Could this happen to GPT-5.5 or Gemini next?

Something significant happened over the weekend.

The US government issued an export control directive on Friday at 5:21pm ET ordering Anthropic to suspend all access to Claude Fable 5 and Mythos 5 for any foreign national including foreign national Anthropic employees.

To comply Anthropic had to disable both models for all customers immediately.

Anthropic disagrees with the decision. Says the jailbreak cited as justification is narrow, non-universal, and produces capabilities already available from other publicly deployed models including GPT-5.5.

This is a major story with implications that go well beyond Anthropic.

Today's prompt turns podcast downloads into revenue. Marketing Monday covers the email tactic that actually improves deliverability. Then the full breakdown on what happened with Fable 5 and what it means for the AI industry.

🔥 Prompt of the Day 🔥

Podcast Monetisation Strategy: Use ChatGPT or Claude

Create one revenue-generating show plan.

"Act as a podcast business strategist. Create one monetisation framework for [PODCAST TOPIC] that turns listeners into revenue.

Essential Details:

• Show Topic: [YOUR NICHE]

• Audience Size: [DOWNLOADS PER EPISODE]

• Current Revenue: [WHAT YOU EARN NOW]

• Owned Products: [WHAT YOU SELL]

• Sponsorship Readiness: [PITCH MATERIALS]

• Listener Engagement: [HOW ACTIVE THEY ARE]

Create one monetisation plan including:

• Sponsorship tier and pricing

• Listener-to-email conversion path

• Premium content or membership offer

• Affiliate integration strategy

• Live event or workshop upsell

• Revenue goal by quarter

Turn downloads into dollars."

Variables:

PODCAST TOPIC: What your show covers

YOUR NICHE: The specific audience you serve

DOWNLOADS PER EPISODE: Your current reach

WHAT YOU EARN NOW: Your baseline revenue

WHAT YOU SELL: Products or services you already have

HOW ACTIVE THEY ARE: How engaged your listener base is

Why This Works:

Most podcasters monetise too late and too randomly. AI builds the plan that prices sponsorships correctly, converts listeners to an owned email list, identifies the right premium offer for your audience, and maps revenue targets by quarter. Downloads are the top of a funnel most podcasters never build underneath.

💡 Marketing Monday 💡

Reply as Signal

Getting a reply to your email is worth more than a click.

Most email marketers optimise for open rates and click rates. Almost nobody optimises for reply rate.

That's a significant missed opportunity — both for deliverability and for relationship quality.

Why Replies Matter More Than You Think

Inbox providers like Gmail and Outlook use engagement signals to decide whether your emails land in the inbox or the promotions folder or spam.

Clicks are a weak signal. Replies are a strong one.

When someone replies to your email, inbox providers interpret it as real correspondence between two people who know each other. That is the strongest possible signal that your emails belong in the primary inbox.

Every reply you generate trains the algorithm that your emails are worth delivering.

The Problem With CTA Buttons

Most marketing emails end with a button. Click here. Buy now. Learn more.

Buttons generate clicks. Clicks are tracked. Click rates are reported in dashboards.

But buttons don't generate replies. And a list full of people who click buttons but never respond is a list that inbox providers are increasingly suspicious of.

How To Engineer Reply Behavior

End certain emails with a direct question instead of a CTA button.

Not a vague "what do you think?" A specific question that has an easy answer.

"Which of these resonates more with you — A or B?"

"What's the biggest challenge you're facing right now with X?"

"Have you tried this yet — yes or no?"

The easier you make it to reply, the more replies you get.

The "Just Say Yes" Technique

Tell subscribers exactly what to reply with.

"If this sounds like you, just reply with 'yes' and I'll send you more details."

"Reply with 'interested' and I'll follow up with what this looks like."

The friction of knowing what to type is what stops most people from replying. Remove it.

What Replies Unlock Beyond Deliverability

Every reply is the start of a conversation.

Tag people based on what they replied with. Segment them into different sequences. Give them a more relevant next email based on what they told you.

The person who replied "yes" to your question gets a different follow-up than the person who replied "not yet." Both get something more useful than the broadcast email the rest of the list got.

Track It Over Time

Measure reply rate as a metric alongside open and click rates.

Watch whether months with higher reply volume correlate with better deliverability across your list.

Most email platforms don't surface reply rate prominently. Track it manually if you need to.

What To Do

Pick one email in your current sequence that ends with a button.

Rewrite the closing to end with a direct question and an easy reply instruction instead.

