As a manager, a big part of our job is to mitigate risk and single points of failure. If enabling someone to work like this increases risk and creates SPOF in our teams, we're screwing up letting them do it.
No, your job as a manager is to understand the trade-offs of taking such risks and make a decision that fits your situation.
I.e. your competitor may capture your market with a single engaged risky engineer before you can produce anything usable with your best practice low risk team. Once they capture the market, they can stabilize later and hire your team, because you went out of business.
More often then not, innovation comes from a reduced number of people doing things differently, and it's very hard to assemble a team of engineers which is willing or even capable of doing this sort of work.
This also comes back to the "decision shielding" from working different hours: it's very common for managers or people averse to risks to try to stick to solutions they know, even if they are suboptimal or do not fit the problem very nicely.
Anyone capable of doing more will soon realize they are spending most of their time solving mismatches between what solution was used/implemented and the problem they were trying to solve. If they don't find an alternative or a way around these limitations, they will leave.
Manager may have some challenging performances lined up and the creative operator has to really toe the line, but the purpose of the team is to support that to the best of their ability.
Anything less, and the desired expression of talent might be dampened. If this is too severe expect the star to move away toward a more functional arrangement.