This makes the witness commitment useless for its intended purpose - a flexible (future) merkle root committing to additional commitments. The nLockTime field is a much neater way of doing this, and I don’t really see why “miners have to slightly tweak their software” is a huge deal as long as they get a year or two of lead time to do it. We could even set this part of the fork to activate on some substantial delay - after the GCCR fork activates, the nLockTime requirement only turns on two years later or whatever.