It's not that bad. "All" we have to do is implement everything in Lisp, since Lisp can be reduced trivially to the lambda calculus.

Once again, Lisp saves the day! ;)

EDIT: I'm totally looking forward to experienced Lisp and Haskell programmers like me being in high demand for their legendary skills... at invalidating patents!

Well if you wrote a program in any Turing complete programming language, you would be done; how "trivial" it is doesn't matter. :)

Indeed. By the Church-Turing Thesis, any program written for any Turing Machine (including any Turing-complete language) can be translated into the lambda calculus through a rote transformation.