In theory any coronavirus could, including the 4 we consider absolutely harmless.
In fact, with existing attenuated virus approaches that eventually becomes the problem, the "last mile" of this approach usually sees you needing a real vaccine to cap off the disease. Otherwise the virus kinda circulates around in a tight enough cycle that you see most if not all eventual mutations including dangerous ones that kill. Communities using attenuated virus vaccination approaches for Polio faced this in recent times, for example. And it becomes a problem in places where you can't get people to afford or accept the real vaccine.
Anyway it's a tried/tested approach to solving a pathogen problem, so it's hardly some crazy idea this person has. They do point out the there will be specific types of mutations that are less likely to mutate back if you read the whole article.
The problem with polio back mutating is limited to one of the three strain in the live vaccine and is due to the mutations not being deletion mutations. The live polio vaccine was developed before people knew this and for regulatory reasons the problem has never been fixed.
I have no reason to believe you're wrong, I just wanted to point out that we (as a whole community of scientists etc.) can observe an entire range and lifecycle of this approach and that we're aware of the risks based not only on theory but on practice.
Personally I see your proposal as a relatively worthy endeavor to consider.
In fact, with existing attenuated virus approaches that eventually becomes the problem, the "last mile" of this approach usually sees you needing a real vaccine to cap off the disease. Otherwise the virus kinda circulates around in a tight enough cycle that you see most if not all eventual mutations including dangerous ones that kill. Communities using attenuated virus vaccination approaches for Polio faced this in recent times, for example. And it becomes a problem in places where you can't get people to afford or accept the real vaccine.
Anyway it's a tried/tested approach to solving a pathogen problem, so it's hardly some crazy idea this person has. They do point out the there will be specific types of mutations that are less likely to mutate back if you read the whole article.