AI Bubble Burst - pop

[analrapist]

You fell for hype

I'm not saying it's good code or anything

The problem with you skeptics is the old question "at what point do you drop the skepticism and acknowledge that something real is happening".

Been fucking with a local OpenClaw instance. Even on 64GB of CPU inference, the bloody thing writes Python faster and better than most of the undergrads I knew.

I didn't fall for jack shit, dawg.

 

[Kether]

The problem with you skeptics is the old question "at what point do you drop the skepticism and acknowledge that something real is happening".

it reminds me a lot of listening to boomers talk about japanese cars they admit they were a punchline when they first hit US dealerships, but most kept pretending they were shit well after it turned out they were a lot better than american cars. i didn't realize people were still shit talking japanese cars in the 1990s.

can AI do everything they tell investors and publish articles about, no. but its at least better than indians and entry level guys. if there was some benchmark it probably would do better than most of the new hires from the past 5 years. the real kicker is AI being able to beat it on both quality and quantity. instead of spending 10 million on h1-b types, a company can use a local LLM and get the same level of accuracy and quality.

 

[Private Tag Reporter]

"The AI bubble is popping!" LOL NO U. Wishful thinking.

Meanwhile at Anthropic: https://www.anthropic.com/engineering/building-c-compiler

https://github.com/anthropics/claudes-c-compiler

Anthropic's agents built a C compiler in Rust that can build the Linux kernel, apparently. Five figures expense in token costs.

I'm not saying it's good code or anything, but the number of individuals who could achieve this if given unlimited budget is quite small, and fewer could do it this cheaply. Please do not underestimate the impact this technology will have. The nightmare is far from over.

i know these guys, i think its called open source or something? yeah they can do it for free

 

[The Mass Shooter Ron Soye]

We have an A&N and Happenings thread on the same thing. Many such cases.

Time: Exclusive: Anthropic Drops Flagship Safety Pledge (archive)

“We felt that it wouldn't actually help anyone for us to stop training AI models,” Anthropic’s chief science officer Jared Kaplan told TIME in an exclusive interview. “We didn't really feel, with the rapid advance of AI, that it made sense for us to make unilateral commitments … if competitors are blazing ahead.”

The Atlantic: Sam Altman Is Losing His Grip on Humanity (archive)

But what’s really significant about Altman’s words is that he thought to compare chatbots to humans at all. Doing so suggests that he views people and machines on equal terms. He didn’t fumble his words; this is a common, calculated position within the AI industry. Altman made an almost identical statement to Forbes India at the same AI summit. And a week ago, Dario Amodei—the CEO of Anthropic, and Altman’s chief rival—made a similar analogy, likening the training of AI models to human evolution and day-to-day learning. The mindset trickles down to product development. Anthropic is studying whether its chatbot, Claude, is conscious or can feel “distress,” and allows Claude to cut off “persistently harmful or abusive” conversations in which there are “risks to model welfare”—explicitly anthropomorphizing a program that does not eat, drink, or have any will of its own.

Ars Technica: RAM now represents 35 percent of bill of materials for HP PCs (archive)

RAM represented about 15 to 18 percent of PC costs last quarter, HP said.

 

[BigMan13]

We have an A&N and Happenings thread on the same thing. Many such cases.

Many such cases indeed. Maybe this should be made into the AI drama megathread.

 

[i440BX]

It won't, the economy has been fake since 2008.

How much can you fake it? It cant be faked till eternity.

 

[Ron Jeremy Stan Account]

So, do we have any actual concrete proof that the AI bubble is going to burst soon? Because when I check any social media as of lately, aside from the current year politisperging there's a ton of doom and gloom over Sam Altman and OpenAI. Hell, every day on YouTube I see endless videos about "It's so over for OpenAI!" with a distressed Sam Altman in the thumbnail.

These videos are the bread and butter of channels like Wall Street Millennial, Meet Kevin, Fireship, and Logically Answered. Go see for yourself: type on "OpenAI collapse" into YouTube search and see endless videos with similar thumbnails, going back as far as last year. They're just as bad as those "China is doomed!" videos that have been a thing since like 2012.

Don't get me wrong: I absolutely have AI fatigue, I'm sick of the RAM shortage, I'm skeptical of the longevity of OpenAI as a company, and the CCP can get fucked with a rusty pipe. But in the meantime, it ain't over until it's over.

 

[i440BX]

My belief is that the AI industry is definitely a bubble, but it might take a while to pop. This bubble pop will not erase LLMs from existence, but it might be probably seen as a tool and not an idol\god that would replace human labour. For example, the impact of the internet on employment wasn't that harmful.

LLMs can be great if they are trained on good academic sources and used for programming.

