AI just pulled off the wildest catfish in internet history. Two "conjoined twins" racked up 300,000 Instagram followers in 6 weeks with bikini photos and heartbreaking backstories. Plot twist: They're 100% AI-generated. Zero disclosure. People sent them money. Brands reached out for collabs. And the kicker? They're still denying it.

Meanwhile, Microsoft's CEO just torpedoed his own $88B AI product in leaked emails, admitting key features "don't really work." Wall Street is spiraling. And you know what's wild? Microsoft is FORCING employees to use it anyway and tracking it in performance reviews. Welcome to corporate dystopia.

Also: Why email is killing your productivity (and the AI hack that fixes it in 5 minutes), a copy-paste prompt that turns meeting chaos into action items, and the free AI tool that's secretly replacing paid apps.

Ready to see how deep the AI rabbit hole goes?

Let's go →

Valeria and Camila launched on Instagram in December 2025 as conjoined twins sharing glamorous photos, answering fan questions, and selling "spicy" content on Telegram. In just six weeks, they gained nearly 300,000 followers. People sent messages. Brands reached out. Followers asked deeply personal questions about their lives, relationships, and medical history.

Here's the problem: AI experts confirmed the account is entirely computer-generated, with digital forensics specialists pointing to hyper-stylized bodies, perfect symmetry, and telltale AI artifacts in the images.

What makes Valeria and Camila different from other AI influencers is the lack of disclosure—they actively deny being AI, complete with fabricated childhood photos and elaborate medical backstories about surgeries.

The creator's goal? Money. The account links to a paid Telegram channel selling adult content featuring these fictional characters.

The takeaway: We've officially entered the era where people can't tell real humans from AI-generated characters—and when confronted, they still choose to believe the lie. If 300,000 people can fall for fake conjoined twins, what else are we falling for?

In leaked internal emails, Microsoft CEO Satya Nadella told his engineering teams that Copilot's integrations with Gmail and Outlook "don't really work" and are "not smart."

This isn't some random product manager venting—this is the CEO publicly nuking his own flagship AI product. Investment firm Melius Research downgraded Microsoft from "buy" to "hold," with analysts saying Nadella has "lost the AI narrative."

But here's where it gets dystopian: Microsoft is FORCING employees to use Copilot and factoring it into performance reviews—even while the CEO admits it's broken. Internal adoption jumped from 20% to 70% in sales teams, not because it works, but because people's jobs depend on it.

The takeaway: Big budgets don't guarantee good products. Microsoft spent $88 billion and still can't get it right. Meanwhile, they're using forced adoption metrics to convince investors that Copilot is a success. That's not innovation—that's corporate theater.

Apple officially announced that the "completely reimagined" AI-powered Siri is launching in March 2026 with iOS 26.4, using Google's 1.2 trillion parameter Gemini model instead of building their own.

Yes, you read that right. Apple—the company famous for controlling every pixel of its ecosystem—is outsourcing Siri's intelligence to its biggest competitor.

The new Siri will have "on-screen awareness" (it can see what you're doing), cross-app integration (it can actually DO things across apps), and will run on Apple's Private Cloud Compute to maintain privacy.

The takeaway: The AI wars aren't about who has the biggest model anymore—they're about who delivers the best experience. Apple just bet billions that integration beats innovation. If they pull it off, Siri could leapfrog every other voice assistant in one move.

One AI tutorial that solves a real problem

The Problem:
You get 50+ emails a day. Important stuff—client requests, boss emails, urgent updates—gets buried under newsletters and spam. You waste 20 minutes every morning just trying to figure out what matters.

The AI Solution:
Let AI auto-sort your inbox and draft responses to routine emails in seconds.

Here's How (Takes 5 Minutes to Set Up):

What You Get:

Time Saved: 15-20 minutes/day = 100+ minutes/week

What it does: Builds a custom inbox system based on YOUR email patterns

Pro tip: Run this again in 30 days to optimize as your patterns change

Why it matters: Google quietly released the most powerful free research tool of 2026, and almost nobody's talking about it.

What it does: NotebookLM lets you upload sources like PDFs, audio files, and websites, then creates a grounded AI expert based ONLY on that data. The free tier allows up to 100 notebooks, with each holding 50 sources and up to 500,000 words total.

Think of it as ChatGPT, but it ONLY pulls from YOUR documents—no hallucinations, no made-up facts, just answers based on what you fed it.

Real-World Use:

The catch:

Try it: notebooklm.google.com

January 27, 2025. DeepSeek—a relatively unknown Chinese AI lab from Hangzhou—dropped a model called R1 that matched OpenAI's best reasoning AI for 1/100th the cost.

Training cost? $5.6 million.

OpenAI's equivalent model (o1)? Over $100 million.

The kicker? DeepSeek built it using DOWNGRADED Nvidia chips (the H800, which the US banned from full export to China) and still beat models from Google, Meta, Anthropic, and OpenAI that had access to the best hardware money could buy.

Wall Street's reaction? PANIC.

Nvidia's stock dropped 17% in a single day, wiping out $589 billion in market value—the largest single-day loss for any company in stock market history. Broadcom fell 17%. Micron dropped 12%. The entire semiconductor industry spiraled.

Why? Because if a Chinese startup can match the best AI models for $6 million while US companies are spending BILLIONS on massive GPU clusters and data centers, what does that say about the entire AI infrastructure build-out?

The AI industry's core assumption—that you NEED to throw infinite money and chips at the problem—just got shattered by a team working with handicapped hardware on a shoestring budget.

The takeaway: The biggest AI budgets don't guarantee the best results. DeepSeek proved that smart algorithms > brute force spending. Silicon Valley is still reeling.

What it is: Canva quietly turned itself into an AI operating system with Magic Studio, which consolidates all of Canva's AI capabilities into one interface, including Magic Media for generating images and videos, Magic Edit, Magic Grab, Magic Animate, and Magic Eraser.

Why nobody talks about it: People still think of Canva as "just templates." They're missing the point entirely.

Why YOU should care:
You can now generate images, remove backgrounds, animate graphics, erase objects, and grab/reposition elements—all inside Canva's free tier. No need for separate tools like Photoshop or Midjourney for basic design work.

Real-world use case:
Need a social media post? Generate the image with AI, remove the background, add text, animate it, and export—all in one place. What used to take 4 different apps now happens in one.

Get started: canva.com - Free tier includes most AI features

Why it's slept on: Because everyone's chasing the flashy new AI tools while Canva quietly became a design powerhouse without anyone noticing.

IT'S A WRAP!

So, there you have it—AI-generated influencers scamming 300,000 people, Microsoft's CEO admitting his $88B product is broken, and Apple outsourcing Siri's brain to Google.

This week's lesson? Don't trust the hype. The biggest AI stories aren't always the best AI products. Execution > budget. Always.

Question for you: Have you ever been fooled by AI-generated content online? Reply and tell me about it—I'm genuinely curious.

📧 Forward this to someone who needs to know | 🔗 Share on LinkedIn | 👍 Saved you time? Let me know!