It feels like a typo in the following line :
We can see there is a rule specifically allowing process transition for anything started using the init_t type to the unconfined_t type, which confirms our original suspicion (there will be more output, but I've omitted this for clarity).
It should be corrected as
We can see there is a rule specifically allowing process transition for anything started using the init_t type to the unconfined_service_t type, which confirms our original suspicion (there will be more output, but I've omitted this for clarity).