To add to this, anyone can fork a blockchain, but not necessarily process all the transactions (a.k.a. "mining"). The forks that matter are those performed by communities or organizations with consolidated mining capacity comprising a large fraction of the total mining capacity. The fact that a fork like this became news-worthy means that a significant portion of the mining community agreed with the protocol changes and switched. Miners are typically interested in only one thing and that is processing as many transactions as possible. They will perceive efficiency bottlenecks in the Bitcoin protocol as threats to their earnings. The CPU power needed for Bitcoin mining grows exponentially with the number of coins / transactions that have been mined. ASIC hardware, low temperature climate and low power costs are prerequisites these days to run a competitive mining farm. However there are ways for speeding up the transaction processing by tweaking the Bitcoin protocol settings. Small mining communities are not in favour of this as it would allow the larger mining communities to push them out.
There are alternatives to the Bitcoin consensus mechanism. Bitcoin uses a Proof-of-Work implementation. Ethereum as well, but it is ASIC-resistant, meaning that you can still mine with a decent GPU. There are others, such as Proof-of-Capacity, Proof-of-Burn and then there is Proof-of-Stake, probably the most interesting "next-gen" consensus mechanism. Ethereum will migrate to a PoS system with the release of Casper.