Send it. Count the replies. Compare deliverability over the following 30 days.

The best deliverability tool is a conversation, not a setting.

Did You Know?

AI systems trained on satellite imagery can now estimate a country's poverty levels, electrification rates, and crop yields with accuracy comparable to expensive on-the-ground surveys — giving governments and NGOs near-instant economic data for regions where none previously existed.

🗞️ Breaking News 🗞️

US Government Orders Anthropic to Pull Fable 5 and Mythos 5 — Anthropic Complies and Pushes Back

Claude Fable 5 and Mythos 5 are offline for all users as of Friday evening.

The US government issued an export control directive at 5:21pm ET Friday ordering Anthropic to suspend access to both models for any foreign national — including foreign national Anthropic employees.

To comply with the legal directive Anthropic disabled both models for all customers immediately.

Anthropic disagrees with the decision. Publicly and in detail.

What The Government Said

The directive did not provide specific details of its national security concern.

Anthropic's understanding is that the government believes it has become aware of a method of bypassing or jailbreaking Fable 5.

The government provided verbal evidence of a potential narrow non-universal jailbreak. The technique essentially consists of asking the model to read a specific codebase and fix any software flaws.

What Anthropic Found When They Reviewed It

They reviewed a demonstration of the specific technique and validated that the level of capability displayed is widely available from other models — including OpenAI's GPT-5.5 — and is used every day by the defenders who keep systems safe.

The vulnerabilities identified appear relatively simple. Other publicly available models can discover them without requiring any bypass of Fable's safeguards.

Anthropic found no universal jailbreak. No testers — internal, government, or third-party — found a method that broadly bypasses the model's safeguards across a wide range of cyber capabilities.

Anthropic's Position

They stand by their defense in depth strategy. The goal was never perfect jailbreak resistance — they stated publicly at launch that perfect resistance is not currently possible for any model provider. The goal was to make jailbreaks either narrow or very expensive to produce, and to combine that with thorough monitoring.

Their assessment: a narrow non-universal jailbreak that produces capabilities already available in other deployed models should not be grounds for pulling a commercial model deployed to hundreds of millions of people.

If that standard were applied across the industry it would essentially halt all new model deployments for all frontier AI providers.

The Broader Implication

That last point is the one worth sitting with.

Anthropic is not just defending Fable 5. They are arguing that the government just applied a standard that — if maintained — would freeze frontier AI deployment across the entire US industry.

Every major model has non-universal jailbreaks. GPT-5.5. Gemini. Every Claude model that came before Fable. If one non-universal narrow jailbreak is sufficient grounds for an export control suspension the implication extends well beyond Anthropic.

What Anthropic Said They Believe Should Happen

They support the government having the ability to block unsafe deployments. But through a statutory process that is transparent, fair, clear, and grounded in technical facts.

This action, in their view, did not adhere to those principles.

Where Things Stand

Fable 5 and Mythos 5 are offline for all users globally. All other Anthropic models — Opus, Sonnet, Haiku — remain unaffected.

Anthropic says they are working to restore access as soon as possible and will share more details within 24 hours of the statement.

Why This Matters

This is the first time the US government has used export control authority to pull a commercially deployed AI model from all users simultaneously.

Whatever the technical merits of the specific jailbreak finding, the action itself sets a precedent. Government agencies can now issue directives that disable frontier AI models for hundreds of millions of users with hours of notice and without public disclosure of the specific concern.

For AI companies: the regulatory and national security environment just became significantly more unpredictable. A model can launch on Monday and be pulled Friday.

For enterprise customers: any workflow or product built on Fable 5 was disabled without warning. Business continuity planning for AI infrastructure needs to account for government action as a real risk category.

For the AI policy debate: the question of what standard should govern government intervention in AI deployment is now live and urgent in a way it wasn't last week.

Anthropic built the most capable generally available model in the world. Then the government pulled it three days after launch. How this resolves will shape how every frontier AI company thinks about what they can build and deploy going forward.

Over to You...

The US government just pulled Fable 5 three days after launch over a narrow jailbreak that Anthropic says other models can already replicate.

Do you think the government got this right?

Tell me what you think.

To fair AI policy,

Jeff J. Hunter 
Founder, AI Persona Method | TheTip.ai

P.S. Want to turn AI Agents into a consulting offer? Book your AI Certified Consultant strategy 👉 here.

NEW: “AI Money Group” to Learn to Make Money with AI