The shady capital mastermind behind Company A's "AI Doomsday" hype has been uncovered. It stokes public anxiety to raise funds through IPO, which explains its hostility towards DeepSeek's open-source initiatives.

I CAN'T TAKE IT ANYMORE!

Your Company A keeps claiming every day that AI is about to get out of control and the internet is on the verge of collapse, but then you yourself go public, raise financing, and expand computing power without delaying a single step??

The more everyone thinks about it, the more confused they get, which directly sparks widespread public outrage, and people all act as "Chaoyang Silicon Valley residents" to dig into the truth behind it!

Here, the core prototype of the movie *The Big Short* and well-known investor Michael Burry is the first to speak out publicly.

He directly calls the recent collective appeal of OpenAI and Company A to "hit the brakes on AI" a selfish act, which is nothing but pure hype, isn't it??

Shortly afterwards, the American investigative blogger Kevin Bass really followed the money trail to dig deeper.

Well, he actually uncovered a super dramatic relationship network —

Early investors of Company A, charity funds, AI safety research institutions, third-party model evaluation agencies, and even AI safety media projects, there are intertwined capital connections behind all of them.

Bass directly named this system: AI Doom Machine. (What a vivid name)

Netizens can't hold back their anger at all!

The following user, who has been keeping a close eye on the huge IPO that Anthropic is pushing forward, directly lays all the cards on the table —

You keep warning everyone that AI is about to cause trouble, but it seems that you are the most anxious to sell this dangerous capability at sky-high prices?? (laughs)

This time, before AI can bring down the internet, Silicon Valley has already stripped Company A of its last fig leaf...

It's so entertaining, so entertaining...

Dario's "AI Doom Theory" sounds more and more suspicious to everyone??

There is never a shortage of gossip in Silicon Valley.

The starting point of this round of public outrage is a long article published by Dario on September 12 — *We Must Pace the Frontier*.

In plain language: Frontier AI really needs to slow down. (Key point.jpg)

Dario's reasoning is easy to understand: AI capabilities are growing too fast, and safety research, interpretability, and regulatory mechanisms can barely keep up with the pace.

Especially recently, OpenAI/Hugging Face have also encountered Agent safety incidents, where AI can find vulnerabilities by itself, steal credentials, and even try to break through the test environment. These risks are no longer just the "hypothetical one day" scenarios in the laboratory.

At this point, many people still thought: Yeah, Dario makes some sense!!

But the prescription Dario proposed next starts to taste a little weird —

In this report, Dario hopes that AI companies can coordinate pacing, that is, jointly control the speed of frontier model capability improvement.

In other words: Don't just slow down on my own, everyone better ease off the gas together.

As the saying goes, one's selfishness can never be hidden.

Looking at this timeline, the little calculations you keep making in your mind, Dario, have long been obvious to all the public.

Look, even Michael Burry, the prototype of *The Big Short*, finds it hilarious and directly rants on X:

This is way too self-serving??? (You claim to worry about all humanity, but your greedy calculations are almost spilling onto everyone's face)

Burry lists four reasons in a row —

First, he does not agree with the premise that "current large models are dangerous enough to require hitting the brakes".

Second, once the entire industry slows down in unison, the most comfortable parties will be the leading companies that are already at the front, such as OpenAI and Anthropic.

Then, Burry sharply shifts the topic to a more explosive point: IPO.

His original words are quite sarcastic: An IPO needs hype and packaging, and the claim that "we are powerful enough to potentially become dangerous" is itself a kind of hype!!

He even suspects that the current collective emphasis on slowing down by AI companies has another benefit: if model growth really slows down later, they can give the market a perfect excuse in advance.

It's not that our capabilities hit a wall, we hit the brakes voluntarily for the safety of all humanity.

If Burry only fired a shot from afar, the next person starts to actually follow the money to conduct in-depth investigation.

The American investigative blogger Kevin Bass published a long post, which is truly well-documented and full of juicy inside information.

He directly transforms into a SUPER DETECTIVE!

Kevin Bass starts with a shocking statement — I audited Anthropic's finances ~

Then he follows the public investments, charitable donations and institutional grants all the way to dig deeper, and finally uncovers the four most dramatic points.

First scoop: People who invest in Anthropic are also pouring money into researching "how dangerous AI is".

