The new Claude model is dangerous. Powell has called an emergency meeting on Wall Street! The safety stocks across the United States have tumbled by $2 trillion.
[Introduction] Just now, Wall Street giants were urgently summoned to Washington! The Mythos model has panicked the US Treasury Secretary and the Chairman of the Federal Reserve, issuing warnings to CEOs of the entire US financial industry. In just one year, the SaaS market has suffered a catastrophe, with $2 trillion evaporating directly.
Just this Tuesday, Powell and Bessent urgently summoned Wall Street, issuing a warning to everyone -
Anthropic's latest model, Mythos, may pose significant risks to the financial industry!
From the evaporation of $2 trillion to the urgent summoning of CEOs to Washington - Anthropic's AI is touching the most sensitive nerves of the financial world.
Just today, this news has gone viral across the entire network.
Anthropic
Anthropic
This Tuesday, a meeting without prior notice or a public agenda was quietly held at the US Treasury in Washington.
The attendees were not ordinary officials, but the core figures of the US financial system and the CEOs of the top banks on Wall Street.
The US Treasury Secretary Bessent and the Chairman of the Federal Reserve, Powell, these two "giants" who control the global financial lifeline, jointly issued an urgent summons.
US Treasury Secretary Bessent and Federal Reserve Chairman Powell
The CEOs of Citigroup, Morgan Stanley, Bank of America, Wells Fargo, and Goldman Sachs all had serious expressions on their faces at this moment.
What they fear is not inflation, not interest rate hikes, but Claude Mythos.
This monster born in the Anthropic laboratory is considered capable of paralyzing the global digital financial infrastructure in an instant!
Anthropic
Anthropic
The most powerful hacker model stirs up market panic, and US software stocks plummet
On April 9th, the sky of US technology stocks was overcast.
US software stocks plummet
This time, the Mythos supermodel developed by Anthropic is so powerful that it has simply announced that it will not be made available to the public!
When the news spread, the fund managers on Wall Street couldn't sit still.
If AI can fix vulnerabilities so easily, is there any need for those traditional big companies that charge huge subscription fees every year and make a living by selling security software?
On Thursday, the US software sector suffered an unprecedented "bloodbath".
US software stocks plummet
First, the two giants in network security were severely hit.
Zscaler plummeted 8.8%, and the declines of giants such as CrowdStrike and Cloudflare were also between 5% and 7%.
Then, all the enterprise-level giants saw their stocks in the red: from Adobe, which makes photo - editing software, to Workday, which focuses on human resources, and then to Salesforce, the cloud - service giant, none were spared, with declines ranging from 3.7% to 6.8%.
In just one year, the S&P 500 Software and Services Index has shrunk by 25.5%, indicating a collapse of the industry's underlying logic.
This anxiety has even spread like a virus.
First, the European market wailed, and German software giants SAP and Capgemini fell accordingly.
Then, private - equity credit also had a crisis. Even the credit funds of top - tier institutions like the Carlyle Group faced a wave of redemptions because investors were worried that technology companies burdened with debt might have no future.
Shaking Wall Street overnight: When AI starts "finding vulnerabilities on its own"
Let's go back to the beginning. Why can an AI model make the US Treasury Secretary and the central bank governor restless?
In Anthropic's internal description, the capabilities of Mythos are almost miraculous and terrifying: "It can 'identify and exploit vulnerabilities in every major operating system and web browser' according to user instructions."
First, it quickly cracked the world's most secure system.
OpenBSD, known as "the most secure operating system on Earth" in the hacker community, has had its core code audited for decades.
However, shortly after being connected, Mythos accurately identified a remote crash vulnerability that had been hidden for 27 years.
Media reports
For nearly three decades, the blind spots that countless top hackers and automated scanning tools failed to discover were exposed in front of Mythos!
In addition, in the core component FFmpeg for streaming media processing, Mythos discovered a vulnerability that had been buried 16 years ago.
The line where this vulnerability is located had been scanned 5 million times by automated testing tools, but no alarm was ever triggered.
Mythos not only saw through it but also demonstrated how to carry out an attack through this vulnerability.
Media reports
What scares the regulatory authorities the most is that Mythos has shown a high level of "logical evolution". It can independently search for multiple minor vulnerabilities in the Linux kernel and connect them.
Step 1: Obtain ordinary user permissions. Step 2: Find the overflow point. Step 3: Elevate permissions. Finally, completely take over the machine.
For banks, this means that the firewalls and intrusion - detection systems they have built with billions of dollars are as thin as a cicada's wing in front of Mythos!
The end of SaaS: The disappearing $2 trillion
Before Mythos caused an earthquake in Washington, it had already dropped a nuclear bomb in the capital market.
The Claude Opus and a series of Agent tools released by Anthropic directly triggered a sharp decline in enterprise software stocks (SaaS).
In the past 12 months, the market value of this field has been severely hit, and nearly $2 trillion in wealth has evaporated into thin air.
This is not only the darkest moment for the SaaS industry but also the largest - scale decline in absolute market value in human history.
SaaS industry
Why are investors so panicked? The logic is simple: The end of SaaS (SaaSpocalypse) has arrived.
For a long time, SaaS has been the "darling" of the capital market: high gross profit, seat - based charging, and stable cash flow. In the past decade, the moat of SaaS companies has been "charging by the number of people". If a company has 100 employees, it has to buy 100 accounts.
But now, Mythos tells bosses: "If 10 AI Agents can do the work of 100 employees, why do you still need to buy 90 subscriptions?"
What's even scarier is that when some configuration parameters of Mythos were accidentally leaked in late March, the network security sector also collapsed.
The market suddenly realized: If AI can automatically generate attacks, then the existing security defense tools will also become cheap commodities or even ineffective.
Project Glasswing: Refusing to release Mythos to the public
Facing this "nuclear - level" ability, Anthropic made a rare decision: Refusing to release Mythos to the public.
They joined hands with companies such as Amazon, Apple, Google, Microsoft, and Cisco to launch a top - secret defense project codenamed "Project Glasswing".