Liste des Groupes | Revenir à col misc |
On Tue, 28 Jan 2025 22:10:02 -0500This is the truth! The only way to make LLMs perform any job with quality and reliability is to hard code rules and off shore to africa job checkers.
"WokieSux282@ud0s4.net" <WokieSux283@ud0s4.net> wrote:
>In any case, China HAS managed to do serious damage to the "AI>
Bubble". That's gonna COST us big-time and undermine future investor
confidence.
The bubble was due to burst anyway. "Generative AI" is a cocktail of
snake oil and party tricks, pumped up by vulture capitalists selling a
dream to corporate ghouls of being able to fire everyone who actually
makes their business function and replace them with ChatGPT - but LLMs
are never going to do what Sam Altman flagrantly fibs about them doing,
and the more they push it as Good For What Ails Ya, the more obvious
its limitations become (anybody want to try my new wood-glue-and-
sawdust pizza? It's the talk of the town!)
We were already seeing the seismic rumblings of an impending bust lastThis is the truth! Based on my dividends, a small setback would be good to release some nervous energy from the markets. Much preferable to havint doped markets for 2-3 years, and then a huge 30%-40% crash.
fall, when Goldman-Sachs put out a report asking "...but wait, what if
all this money we're burning on AI *doesn't* actually pay off in the
end?" (And just stop to consider: when the people who were just fine
with *subprime mortgage bundling* think an investment might be too
stupid and risky...!) The longer this farce plays out before the
collapse, the worse the damage will be; far better to put a stake
through its heart *now.* If anything, we should be thanking them for
ripping the bandaid off, nevermind that they did it to mess with us.
(I'd highly encourage anyone trying to wrap their head around this
whole circus to check out Ed Zitron - https://www.wheresyoured.at/ - who
has done a ton of good coverage on this over the last couple years.)
Les messages affichés proviennent d'usenet.