Lasers, Data Centers, Deepfakes
· The Fluency Briefing
The Fluency Briefing
Your Guide to What's Happening in AI and Why It Matters to You
Sunday, May 31, 2026
Fast one for your Sunday: someone spent four months building a laser cannon that uses computer vision to snipe mosquitoes mid-flight, SoftBank just pledged $87 billion to blanket France in data centers, and AI-generated fake Black influencers are crying on TikTok to sell you dropshipped junk from Shein. The speed at which AI is showing up in places you never expected - your backyard, your shopping feed, entire national energy grids - is accelerating by the week.
Today in AI:
- Mosquito? Meet Your Laser-Guided Nemesis - Engineer Steven Cheng spent four months building an AI-powered laser turret that detects mosquitoes with a custom deep-learning model and a DSLR zoom lens, then zaps them mid-air. It even has a safety camera that cuts the laser if humans or flammable objects are in the line of fire. Tom's Hardware
- SoftBank Drops $87 Billion on French Data Centers - SoftBank announced plans to build up to 5 gigawatts of data center capacity across France, its largest European AI infrastructure bet. The first phase targets 3.1 gigawatts in northern France by 2031, as U.S. data center opposition grows. TechCrunch
- AI Deepfakes Are Selling You Shein Junk in Blackface - AI-generated Black women are appearing on TikTok, crying and begging for views to "save" their small businesses, but the people aren't real and the products are mass-produced dropshipped goods. The grift weaponizes empathy and race. The Verge
- Turkey's Hair Transplant Empire Runs on AI - Turkey's multibillion-dollar hair transplant industry now uses machine-learning algorithms to plan graft placement and customized motors for extraction tools. Translation: your barber might not use AI, but your hair surgeon probably does. Wired
- MeMo Lets You Upgrade an LLM's Brain Without Retraining It - Researchers built a modular "memory model" that sits alongside any LLM and feeds it new knowledge on demand, boosting performance by 26 percent. It works with both open- and closed-source models and sidesteps the cost of full retraining. VentureBeat
- Pearl Crypto Turns GPU Mining Into AI Busywork - A new cryptocurrency called Pearl uses matrix multiplication (the math behind AI) as its proof-of-work. RTX 5090 daily revenue already halved from roughly $34 to $17 as miners flooded in, and most of the "useful" compute is going nowhere. Tom's Hardware
- AI Designs Entire Battery Recipes From Scratch - University of Chicago researchers used AI to generate complete electrolyte formulations for lithium metal batteries, not just picking individual ingredients but balancing the full chemical cocktail. Results matched top-performing lab-designed recipes. TechXplore
- Anthropic Shows Exactly How It Keeps Claude in Its Box - Anthropic published a detailed breakdown of how it sandboxes Claude across products: gVisor for Claude.ai, macOS Seatbelt for Claude Code, full virtual machines for Cowork. The transparency is notable because most AI companies don't document their containment this thoroughly. Simon Willison
Today's Takeaway:
Here's the thing about AI's physical expansion: it isn't a single wave - it's two moving in opposite directions. On one side, SoftBank is committing $87 billion to carpet northern France with data centers delivering 5 gigawatts of capacity, according to TechCrunch. That's happening precisely because U.S. communities are pushing back against data center construction over grid strain and environmental costs. France is betting that welcoming AI infrastructure creates jobs and geopolitical leverage. The infrastructure arms race has gone international, and countries are now competing to host the raw compute that powers everything else on this list.
On the other side, you've got Pearl - a cryptocurrency that supposedly makes GPU mining "useful" by running AI math instead of pointless hashing. But as Tom's Hardware reports, most miners are running inference nobody asked for, which means the compute is AI-shaped but still fundamentally wasted. Through an economic lens, these two stories reveal the same tension: the demand for AI compute is so enormous that billions are being poured into building it and mining it, but not all of that spending is creating real value. The companies and countries that distinguish productive compute from performative compute will be the ones that win.
🔍 Myth Buster
The myth: "AI can only do what humans explicitly program it to do - it has no ability to learn or adapt on its own"
The reality: Modern AI systems like computer vision models demonstrate clear autonomous learning: Steven Cheng's mosquito-killing laser system trained itself to detect and lock onto specific insects by processing thousands of visual examples, developing targeting precision no human explicitly coded rule-by-rule. Similarly, large language models develop emergent capabilities - like multi-step reasoning - that their designers did not directly program, a phenomenon documented across GPT-4, Claude, and Gemini benchmarks where models spontaneously solved tasks they were never specifically trained on.
The nuance: The concern behind this myth is valid in one respect: AI systems are still fundamentally shaped and constrained by their training data and objectives, meaning human choices upstream (what data to use, what to optimize for) deeply influence AI behavior even when the outputs feel surprising or autonomous.
The Bottom Line
The Pattern: AI is no longer confined to software - it's reshaping physical infrastructure, consumer trust, and even national economic strategy. This Sunday's stories span laser turrets, $87 billion construction projects, fake humans selling real products, and cryptocurrencies burning GPUs for math nobody requested. The thread connecting them: AI has escaped the lab and is now bumping into the real world's messy constraints.
Why It Matters: When AI manipulates your empathy on TikTok, rewires a country's power grid, or turns GPU mining into an arms race for compute that serves no one, the stakes aren't theoretical. These aren't demo-day curiosities - they're the early friction points of a technology integrating into supply chains, energy markets, and social trust simultaneously.
Your Move: Pick one story from today that touches your world - maybe it's the TikTok fakes eroding trust in the creators you follow, maybe it's wondering whether your company's LLM needs a MeMo-style memory upgrade. Then ask: am I watching this trend, or is it already affecting me?
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
✨ How AI-Fluent Are You? - Test your AI fluency with our interactive quiz. See how you stack up and discover what to learn next. → Take the Quiz
💬 Community | 📞 Book a Consultation | 🌐 Website
Fluently yours, The My AI Fluency Team