secure" network. If a majority of hashing power is priced out of operation because of low block rewards and low fees, that security flies right out of the window. This situation would have a far more significant impact on price than a little bit of controlled inflation.
If bitcoin cannot be sustained by fees alone, the experiment is a failure. There is little advantage to "digital fiat" which cannot be done in other ways without the inefficient blockchain.
However, there is also the acute possibility that mining is not profitable, but still performed. The reason is simply that it is a way to recover some money from energy which would otherwise be completely wasted. Consider production flaring in the extraction of crude oil. Energy companies can recover some of the loss by deploying mobile bitcoin miners, which already exist on the market today. There is also the potential for miners to be data furnaces, whose electricity costs can be partially recovered through fees, which may be utilized anywhere that heating is required.
> It's not going to be alternative software, it's going to be a software update, through the official channels. It also likely would only affect full nodes, and the majority of Bitcoin users haven't been running full nodes for a while.
The majority of users don't need to run full nodes, but there are sufficient full nodes that they don't have to. It also happens that the people running full nodes are the bitcoin maximalists who understand bitcoin and are even less likely to accept any inflation attempts than those who are running SPV nodes.
> It's really up to a few key stakeholders, not Bitcoin users in general. When faced with the decision of giving up network security versus maybe losing a little bit of value to inflation, they will choose the former, because they're not that stupid.
If multiple software clients are running on the network, then none of them will have an absolute majority. It is questionable whether developers of Bitcoin Core will roll an auto update feature into the software, which means there will always, very likely, be a majority of users running old software. The new software can only introduce backward compatible features or it will fork the minority of users who upgrade away from the network. Auto-update is an implicit backdoor and unlikely to be installed by competent users, especially where significant amounts of money are concerned.
> A similar thing already happened with Ethereum, the rules were changed mid-game in the official client, the loser fork (Ethereum Classic) was the one that kept the old rules intact.
That's because Ethereum is a personality cult and not an experiment in sound money.
Bitcoin maximalists have developed a culture of weeding out such personalities.
It is unlikely that a hard-forked bitcoin will ever gain traction. The SegWit rollout demonstrated that it is possible to perform sophisticated upgrades without breaking backward compatibility, and this mindset is now the default for all developers left on Bitcoin. Those with the mindset that they are in charge, rather than the market, have all left.