As someone who has fallen into a transition from software engineering to science I think I can say this is mostly wrong.
It seems to me that any given field of "science" isn't any harder than dropping into a new area of software engineering. There's a lot to learn, sure, but there's a lot to do as well and there are areas where an complete rank amateur can make significant contributions.
It's still true that every day I discover something I have no idea at all about. But it seems that this is fairly normal in science - people know a lot about a single specialized area and not much outside it.
The crucial difference is that an amateur can provide a meaningful contribution to an open-source software project while the same is
unlikely in (modern) science.
It seems to me that any given field of "science" isn't any harder than dropping into a new area of software engineering. There's a lot to learn, sure, but there's a lot to do as well and there are areas where an complete rank amateur can make significant contributions.
It's still true that every day I discover something I have no idea at all about. But it seems that this is fairly normal in science - people know a lot about a single specialized area and not much outside it.