Exclusive - Meta plans to lay off employees as AI costs rise

And lobby for children online safety laws and age verification scams with millions they made from our data already

Hold up, that’s where the water issue comes from?  How is that allowed??  There should be a guaranteed amount of water for public use before any company can get any. 

"Ah shit, here we go again"

what number of "one in a lifetime crisis" would this be for millenials? 4th? 5th? though this one has the worst outlook compared to the previous ones, Ill give it that

You can't tell me the gov is just gonna let AI or specific companies go bankrupt. The government will bail them out and then fully consume whichever company is the most successful/beneficial to the government. With trump we can just say he'll get his cut from the highest bidder. If the companies can keep going like this until the next electorate, then it'll be wallstreet all over again. They'll get slaps on the wrist while thousands of folks lives get completely fucked up.

You mean revolution, right?

What to do ? Stockpile cash and move to new zeland ?

and the ironic thing is the only thing preventing a full scale recession is the AI bubble

It will be far, far worse than Netflix. Claude Code $200 plans are thought to cost anthropic about $5000 in compute.

Oh there doing more than Just chasing trends. They p****** a lot of content from books to train their ai. word going around. They might have been funding some of the online age verification cases.

We're finding out that this is just objectively not true lol

Thank you. AI is not a single product, it's a process for creating tools. Some of those tools are legitimately revolutionary, potentially world-changing. Some of those tools are "plausible bullshit" generators. None of those tools are even close to the AGI overlord of Silicon Valley fever dreams.

Can you teach me about what AI should add to my resume that's actually useful?

>>Extractive AI Is that what has been helping me grow really great cannabis for c****? Because aside from the occasional hallucination ChatGPT has been biologically accurate *and* good at finding frugal-value items to add to my watering can.

Sarcasm tags ruin all sarcasm.

Yeah real AI researchers had to do a massive analysis that involved separating the "philosophy out of AI and putting the science back," so that it's possible to make forwards progress again. So, some of us are aware of "how little they know about what they're doing as we can clearly see that they made mistakes because they followed the advice of philosophers instead of scientists." That's "not how math works." The math is legitimately wrong and that's why it hallucinates.

>Well, to be fair, they did introduce a couple of open source models that were pretty good for its time. Not really, they tricked people. It's not AI. It doesn't understand anything. So, it's just a plagiarism parrot like scientists have pointed out multiple times. >Also, keep in mind, Zuckerberg is not very clever. He copied (stole) the idea of Facebook. His own idea of metaverse (technically a copy of playstation home?) was a flop. Yeah he's a "copy catter." He just copy cats other people's work, figures out a way to make gigapiles of money from it, and by turning it into a giant scam. Then apparently he covers up the crimes that are being investigated by law enforcement, because he just got caught doing that in Japan. So, he's a criminal. >This will be another flop because they fundamentally don't understand the technology and its inherent limitations. The main problem with LLM tech is: They don't understand what language is in the first place. They think "people using words is language. That's language usage..." They're unaware that it's a standardized system because they refuse to read the history of language creation... They're so "anti education that they refuse to look there." So, they're legitimately building language technology with no knowledge of language at all... They have absolutely no idea what they are doing and they are way off course... I assure you: There is nobody at Meta that can even accomplish the extremely basic task of classifying words by their type. Obviously, they have no entity detection algo, so, I don't even know what they're doing, because that's how the language works... You're just using words to describe things, that's "how it works." So, they don't have a system to get "the things" or "the words that are used to describe the things." Which, that's a big problem, because "that's the whole language." "That's what human communication is..." So they have nothing... To be clear: You were taught how to do that in kindergarten... It's a two step process. They teach students objects (entities, which are basically just nouns, so "apple", "cat", and many more) and then they teach the students how to describe the properties of the entities. So, "it is just a math equation." First students learn A+Entity (big cat), then they learn A+B+Entity (The young boy), and eventually, A+B+C+Entity (the little young boy). Then the process just repeats three or more times with the (S,V,O) order to form a complete sentence. Because, to save energy, humans combine, typically 3, clauses into one sentence.

Yeah it’s become quite obvious that Zuck just copies ideas, throws a bunch of money and resources and hopes it sticks. It worked out well for Facebook, not so much for the Metaverse and his AI efforts.

Yeah well, after people found out their cool AI glasses are really just spy cameras for Mark, you know, I don't know. I think people are starting to get sick of being lied to, but I'm not sure yet. Maybe a few more scams from big tech and then people will finally figure out that it's a bunch of crooks. They don't even create normal products anymore, it's always some kind of weird trick. I don't even think these companies know how to create a decent product anymore at all because they're going to start the product creation process with some kind of weird scam, that they're trying to hide in the product. So, "the problem their product solves, is obfuscating their weird scam and not actually solving a problem for their customers." That's exactly what LLM tech is... The problem they're hiding is that it's not AI, it's a plagiarism parrot... They thought that by jamming everything into a matrix, that "it's a black box and we can't figure it out." Newsflash: Scientists figured it out. They're a bunch of crooks and liars... What's going on is absurd and those people should be arrested for fraud...

>They started FAIR, which was considered one of the best AI research labs in the industry, over a decade ago. That has no relevancy to the current discussion. But, to "be Fair" pun intended, uh, what they produced sucks. So, I don't know why you are even bringing that up. Compared to graph based techniques that I've seen both, being actively demoed, and with more coming, obviously, what they produced is not "correct." Personally, I'm thinking "beyond graphs", but you know, that's me. I mean obviously when people are soloing these projects and then you compare that to what they were trying to do, clearly they were wasting their time and their parent company's approach to, honestly basically everything, is terrible.