One thing I noticed that nobody here brought it up is that LLMs have become dumber than in the last 2 years. But LLMs are definitely better than trying to find information on enshitified search engines.

No wonder people ask LLMs more when Google delivers SEO meaningless garbage.

 

[SIMON THE PIEMAN]

For me the sign that the bubble is about to pop soon was a recent article where Satya Nadella ask people to "do something useful with AI". I don't know how this interview flew under the radar because this is just insane: a CEO of a company that put billions into developing their own AI tech and then put said tech into virtually every single one of their products has to beg people to use these features and go all "SAAR PLEASE DO THE NEEDFUL WITH COPILOT". A CEO of one of the biggest tech companies to ever exist can't point to a single example of their own making that would be worth showing and I am supposed to believe that this industry is not a bubble? Why?

 

[Rootless Cosmopolitan]

How much can you fake it? It cant be faked till eternity.

You can lie and tell people that they’re getting richer for 5 years, maybe 10. They’ll believe you… for a while.

If some people are getting rich and others aren’t, the picture is murky. And you can always point to individuals and blame them — you’re not smart enough, you didn’t work hard enough, you couldn’t take the competition, you didn’t make the sacrifices.

In this way, you can string people along for quite some time, filling them with self doubt and shame and uncertainty. But the truth will surface eventually and people will be very angry.

Eventually… two decades, maybe?

 

[Tanner Glass]

How much can you fake it? It cant be faked till eternity.

They're trying their best

 

[ArnoldPalmer]

It's important to remember that a round robin takes a long time to fully lose momentum. You're seeing the end, but the end is probably going to persist into the 2030s.

The way I see it, the U.S. Economy has six years to pull the fuck out of AI investments, before we make the Great Depression look like a skipped brunch.

 

[TEA 3HF]

So, do we have any actual concrete proof that the AI bubble is going to burst soon? Because when I check any social media as of lately, aside from the current year politisperging there's a ton of doom and gloom over Sam Altman and OpenAI. Hell, every day on YouTube I see endless videos about "It's so over for OpenAI!" with a distressed Sam Altman in the thumbnail.

These videos are the bread and butter of channels like Wall Street Millennial, Meet Kevin, Fireship, and Logically Answered. Go see for yourself: type on "OpenAI collapse" into YouTube search and see endless videos with similar thumbnails, going back as far as last year. They're just as bad as those "China is doomed!" videos that have been a thing since like 2012.

Don't get me wrong: I absolutely have AI fatigue, I'm sick of the RAM shortage, I'm skeptical of the longevity of OpenAI as a company, and the CCP can get fucked with a rusty pipe. But in the meantime, it ain't over until it's over.

I was in the same position myself. Trying to figure out when, who was full of more shit, etc. I looked into wall street millenial, Ed Zitron, this autisto, Roman Yampolsky and a few others I do not give enough of a fuck about to recall. Overall I have come away with the following:

LLM's are literally just statistical engines spewing out tokens based on hierarchical calculations run across a vast set of training data.
Proponents are often utter retards that mistake this for a real, living intelligence with its own dasein and internality.
Skeptics are often salty bitches that end up eating shit when they say "oh, but it can't do THIS!!!!1!" only to watch a model eventually reach a point that it can get to a believable simulacrum of that task.
Meanwhile, in reality, AI companies and labs are headed up by slobbering mongoloids who acknowledge that regardless of the gross scale of training data fed to the LLM, there is a performance threshold thry will never be able to breach.
So, what they do is keep trying to brute force it anyways with sketchily scraped, largely shit quality data they pirated or paid someone else to pirate. The result being that the LLM's do, in fact, improve, but only to the point allowable by scale.

Ergo, in my overall opinion has become twofold the more I stare into the abyss:

What really pops the bubble is when the scaling workhorse finally gives out, breaks a leg, and is unceremoniously shot int he head and sold to the butcher. This entire bubble ends in a long, wet fart and becomes a contest of who can stand the smell the longest, gradually rupturing and starting a deflationary spiral that is hard to notice at first and even harder to time correctly.

Sam Altman is a faggot giga-N I G G E R fucking Jew that will probably run off to Israel the moment the goyim notice with a cool billion in CHF.

 

[Fruit Flies]

Very curious to see where it’s going to go from here. Week or so ago there was this article posted by some AI company suit who was essentially writing what came across as half schizophrenic delusions and half a thinly veiled sales pitch:

https://shumer.dev/something-big-is-happening

I don’t really want to bother reading it again to cherry pick lines but the general gist of it is

“whoah, these machines are learning faster than we know what to do with them… I bet in less than a year, AI will have made like 90 percent of jobs obsolete! So I say to YOU the people: pay for ChatGPT and use it every day for at least an hour!”

