• Smarter with AI
  • Posts
  • SunBrief#84: ️ US forces Anthropic to shut down its strongest model

SunBrief#84: ️ US forces Anthropic to shut down its strongest model

Meta and YouTube face youth addiction pressure, while SpaceX prepares the biggest IPO in history.

Smarter with AI banner

 

Welcome to the SunBrief

Today in SunBrief 🌞

  • AI agents that run your influencer marketing end-to-end.

  • Anthropic Suspends Fable 5 and Mythos 5 After U.S. Export Control Order

  • Stock Updates

  • Judge Rejects Meta and YouTube’s New Trial Bid in Youth Addiction Case

  • SpaceX Raises $75B in Record-Breaking IPO

  • AI Highlights of the Week

  • Too Important to Miss

AI agents that run your influencer marketing end-to-end.

Sponsored

Most platforms give you data. Creally does the work. It is the AI platform for influencer marketing and creator partnerships. Its agents run the full pipeline from discovery to signed deal automatically. No manual outreach, no missed follow-ups, no spreadsheets. One client went from 200 to 5,000 monthly outreaches without a single new hire. Same team. 100х more creators.

Anthropic Suspends Fable 5 and Mythos 5 After U.S. Export Control Order

Company disables access globally after national security directive raises concerns over potential jailbreak risk

Anthropic temporarily suspended Claude Fable 5 and Mythos 5 after a US export control order raised national security concerns. The directive requires blocking foreign nationals, so Anthropic disabled both models for all customers while it works on compliance.

Key Points:

  • Models pulled globally: Anthropic disabled access to Fable 5 and Mythos 5 for all customers after receiving the directive.

  • National security concern: The government letter reportedly did not provide details, but Anthropic believes the concern may involve a possible method to jailbreak Fable 5.

  • Foreign national restriction: The order blocks access by any foreign national, whether inside or outside the U.S., including Anthropic’s own foreign national employees.

  • Older models unaffected: Existing Anthropic models like Opus and Sonnet remain available and are not covered by the directive.

  • Anthropic pushes back: The company argues that a narrow potential jailbreak should not justify recalling a commercial model used by hundreds of millions of people.

  • Safeguards already tested: Anthropic says Fable 5 underwent extensive red-teaming with U.S. and U.K. government partners, private organizations, and internal teams, and that no universal jailbreak has been found.

Why It Matters:
This is a major escalation in government control over frontier AI. If export officials can force labs to pull models over jailbreak risks, companies may face stricter rules for releasing advanced AI, especially in cybersecurity and biology.

Was the U.S. government right to force Anthropic to suspend Fable 5 and Mythos 5?

Login or Subscribe to participate in polls.

Stock Updates

Judge Rejects Meta and YouTube’s New Trial Bid in Youth Addiction Case

Court says there was substantial evidence Instagram’s design features harmed the plaintiff

A California judge has denied Meta and Google’s YouTube request for a new trial after a jury found both companies liable for designing social media platforms that harmed a young user. The ruling keeps the landmark verdict in place as both companies prepare to appeal.

Key Points:

  • New trial denied: Los Angeles Superior Court Judge Carolyn Kuhl rejected Meta and YouTube’s attempt to overturn the jury’s decision.

  • Platforms found liable: Jurors previously found that Instagram and YouTube were negligently designed or operated in ways that contributed to harm suffered by the plaintiff.

  • Meta took most blame: The jury assigned 70% of the liability to Meta and 30% to YouTube.

  • Section 230 defense rejected: The judge said Section 230 does not shield the companies because the case focused on platform design choices, not user-generated content.

  • Design features at issue: Judge Kuhl wrote that there was substantial evidence the plaintiff was harmed by Instagram’s design features, regardless of the content on the platform.

  • Appeals coming: Meta and Google both said they disagree with the ruling and plan to appeal.

  • Broader legal pressure: The decision follows another major case in New Mexico, where a jury ordered Meta to pay $375 million over consumer protection violations.

