Yes, take a look at the appendix for my realization about Python from 1991. The comments about Clang and LLVM are also related -- with the Linux kernel and distros and distros in the early stages of migration 15-20 years later.
But to clarify, I don't want implementing Oil to take decades :) I was talking more about adoption. Mass adoption will take decades, just as bash actually wasn't that popular until 15 years ago, but it's also 30 years old. The adoption curve is very long just as it was for Linux, Python, sqlite, etc.
A major reason I write the blog is so that if I decide to stop working on Oil, future maintainers can pick up the knowledge. But I'm also saying it will be embarrassing if I quit before Oil replaces bash for my own use cases. Replacing bash for other people will take a lot longer.
I thought I was the last person who learned Python in 2003, but no, it turns out that several million people were still yet to hear of it :)