johnnydoe2 Is there a strong reason for NOT making the optimization an opt-out feature?
In theory, like anything else, it's a simple matter of writing the code. But the code probably isn't all that simple, and then it must be tested. Should the AOT toggle default to off or on? If it's on and the user turns it off, should the already-compiled apps have the binaries deleted? What if the toggle is turned off while AOT compilation is in progress? Should the user be warned when the toggle is flipped off? Should the setting be global or per-profile? I suspect it needs to be global, but if it's deployed that way probably people will complain.
It would of course be possible to address those questions, and others, and to implement something, and test it. But how should that be prioritized compared to other things users have asked for, and things the developers know are issues that users haven't yet asked for?
So there may not be a strong reason to not implement the feature, but that's probably not enough. It's probably necessary for there to be a strong reason to implement it. It's not as if features implement themselves automatically and the only question is which ones to reject instead of deploying.
Please note that I don't speak for the GrapheneOS project.