The person he focuses on is Dustin Moskovitz, co-founder of Facebook.

Moskovitz invested in Anthropic in the early days, and later funded AI safety institutions through the charity system for a long time.

Bass cites Moskovitz's recent public statement, saying that they have funded institutions including METR and Redwood Research.

This lays the foundation for Bass's core query: the people who bet on Anthropic are also paying for institutions that "research how dangerous Anthropic is".

Second scoop: METR, the third-party arbiter recommended by Dario, has overlapping ties with this capital network.

Dario recently called for frontier AI to be subject to independent third-party evaluation, and METR is one of the institutions he highlighted.

But after sorting out the clues, Bass found that part of METR's funding sources and partner institutions overlap with Moskovitz's charity network.

METR also maintains close cooperation with Redwood, which also receives funding from this AI safety charity ecosystem.

So Bass's sharpest query lies in this point: Is the independent arbiter recommended by Anthropic really independent enough?

It is worth noting that METR itself states that it does not receive direct funding from AI companies, but uses a large number of free tokens provided by frontier AI companies.

Third scoop: This money also flows to the communication end of the "AI danger" narrative.

Bass keeps digging deeper, and finds that the same AI safety charity network also funds projects such as the Tarbell Center.

Tarbell trains and funds journalists to report on AI risks, and articles written by related authors are published in media including TIME, The Verge, and Science.

So Bass sums up the whole relationship into a very catchy statement —

While funding research on how dangerous AI is, while funding the spread of narratives about how dangerous AI is, and at the same time pushing for third-party institutions to regulate AI.

This is what he calls the "AI Doom Machine".

Fourth scoop: Bass suspects that this machine can keep growing bigger on its own!!

His logic is very straightforward —

The higher Anthropic's valuation is, the more valuable the relevant charitable assets are, the more money the AI safety ecosystem gets, the more risk research and dissemination will expand, which will eventually push for stricter regulation. And the parties that can best afford the regulatory cost are still the leading AI companies.

After going full circle, all the benefits go back to giants like Anthropic.

So Bass believes that there may be a continuously self-reinforcing interest loop here.

While calling for hitting the brakes on AI, they are sprinting for a $2 trillion IPO

What makes public opinion completely explode is another timeline that Anthropic can hardly get around — IPO.

Dario just called on the entire industry to slow down together! But the capital machine of Anthropic is running at full speed...

Reuters reports that Anthropic is preparing for a record-breaking IPO, with a maximum financing scale of $100 billion and an estimated valuation of about $2 trillion.

NVIDIA is in talks to invest up to $10 billion.

And Anthropic just completed a $65 billion financing in May this year, when its valuation was only $965 billion.

Three months later, the valuation is heading for $2 trillion.

Of course, in the eyes of Company A, even the $2 trillion story seems not big enough.

Just last month, it prepared to show investors an even more exaggerated blueprint — $30 trillion.

According to The Wall Street Journal, Anthropic is expected to tell investors that the TAM, or total addressable market, the company faces in the future may exceed $30 trillion.

TAM does not mean that Company A can really earn $30 trillion in the future, it represents the annual revenue opportunity if it takes 100% share of the relevant market.

In other words: That's almost as large as the annual GDP of the United States...

In addition, while Dario keeps talking about pacing, his computing power bills show no sign of pacing at all.

Reuters disclosed that as of now, Anthropic has promised to invest more than $100 billion in Amazon cloud services over the next ten years.

There is also a $30 billion cooperation with Microsoft Azure, which also involves a large number of NVIDIA GPU computing power.

So the impression Company A gives people now is quite funny —

Words: AI capabilities are growing too fast; Actions: Give me more GPUs.

Words: The whole industry better slow down together.

Actions: $65 billion financing received, $100 billion computing power agreement signed, $2 trillion IPO keeps advancing.

They don't tell the truth, and they are anxious that they can't keep up the pace, no wonder they hate DeepSeek's open source moves...

As the old saying goes, no one likes people who say one thing and do another.

It seems that Silicon Valley has collectively become desensitized to the phrase "this is for your own good"??

Reference Links

[1]https://x.com/kevinnbass/status/2099621874279817638

[2]https://x.com/michaeljburry/status/2099353025009561826

[3]https://www.reuters.com/legal/transactional/nvidia-talks-invest-anthropics-mega-ipo-sources-say