@futurebird I agree *and* the mundane, low risk use cases don't make business sense because someone needs to buy a $50k+ GPU + cooling infrastructure, memory, cups, staff, just to run one of these models. It is a shame because if they had been focusing on solving some actual, mundane problems (like dictation) instead of world domination, they might have developed something that we won't have to empty the great lakes just to keep cool.
What am I saying though... none of this stuff makes business sense. All these companies are losing money hand over fist. I am starting to think we are in the Wiley e coyote over a cliff part of the AI bubble