I remain opposed for the reason I gave last time this was proposed: The IETF should retain control of IPv4 and any statement to the effect that the IETF will no longer work on IPv4 will inevitably lead to formation of an IPv4 legacy standards group in competition with IETF.
Like it or not, FORTRAN and COBOL are still in common use a full 40 years after they were functionally obsolete. I see no reason to believe that anyone will need more than 32 bits of addressing for their home network. There being no compelling reason for my coffee pot to be able to talk to the entire Internet, I have a compelling reason to prevent it doing so.
Rather than sunset IPv4, I would sunset IPv4 as an Internet protocol and relegate it to use as a network protocol only.
As a network protocol, IPv4 remains far superior to many other protocols that the IETF continues to support and should begin supporting. I do not believe in a model where every device connects directly to the Internet. There is a proper role for the local loop and RS485, SPI, IC2 etc. devices in IoT and there is a role for IPv4.