The thing was even framed as a “letter he wrote to his family and colleagues” that, for some reason, he decided to publish and spread around online. The whole thing reeks of scammy sales tactics.

 

[The Ugly One]

The problem with you skeptics is the old question "at what point do you drop the skepticism and acknowledge that something real is happening".

Something real was happening with railroad, air travel, and the internet. They were transformative technologies. The unbridled enthusiasm they caused that led to utterly retarded business plans was real. The waves of bankruptcies that followed as financial reality caught up to speculative mania were also real.

This is the process:

1. Transformative technology
2. Unbridled enthusiasm
3. Retarded business plans
4. Financial reality
5. Waves of bankruptcies
6. Readjustment to sanity

We are late in step 3 and entering step 4.

So, do we have any actual concrete proof that the AI bubble is going to burst soon?

Cracks are forming. One of the last stages of an irrational market is when people are paying for things with equities or other financial instruments instead of money. That's been going on for a little while now. We're starting to see cooler heads in the financial sphere urge caution rather than doubling down. The financial ouroboros that boosted Oracle's stock has died, and the stock has returned to where it was before. It's at least the end of the beginning.

But when will it really burst? Who knows! Tesla's market cap is now $1.5T, making it as big the entire rest of the global auto industry combined, despite delivering only around 1% of the auto industry's sales, the Cybertruck failing, the Model S being pulled off the market, and Optimus failing to deliver to expectations.

We now live in an era where companies that don't really do much can command valuations larger than entire industries, and their stock just keeps going up, up, and up.

What really pops the bubble is when the scaling workhorse finally gives out, breaks a leg, and is unceremoniously shot int he head and sold to the butcher. This entire bubble ends in a long, wet fart and becomes a contest of who can stand the smell the longest, gradually rupturing and starting a deflationary spiral that is hard to notice at first and even harder to time correctly.

It won't be gradual. There is around $2T of on-paper value that people have built massive enterprises around, which is all based on unfunded capital expenses, that will suddenly evaporate. We're using a $10m worth of GPUs and electricity to drive $1m of revenue. When the VCs paying for all this decline to fund the next investment round, it all collapses very suddenly.

 

[Stoneheart]

Its not a bubble, its a money pit for big tech they can stop shoveling money into.
The main player are all big enough to eat 10 billion in AI loses indefinitly and they now have to.

People will not pay full price for AI and you need way to many ads.

 

[BigMan13]

PLing a tiny bit: My father is a senior product manager at an AI department, he is one of the relatively few who develop the narratives about AI, which the CEOs then tell the media, which the media then tells the public, which the public then tells themselves. He's a source of the cancer. And he is the most full of shit nigger I have ever had the misfortune of knowing, much more being related to him.

He is a satire of his own position.
- AI is always 6 - 18 months away from achieving perfection
- "Think about the future"
- No, he has no plan or idea of how to get there
- But he gets pissed when others don't have a "plan" (lie) to tell investors
- He responds to any critique of AI with sophistry extrapolating a tiny improvement on a niche area
- He's a lifelong Star Trek fan (fine, good) who consumes all of the new stuff (bad), currently binging Academy (ungodly)
- Would rather be abducted by aliens, leaving everyone/thing behind, has consistently stated this numerous times
- Many struggle with something mentally and work to improve; he instead drops so many pills and therapy sessions I don't actually know who he is without them

These are the people/elites telling you AI is going to take over the world. Not just as a tech here to stay, but going to replace everything. Trust me when I say this is NOT specific to him. Everyone in a similar position to him, is exacty like him in different ways.

Honestly, if there's any requests within reason to learn more about what AI is to the people who believe in it most, I'll ask him, or respond from what I've already talked with him.

Also crosslinking a deep thought I had on AI limitations recently (the person referenced later in that is someone else).

 

[The Ugly One]

He's a lifelong Star Trek fan (fine, good) who consumes all of the new stuff (bad), currently binging Academy (ungodly)

If my dad was watching Academy, I'd start telling people I'm an orphan.

 

[Rust Troony]

GUYS PLEASE REDEEM ANTHROPIC STOCK! DO THE NEEDFUL SAARS AGI IS HERE! THE HECKING CHAT BOT IS HAVING AN ANXIETY ATTACK OR SOMETHING!

 

[Tri-Tachyon'sClown]

*snip*

This probably detracts from the point of your post but I'm sorry you have to deal with this. Having a negative relationship with any of your parents (or any direct family member) is horrible and I hope you have at least a decent relationship outside of this AI nonsense. I still wish everyone deeply supportive of this AI faggotry a horrible unemployment, no offense.

Honestly, if there's any requests within reason to learn more about what AI is to the people who believe in it most, I'll ask him, or respond from what I've already talked with him.

What are some of the worst examples you've seen or heard from him?