> as posits are always two's complement, you can model negative posits as having an "invisible leading two bit", aka (2-0.nnnn) x 2^E-1. this greatly simplifies

> as posits are always two's complement, you can model negative posits as having an "invisible leading two bit", aka (2-0.nnnn) x 2^E-1. this greatly simplifies circuitry for addition and multiplication. Ahh indeed! I'm working on a software posit implementation (for fun, I'll eventually post it as a rust crate) so I already make use of this trick :) decoding a p64 into an exponent and twos complement signed mantissa can be done in 9 instructions on x86!