LISP evolved on mainframes of the late 1950s. Many of the ideas underlying LISP predate digital computers and were used in mechanical tabulators and analog computers. Those system requirements are fine. The more interesting question is what would a LISP or Smalltalk make better in a microcontroller? If you had to model new behavior and the cost of programming relative to the cost of the device was high then it would make sense.
On the larger end I've yet to see anything even hinting of replacing C, C++ or FORTRAN for numerics and scientific code, especially back-end performance intensive libraries, because nothing can touch them for speed.
I have. Mathematica, Matlab are commonly used. I certainly use PIG all the time for numerics. C and C++ have serious problems with the complexity of implementing parallelism that quite often make them unusable.
Your arguments cut both ways.