Mythos, Misinformation, Medical AI
· The Fluency Briefing
The Fluency Briefing
Your Guide to What's Happening in AI and Why It Matters to You
Saturday, April 11, 2026
An AI model that can hack any operating system, an AI Jesus charging $1.99 a minute, and a hospital CEO saying AI can replace radiologists. Three very different stories, one uncomfortable pattern: AI is no longer just changing what we can do - it's changing what we trust, who we worship, and who gets to keep their job. Here's what you need to know this Saturday.
Today in AI:
- Anthropic Built a Hacker and Then Called the Government - Anthropic's new Claude Mythos Preview model can reportedly discover vulnerabilities in virtually any software and autonomously build exploits. The company pre-briefed the White House and limited release to a consortium including Microsoft, Apple, and Google called Project Glasswing. Wired
- Vance and Bessent Speed-Dialed Big Tech Before Mythos Dropped - Vice President JD Vance and Treasury Secretary Scott Bessent held a private call with CEOs from Anthropic, OpenAI, Google, Microsoft, xAI, CrowdStrike, and Palo Alto Networks to discuss AI security posture and what happens if models start favoring attackers. CNBC
- AI Lego Propaganda Is Winning the Information War - An Iran-linked group called Explosive Media is producing AI-generated Lego-style war videos in about 24 hours, and they're going viral on TikTok, X, and Instagram. The videos reframe U.S. military operations from Iran's perspective in a format engineered for maximum shareability. Ars Technica
- Your Bullshit Detector Is Officially Broken - Automated traffic now accounts for an estimated 51 percent of internet activity, scaling eight times faster than human traffic. The old signals of authenticity - like a zero digital footprint - have inverted, and verification can't keep pace with synthetic content. Wired
- AI Jesus Wants Your Credit Card Number - A company called Just Like Me offers video calls with an AI-generated Jesus avatar at $1.99 per minute, complete with remembered conversations and prayer guidance. The faith-based AI boom now spans Hindu gurus, Buddhist priests, and Catholic chatbots, raising hard questions about spiritual authority and accountability. Fast Company
- NYC's Biggest Hospital System Eyes Replacing Radiologists With AI - The CEO of NYC Health + Hospitals said AI is now better than humans at spotting breast cancers and could replace a great deal of radiologists if regulators allowed it. This comes a decade after Geoffrey Hinton first predicted the replacement - a prediction that initially proved wrong as AI-human collaboration actually grew the field. Semafor
- Valve Is Quietly Building SteamGPT - Leaked files from a recent Steam client update reveal references to an internal AI system called SteamGPT, apparently designed to auto-categorize multiplayer incident reports and flag suspicious accounts. Think automated moderation and anti-cheat, not a chatbot that recommends games. Ars Technica
Today's Takeaway:
Here's the thread connecting Anthropic's Mythos, Lego propaganda, and broken bullshit detectors: we're watching AI erode the verification layer that holds modern institutions together. Anthropic built a model that can find and exploit software vulnerabilities autonomously - then restricted access because the defensive infrastructure isn't ready. Meanwhile, as Wired reports, bots now drive 51 percent of internet traffic, and synthetic Lego war videos produced by Iran-linked groups travel faster than anyone can fact-check them. The speed advantage belongs to the attacker in both cases - whether you're hacking code or hacking perception.
Think of verification like a lock on your front door. It was never perfect, but it kept honest people honest and slowed everyone else down. AI just handed out master keys. Anthropic's response - a private consortium of Big Tech companies and a White House briefing - is essentially saying the old locks don't work anymore, so let's build new ones before we open the door wider. But as The Guardian points out, the real danger isn't any single model. It's that these capabilities will soon be widely available across many models, and our institutions - from hospitals to newsrooms to software companies - are still operating with yesterday's assumptions about what's real and what's secure.
💡 Fluency Moment - Building your AI fluency, one term at a time.
"AI Agent"
In plain English: An AI that takes actions on its own, not just answers questions.
Think of it like: Like hiring an intern who doesn't just give advice - they actually log in and do the task.
Why you'll hear about it: Claude Mythos doesn't just find hacks - it autonomously builds and executes them.
🧰 Your Toolkit
- Your AI Security Checklist: Staying Safe in a World of AI Threats - [ ] Check your phone's notification settings and turn off previews for sensitive apps like banking or email.
- Before believing a shocking photo or video online, use Google's reverse image search to check if it's real.
- Turn on two-step verification (a second code sent to your phone) for your email and social media accounts.
- Ask ChatGPT: 'What are three signs that an image or news story might be AI-generated?' and review the list.
- Search your email address on HaveIBeenPwned.com to find out if your data was ever leaked in a breach.
- Review which apps on your phone can send you push notifications and remove access for any you don't recognize or use.
- When an AI tool asks for personal information, ask yourself: 'Does it actually need this?' and share only what's necessary.
Taking even one of these steps today puts you ahead of most people online. Small habits like these are your best defense as AI makes scams and fake content more convincing than ever.
The Bottom Line
The Pattern: Every major story today - from autonomous hacking models to synthetic Lego propaganda to AI radiologists - shares one trait: AI is outrunning the human systems built to verify, regulate, and judge what's real and trustworthy. The tools aren't waiting for the rules.
Why It Matters: This isn't abstract. If verification fails, your hospital misdiagnoses, your news feed lies to you, and your company's software gets exploited by a machine that works faster than your security team. The gap between what AI can do and what our institutions can handle is the defining risk of this moment.
Your Move: Pick one thing you trusted this week - a news story, a scan result, a security audit - and ask yourself: how would I know if this were wrong? That question is about to become the most valuable habit you can build.
What We're Working On
✨ Founding Cohort Special - 60% Off! - Use code MAF20 to join for just $20/month (regularly $50). Get weekly group sessions & workshops, self-paced courses for all levels, access to tools & templates, challenges with peer feedback, and 24/7 support community. → Join Now
✨ Free 30-Minute AI Consultation - Discover how My AI Fluency can help your business unlock the potential of AI. We'll discuss your goals, explore practical AI opportunities for your industry, and outline clear next steps. → Schedule Free Call
💬 Community | 📞 Book a Consultation | 🌐 Website
Fluently yours, The My AI Fluency Team