To me, the whole BCacheFS thing feels like ReiserFS all over again. Including the borderline insane, self-proclaimed genius lead developer with an out-of-control ego, the massive over-presentation of how important this filesystem, or even filesystems in general are to Linux as a whole, the complete refusal to work within established structures, both in terms of process and in terms architectural structures in the software or to even have a mature discussion about how these ought to be like, and even the ludicrous claims about how Linux will be hopelessly outcompeted if it doesn't put his genius front-and-center before anything else.
At least Overstreet has not murdered anybody.
There's also the whole selective perception about software stability. You cannot claim at one point that your fs is completely ready for production use, everyone saying otherwise is a hater, and marking the thing as experimental in the config is basically slander against your person, and then shortly later demand that Linux merge some particularly complex and hard to review several-thousand lines patch in a minor bugfix release because your users could experience horrible data loss otherwise. Those are two things that cannot be true at the same time.
Note, filesystems really are hugely important to an OS like Linux. Even so, both Reiser and Overstreet managed to overstate that to an outright comical degree.