Why It Matters:
This ruling could put more pressure on social media companies by making platform design a legal risk, not just content moderation. If courts keep treating addictive features as liability, Meta, YouTube, and others may face stricter limits on products built for young users.

What should be the priority for teen social media safety?

Login or Subscribe to participate in polls.

SpaceX Raises $75B in Record-Breaking IPO

Elon Musk’s rocket and Starlink company prepares for Nasdaq debut under ticker SPCX

SpaceX is set to go public in the largest IPO on record, raising $75 billion by selling 555.6 million shares at $135 each. The deal values the company at $1.77 trillion, placing it among the most valuable U.S. companies as it prepares to debut on the Nasdaq under the ticker SPCX.

Key Points:

  • Largest IPO ever: SpaceX is raising $75 billion, roughly three times larger than the previous biggest U.S. IPO.

  • Massive valuation: The IPO values SpaceX at $1.77 trillion, putting it ahead of Tesla by market value.

  • Nasdaq debut: SpaceX is expected to begin trading on Friday under the ticker symbol SPCX.

  • Starlink drives revenue: Starlink remains SpaceX’s main revenue source and its only profitable business unit.

  • Heavy losses continue: SpaceX reported a $4.28 billion net loss in the latest quarter and warned it may not achieve profitability.

  • AI spending surge: Capital expenditures reached $10.1 billion in the first quarter, with most of that spending tied to AI after SpaceX merged with xAI.

  • Musk keeps control: Elon Musk controls more than 82% of voting power, giving him near-total control over the company’s board.

  • Trillionaire moment: Musk’s SpaceX stake is now valued at about $866.5 billion, pushing him closer to becoming the world’s first trillionaire.

Why It Matters:
SpaceX’s IPO is more than a stock market debut. It turns the company into a public bet on rockets, satellite internet, AI infrastructure, and Elon Musk’s leadership. The huge valuation shows strong investor confidence, but its losses, capital needs, and Musk’s control will make SpaceX one of the most closely watched stocks on the market.

What does the SpaceX IPO represent most?

Login or Subscribe to participate in polls.

AI Highlights of the Week

  • Ukraine Warns AI Could Outpace Humans in War
    Ukraine says AI is already changing warfare, from drone operations and combat planning to analyzing Russian missile and drone attacks.

    Officials warn future conflicts could become a war of operating systems, where the side that processes battlefield data faster gains the advantage.

  • SpaceX Makes History With Record-Breaking IPO
    SpaceX priced its IPO at $135 per share, raising $75 billion and making it the largest public offering in history.

    Shares opened higher on Nasdaq and kept climbing, turning the debut into a massive win for Elon Musk, employees, early investors, and major banks.

  • Anthropic Pushes for Faster AI Rules
    Anthropic warns that AI is moving faster than policy, raising risks in cybersecurity, biology, and autonomy.

    It calls for stronger frontier AI testing, job-displacement planning, and a strategy to keep democracies ahead.

  • Google Looks to Samsung for Future AI Chips
    Google is reportedly talking with Samsung to make part of its next-gen AI chip, codenamed Icefish.

    The move could help Google reduce reliance on TSMC and give Samsung a stronger role in the AI chip race.

Too Important to Miss

Last Week’s Poll Result

  • Should AI be allowed to design and improve future AI systems?

    No, humans must remain in control → 41.03%

    Yes, that's how progress happens → 33.33%

    Maybe, with strict oversight → 25.64%

  • What's the biggest risk of government ownership in AI companies?

    Political influence over AI → 54.55%

    Conflicts of interest → 45.45%

    Reduced innovation → 0.00%

  • Will memory turn AI from a chatbot into a real long-term assistant?

    Yes, this is the key shift → 45.45%

    No, memory alone isn't enough → 36.36%

    Maybe, but agents matter more → 18.18%

Feedback

We’d love to hear from you!

How did you feel about today's SunBrief? Your feedback helps us improve and deliver the best possible content.

Login or Subscribe to participate in polls.

Know someone who may be interested?

And that's a wrap on today’s SunBrief!

Reply

or to participate.