Coordinated hard-forks have been a well established upgrade protocol of other cryptocurrencies such as Monero. When they have not been coordinated, such as was the case of Monero's younger sibling Wownero, the network effect of these forks is not worth considering. Hard forks of existing cryptocurrencies are mostly miserable, as in the case of BCH, or irrelevant, as in the case of Wownero. Other than the risk of network split, a nice problem to have that can be solved easily with rough consensus and running code, are there any downsides to hardforks? Let me know in the comments. Their critics say that well-coordinated hard forks require a more centralized coordination than soft forks, the latter method being better suited to decentralized systems. Even if this were true, that Bitcoin's developing ecosystem is so decentralized that it cannot well coordinate to hardfork the software, then Bitcoin can just have poorly coordinated hard forks which may turn out to be way better for the software than upgrading it with a soft fork.

Now a few words on soft forks. A soft fork involves adding more code to account for new rules. Soft forks tend to make the software more complex and their downsides of the forks accumulate with time, making it perhaps unsustainable to upgrade in the future. The fact that the developers have to keep the changes compatible with the old software means that they have to make more technical workarounds, and this makes the software more complex than if they did changes that would be incompatible with old nodes. To this technical debt add the relaxation of validation - as non-upgraded clients see transactions as valid and the problem of network-split risk pales in comparison to those disadvantages of soft forking. Seriously, it seems like a good problem to have, a problem that time will cure anyway.

I find it hypocritical to hear from many Bitcoiners who are advocates of soft-forks when they complain of high-time preference lifestyle. If they are so low-time preference why do they worry so much about the network split, a non-existing problem in the real world. A problem only for those who speculate on BCH, and even for them it may not be a problem at all. You might as well worry about alleged Bitcoin's unsustainability, instead of worrying about the unsustainable development of the code base, in the face of global warming and what not.

P.S. I've been running a full Bitcoin Lightning node and a Monero node for the last five years. I have had many more serious problems related to that then the risk of the network splits.