That's basically the core idea of ML, though, at least that was my main takeaway from reading Lenin. As an anarchist myself, my main problem with ML is that it doesn't seem to provide any actual mechanism for that last crucial step. I don't think a state can be both stable and allow itself to be gradually abolished, it has to be destroyed by revolution.
I think if you want to build a state that you want to disappear later, it needs built-in breaking points for the anarchist followup revolution, while still being resistant to capitalist counterrevolution. And I have no idea how to do that.
That said, I think ML(M) is remarkably effective at what it does, and will form a central part of whatever theory synthesis we ultimately end up with.
It seems absolutely insane to me that you would ever use electronic voting machines. It's just impossible to trust them. Voting needs to be low-tech, paper ballots marked with a pen (pencil would be even better tbh), and a sealed urn. What's so difficult about this?