• Uncovering AI
  • Posts
  • 🔐 Mythos is back — but only for 100 companies

🔐 Mythos is back — but only for 100 companies

Washington's new AI gatekeeper era is officially here... and your job title might matter more than you think.

June 27, 2026

In partnership with

My fellow AI explorers

What a week. Between government letters dropping on a Friday afternoon, Nobel laureates sparring with Stanford professors over payroll data, and the US essentially inventing a new framework for controlling frontier AI on the fly… There was a little bit of everything.

Grab your coffee. Let's get into it.

In today’s edition:

HubSpot AEO

Picture this. A buyer opens ChatGPT and asks for a recommendation in your category. Your competitor's name comes up. Yours doesn't. And that buyer never makes it to your website.

That's happening right now in markets everywhere. And most teams don't know it's happening because it never shows up in their analytics.

HubSpot AEO shows you exactly where your brand stands in AI search, where competitors are getting recommended instead of you, and tells you specifically what to fix. No expertise needed.

Try it free for 28 days. Just $50 a month after.

Claude

🔐 The US Just Created a New AI Gatekeeper, and Anthropic Is the First Test Case

Anthropic's two-week standoff with the Trump administration ended Friday with a partial win.

The US government lifted its block on Claude Mythos 5, allowing Anthropic to release the model to more than 100 companies and institutions, including many Fortune 500 firms and federal agencies. The decision came via a Commerce Department letter from Secretary Howard Lutnick, citing "significant progress" in two weeks of intense daily talks.

Here's what you need to know:

  • The trigger: Concerns were raised, reportedly after Amazon CEO Andy Jassy flagged a potential jailbreak, that Mythos could be misused by hackers and bad actors to find and exploit vulnerabilities exceptionally quickly.

  • The penalty: Anthropic was forced to disable access for all foreign nationals, including its own employees. Every customer. Gone overnight.

  • The resolution: A license is no longer required to export Mythos 5 to entities listed in the government's annex, including their foreign national employees and Anthropic's own foreign national staff.

  • The catch: The letter does not include permission to release Fable 5, the less powerful version of Mythos. Discussions are expected to continue.

But here's the real story that most people are missing.

While Washington was negotiating with Anthropic, Asian competitors were not sitting still. Tokyo-based Sakana AI launched Fugu, a model it says stands shoulder-to-shoulder with Fable 5 and Mythos Preview. And its website explicitly advertises "delivering frontier capability without the risk of export controls." Meanwhile, Chinese cybersecurity firm 360 unveiled Tulongfeng, claiming it rivals Mythos directly.

Anthropic had been on a historic growth trajectory, with run-rate revenue crossing $47 billion in May 2026. How much of that depends on Asian enterprise customers is not publicly known. But the longer these bans drag on, the more those customers find permanent alternatives.

🔮 Prediction: What happened to Anthropic is a preview of a new normal. Washington is building an ad hoc regulatory framework for frontier AI in real time. And every lab with a powerful model is now one security concern away from a two-week shutdown. The winners won't just be the best at building AI. They'll be the best at navigating Washington. How do you feel about AI being treated as a strategic export? Hit reply. I'm genuinely curious.

AI in College

👔 Stanford Has the Receipts on AI and Entry-Level Jobs. It's Not a Blip.

This one is personal for a lot of readers, and it’s honestly not easy to dismiss the data anymore.

Stanford economist Erik Brynjolfsson found a steep decline in employment among people aged 22-25 in AI-vulnerable occupations. He’s now stress-tested his original finding against every major counter-argument. He removed the entire tech sector. He isolated remote-work effects. He controlled for interest rates. The pattern held every time.

The effect has grown by roughly half a percentage point per month, consistently, month after month… and it hasn't mean-reverted. It's grown.

Here's what the Canaries Dashboard (a partnership between Stanford's Digital Economy Lab and ADP Research) is now showing:

  • The data set: 4.6 million total workers across more than 730 occupations, updated in near real-time through April 2026.

