Two "mores" so you know I'm being as forcible as I can here. I am tired of seeing the most OBVIOUS traits of the "Generative #AI" fad go without commentary, because they fall into that vast region of taboo and ostracism pertaining to the designed-in stupidity and inefficiency of corporate #business practices and leadership culture. This is the inevitable result of caring only about making numbers go up, especially anything pertaining to the dollars-per-second #tech sector obsession with a particularly stupid definition of "efficiency". The people who get the furthest in this pathological culture are the ones who have nothing in their heads other than making numbers go up. To them, all other considerations are "inefficient."
Intelligence itself is inefficiency, as I've pointed out. It takes time to work out chains of reasoning. Pausing for second and third thoughts, rather than simply acting upon impulse or "gut instinct" or whatever, is very easily construed as a waste of time and damaging to "efficiency", i.e. dollars per second in the John Carmack or DOGE kiddie definition of "efficiency".
They shouldn't call the #LLM "artificial intelligence". If "AI" stands for anything in LLM terms, it's something like applied impulsiveness: the LLM is designed with the needs of terminally impulsive tech execs in mind. They want instant answers they don't want to think about, and thus they think it's natural to get all possible answers from a black box that's simply declared to be a box stuffed to bursting with "AGI superintelligence."
The false claim of supreme intelligence is what supports the impulsiveness: the black box is SO impossibly brainy, think the tech weenies, that it's got all the answers right off the top of its head, able to be served up by the billions from some vast GPU farm or other, an "AI factory" (they don't like "data center" so much any more, too many people hate that phrase now.)
(cont'd)