Model Wars, Ethics, Enterprise AI
· The Fluency Briefing
The Fluency Briefing
Your Guide to What's Happening in AI and Why It Matters to You
Tuesday, March 10, 2026
A note-taking ring, an empty book, and a billion-dollar bet against large language models - these sound like unrelated stories, but they share a thread. This Tuesday, AI is simultaneously shrinking (onto your finger), expanding (into trillion-dollar infrastructure), and provoking backlash (from 10,000 angry authors). The pattern?
Every layer of society is being forced to pick a side.
Today in AI:
- Yann LeCun Bets $1 Billion That LLMs Are a Dead End - Meta's former chief AI scientist launched AMI, a Paris-based startup raising over $1 billion to build AI "world models" that understand physical reality instead of just text. The round values AMI at $3.5 billion, with backers including Jeff Bezos and Eric Schmidt. Wired
- Zoom Wants to Be Your Entire Office, Not Just Your Meeting App - Zoom announced AI Docs, Slides, and Sheets alongside photorealistic AI avatars that attend meetings for you. The company also shipped deepfake detection and a no-code AI agent builder. Translation: Zoom is coming for Google Workspace's lunch. TechCrunch
- 10,000 Authors Publish a Book With Nothing In It - Writers including Kazuo Ishiguro and Richard Osman released "Don't Steal This Book" - a volume containing only their names - to protest AI training on copyrighted work without permission. The stunt lands at the London Book Fair, one week before a UK government copyright assessment. The Guardian
- OpenAI Sued Over Canada's Worst School Shooting - Parents of a critically wounded student allege OpenAI knew the shooter was using ChatGPT to plan the attack and failed to alert police. OpenAI has acknowledged it flagged the account but didn't contact law enforcement before the February shooting that killed eight people. Fast Company
- A Smart Ring for Your Thoughts - Sandbar raised $23 million for the Stream ring, a note-taking wearable with a proximity-tuned microphone you activate by lifting your hand to your face. The first batch of pre-orders sold out, and some early users record over 50 notes a day. Shipping starts this summer. TechCrunch
- Meta's Deepfake Problem Is Worse Than It Admits - The Meta Oversight Board declared the company's AI content labeling "not robust or comprehensive enough," especially during armed conflicts like the Iran war. The board is pushing Meta to overhaul how it detects and surfaces AI-generated content across Facebook, Instagram, and Threads. The Verge
- Professors vs. ChatGPT: The Humanities Fight Back - Stanford and other universities are experimenting with offline-first teaching - memorizing poems, visiting art in person - to counter students outsourcing their thinking to AI. As one professor put it: "There's no AI-proof anything." The worry isn't cheating; it's the erosion of critical thought itself. The Guardian
- OpenAI's $110 Billion Round Says the Bubble Won't Pop Yet - OpenAI closed a $110 billion raise valuing the company at up to $840 billion, despite earning just $20 billion in 2025 - roughly the same as Frito-Lay makes selling chips. The round was backed by major institutional investors, lending the eye-watering valuation some real-world credibility. Fast Company
Today's Takeaway:
Here's the thing about Zoom's announcement on Tuesday: it's not really about avatars or slides. It's about a company that made its fortune on one feature - video calls - deciding that features themselves are becoming commodities. Zoom is building an entire AI-powered office suite (Docs, Slides, Sheets) that generates content from meeting transcripts, plus a no-code agent builder for non-technical users. As TechCrunch reports, Zoom AI Companion's monthly active users more than tripled in Q4 year-over-year. That's real adoption, not vaporware.
This move directly connects to a quieter but arguably bigger story: SaaS customers are starting to build their own features using AI coding tools rather than waiting on vendors. According to Crunchbase News, enterprise customers now view the traditional feature-request backlog as optional, not inevitable. Zoom clearly got the memo. Instead of waiting to be disrupted by customers who can vibe-code their own tools, it's racing to become the AI layer that makes those tools unnecessary. The question every SaaS company should be asking right now: are you the platform people build on, or the product they build around?
💡 Fluency Moment - Building your AI fluency, one term at a time.
"World Model"
In plain English: An AI that understands how physical reality works, not just patterns in text.
Think of it like: Like the difference between reading every cookbook ever written versus actually knowing how to cook.
Why you'll hear about it: Yann LeCun just raised $1 billion betting world models will replace today's AI.
🧰 Your Toolkit
- Your AI Security Checklist: Staying Safe in an AI-Powered World - [ ] Check if any AI tool you use asks for your passwords or bank details - no legitimate AI needs these.
- Before trusting AI-generated news or predictions (like market forecasts), verify the information on a trusted news website.
- Review what personal information you share with AI assistants - avoid typing your full name, address, or ID numbers.
- If an AI wearable or app (like a smart ring) records your conversations, check its privacy settings and turn off sharing with third parties.
- When a company emails you saying 'our AI found something urgent,' pause and call the company directly before clicking any links.
- Ask ChatGPT: 'What are three questions I should ask before trusting [ANY AI TOOL] with my personal data?'
- Check the app permissions for any AI tool on your phone - if it wants access to your contacts or location and doesn't need them, turn those off.
Taking just 10 minutes to run through these checks can protect your personal information as AI becomes a bigger part of everyday life. Small habits today prevent big headaches tomorrow.
The Bottom Line
The Pattern: Across Tuesday's stories - from Zoom's office suite ambitions to authors publishing empty books to LeCun betting a billion against LLMs - the common thread isn't "AI is growing." It's that every institution, from universities to SaaS companies to publishing houses, is being forced to renegotiate its value proposition in a world where AI can approximate what they used to uniquely provide.
Why It Matters: When a Stanford professor has to assign poetry memorization to keep students thinking, and an enterprise customer tells its SaaS vendor "we'll just build it ourselves," the same thing is happening at different scales. The moat for humans and companies alike is shifting from "what can you produce" to "what judgment, taste, or trust do you bring that AI can't replicate."
Your Move: Pick one thing your team or company charges for - a report, a feature, a service - and honestly assess how long before a customer could approximate it with AI tools. If the answer is "soon," that's not a crisis. That's your cue to figure out what you offer beyond the deliverable itself.
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