  • The finding: Early career workers aged 22 to 25 in certain fields (software engineering, marketing, customer service) have experienced a 16% relative decline in employment since late 2022, even after controlling for firm-level impacts.

  • The nuance: ADP chief economist Nela Richardson argues the key variable is whether AI automates or augments. Occupations where AI augments human work show employment growth; those where AI automates tasks show contraction. Early-career workers sit squarely in the second category.

What makes this uncomfortable is the implication for pipelines, not just payrolls. If companies stop hiring the 22-year-olds who become the 35-year-old senior managers, they're not just cutting costs today. They're hollowing out their own bench.

If the trend continues for young workers in AI-exposed roles, "we're going to see it affect the broader labor market more," Brynjolfsson warned.

🔮 Prediction: This becomes a board-level conversation in 2027. The companies that quietly protect their entry-level pipelines now, even when AI makes it tempting not to, will have a structural talent advantage in three to five years that money can't easily buy back. Are you seeing fewer entry-level roles at your company or in your industry? Reply and tell me what you're seeing on the ground.

30-Second AI Play

🎵 Make a Full Original Song in 10 Minutes

You don't need to play an instrument, write lyrics, or know anything about music production. Here's how to go from zero to a full track:

  1. Go to Suno AI. It's free to start.

  2. Click "Create" and choose "Custom Mode" for more control.

  3. Write a short lyric idea or paste a theme (e.g., "upbeat song about grinding late at night with big dreams").

  4. Pick a genre. Try lo-fi hip hop, indie pop, or cinematic.

  5. Hit Generate. In 30 seconds, you get two full versions with real vocals and instrumentation.

  6. Download the one you like. Done.

Why it matters: Suno's v5 model produces radio-ready tracks up to 4+ minutes long with coherent structure, choruses, and bridges. In 2025, Deezer reported over 20,000 AI-generated tracks being uploaded to its platform daily. Warner Music Group settled its copyright lawsuit with Suno and entered a licensing deal. It’s a sign the industry is shifting from resistance to collaboration.

💡 Pro tip: Use it to score your YouTube videos, podcast intros, or social content instead of paying for stock music subscriptions. Pair it with a simple prompt, and you'll have something unique every time.

Advertise to the 166k engineers and CTOs choosing what tools their companies build with.

Other Relevant AI News!

🧠 Big Short legend Steve Eisman says investors are buying the wrong AI stocks. He argues there are no moats in AI, hyperscalers are in a losing race to the bottom on price, and the real winners are the picks-and-shovels players like Nvidia, Arista, and Cisco.

🌏 Asian AI startups are capitalizing on Anthropic's export ban, with Tokyo's Sakana AI and Beijing's 360 Security positioning their new models as frontier alternatives that come without Washington's regulatory strings attached.

📊 Anthropic's own Economic Index June 2026 report dropped this week, offering a rare look at how Claude is actually being used across industries. It’s a must-read for anyone thinking seriously about AI adoption strategy.

🇮🇳 OpenAI just poached Uber India's president, Prabhjeet Singh, to become its first Managing Director for India, which is its second-largest market after the US. India sees over 100 million weekly ChatGPT users.

👓 Qualcomm's CEO says 6G will turn all of us into "walking cameras." Smart glasses will stream everything you see to AI models in real time, and he's already working with 40+ wearable device makers to make it happen.

Golden Nuggets

  • 🔐 The US government is now the gatekeeper for frontier AI. Anthropic's Mythos standoff just set the template for how Washington will control who gets access to the most powerful models going forward.

  • 👔 The entry-level jobs data is no longer deniable. Stanford and ADP's Canaries Dashboard shows a 16% relative employment decline for young workers in AI-exposed roles, and the effect is growing, not fading.

  • 🌏 Export control bans are gifting markets to Asian competitors. Every week, Mythos and Fable stay locked out of global markets, local alternatives get stickier and harder to displace.

Would love to hear your thoughts! Send me your thoughts by replying to this email (yes, I read them all :)

Until our next AI rendezvous,

Anthony | Founder of Uncover AI