NVidia (Finally) Admits The Obvious - Denninger

 Their "monopoly" is gone.

No, their stock price didn't crash, but they all but admitted Huawei is going to be the lead in China when it comes to the chips run there -- particularly for AI, but not only for AI.  Their CEO, in other words, admitted in public what anyone who actually read any of the material related to Huawei, China and others over the last several years already knew (and which I've pointed out repeatedly.)

Whistling past the market's detonation everyone is.

Why?

Because the same is going to happen domestically, and "AI" isn't going to actually be what people believe......

......Even people who I know are quite intelligent have bought into the BS of those pumping this stuff: The tasks that we as humans find trivial -- and I remind you that the human brain consumes about 20 watts to do everything it does, most of which is consumed keeping your organism alive -- are those which computers require terawatts of power to emulate and they still cannot manage it because there is in fact no actual inference going on.  We don't know how the brain does it (even if the brain is in the head of a cat) and thus we can't reproduce it.......

https://market-ticker.org/akcs-www?post=255447 

.....Oh by the way this does not demonstrate out-of-scope thought since it is simply a search through existing facts -- albeit a very fast one which is what computers are good at.

But the claims of firms like Microsoft's recent pronouncement that 60% of all jobs -- all white collar work -- will be replaced by AI within the next 18 months -- do not pass the giggle test.

That is not going to happen and yet the entirety of the market's current levels are based on that being true.

It is not true.

No computer has ever acted out of scope.

Not even once in the history of computing, yet humans do it every single day.

That humans do so is why with roughly 20 hours of training and a tenth of a kilowatt-hour of energy expenditure a 16 year old human can operate a motor vehicle with 99.9% success thereafter (and improving) while the computer, which cannot, is forced to have a corpus of billions of miles of travel in order to understand that a semi trailer across the road is dangerous -- and when the corpus is missing the current specific danger it decapitates you by driving under said trailer because it cannot actually infer anything.

The "AI" is not hallucinating; it is following its programming.