Yeah, where did I read it, I think John McCarthy checked it out along with...who could it have been...definitely him.
Today one Lisp Machine CPU emulator written in assembler is 80 times faster than the original Lisp hardware. This could be made faster by a JIT compiler. Native code Lisp compilers for x86-64 or Apple's M1 are roughly 800 times faster than the original Lisp hardware (the third generation Ivory processor).
Wow! Really?! So the M1 really is faster in a non-gimmicky way? I could cut out this X86-64 assembly bullshit, at 800x I could be happy just with Lisp. Assembly all based on C by this point anyway...