I think it's far more useful than you seem to, though. Reducing the problem space is huge. It's why macro programming (ala Lisp) is so dangerous. It's why I left Haskell for Python, and why I often wish I was using Idris[1], which pretty much reads your mind and writes what you meant. It's true of coding generally, but SQL programming especially, that there's a conceptually obvious, small space of reasonable programs, but because we're writing in text, we're free to roam over an enormous superspace of (occasionally) wrong or (usually) meaningless programs.
Not sure exactly what you mean by wrong or meaningless programs. (edit3) I'm not sure it's possible to write a meaningless SQL statement.
Edit: it seems like the authors don't see it as that useful. I think the page said they are still looking for use-cases, or something. Have you contacted them?
Edit2 (fta): "However, as of Jan 2019 we’re tabling the project until we have a concrete, motivating use case to inform further development."
It would be more useful if it weren't an OSX-only closed source demo.