about ai and its 'evolution' as 'tool'...
...among all the smoke and mirrors of 'progresss'..
i think dere is a great deal of synergy in fpga and llms...
if llms : can be - 'burned' into a big enough fpga..
it is just the way - fpga works..
not bound by an instruction at a time..
but - all 'logic' - executed at a 'clock'..
llms - are 'logic' as traversable by - 'fast' of it..
cpu or parallel matrix ops - are the 'limit' of it..
mee - not a hardware guru - yet..
but i am sure - dere are many..
theoretically - dere seems to be much 'synergy' dere..
and that is 1 kind of module - which is quite - helpful to have..
much generic - many use cases...
kinda - having ur own - cpu for simple cases..
ask 90s pioneers - dey noe - or maybe - 'cant'.. 😵💫