🍭 Perplexity, Paywalls, Power

Good morning. We almost let an AI write today’s issue, but it tried to unionize the paragraphs and demand residuals from your eyeballs.

Let’s dig in. 👇

🍭 What’s Cookin’:

• Publishers vs AI: who decides what models read

• Hollywood’s newest “actress” is fully synthetic

• Humans are starting to talk like chatbots

• Apple’s VP exodus and Meta’s AI news feed

• ToolBox: turn any photo into a product ad

_ToolBox™
🧰 5 BRAND NEW AI LAUNCHES

✍️ BlogBowl
→ Drop in keywords and get AI-written articles auto-posted to your blog so it looks like you “care about content” on autopilot.

🎓 Strater AI
→ Turn any video, PDF, or web page into a personal AI tutor that quizzes you, explains concepts, and actually helps stuff stick.

🗣️ Fluently: AI English Coach
→ Real-time AI English coach that fixes your pronunciation and phrasing so work calls stop feeling like oral exams.

🧠 Infira
→ Multidimensional learning engine that turns complex ideas into living visuals—breathing timelines, pulsing flowcharts, and self-building 3D concept maps.

🌍 Algebras AI
→ Full-stack AI localization agent that translates, rewrites, and checks your product for 300+ languages without 300+ freelancers.

AI Copyright
👨‍⚖️ Who Gets To Decide What AI Knows?

The Bite:
Perplexity just got smacked with fresh lawsuits from The New York Times and the Chicago Tribune for allegedly scraping their work to train its models and then spitting their journalism back out almost word-for-word.

The Times says it even sent cease-and-desist letters and Perplexity kept going anyway.

Oh, and there are claims the AI made up stories and slapped the Times’ name on them.

But this isn’t really a “Perplexity bad, publishers good” story.

It’s a much weirder question:
If you need permission (and big checks) to train on the world’s information, who gets to decide what tomorrow’s AI is allowed to know?

Snacks:

• The lawsuits: NYT and Chicago Tribune say Perplexity scraped their content, trained on it, and reproduced their journalism inside its answers.

• The double hit: They’re mad about both the training fuel and the AI answers that look a lot like paywalled articles.

• The hallucination problem: One allegation: Perplexity invented a scandal and credited it to NYT reporting.

• The gatekeeping upgrade: Big media is now cutting licensing deals with AI labs — basically “pay to be in the model’s brain.”

• The bias twist: If only a few big publishers make it into training, their blind spots and politics shape what the model calls “truth.”

• The open-web reality: The original scoop gets summarized and reposted everywhere anyway, so total control over data is mostly a fantasy.

Why It Bites:
Everyone says they want “unbiased AI,” but you can’t get that if the training data is basically whoever could afford the best lawyer or cut the biggest licensing deal.

If models are blocked from huge chunks of reality because of who owns it, you don’t just protect journalists, you also create an information filter that users never see.

The scary future isn’t just AI stealing content.

It’s AI that politely follows all the rules, trains only on pre-approved sources, and still ends up giving you a censored version of the world.

Steal This Prompt
📸 Any Photo → Product Ad

Turn random product shots or scrappy phone pics into polished, conversion-ready ad concepts without touching Photoshop.

Use it to make your roommate’s sad desk soda look like a Super Bowl spot, or to pitch your stapler as a $200 minimalist productivity object.

Workflow:

1. Click this link (Prompt)

2. Paste it into your favorite GPT

3. Replace the #’s with your product and brand

4. Watch it cook up ad copy and angles your CMO will pretend were their idea

   
   🍭 Try the prompt on Snack Prompt 🍭

_{Everything Else}
🧠 You Need to Know

🗣 Humans Are Starting to Talk Like Chatbots
→ Researchers and moderators report growing evidence that people are unconsciously adopting LLM-like vocabulary and phrasing in real-world speech and online posts.

🍏 Apple Is Losing a Wave of Top Executives
→ Multiple Apple vice presidents, including leaders in AI, hardware, and design, are retiring or exiting amid ongoing speculation about Tim Cook’s eventual departure.

☁️ Salesforce Weighs Rebrand to ‘Agentforce’
→ CEO Marc Benioff is seriously considering renaming Salesforce to Agentforce to align the company’s identity with its AI agent products.

📰 Meta AI Will Pipe In Live News From Major Outlets
→ Meta’s chatbot is adding real-time headlines and links from CNN, Fox News, USA Today, Le Monde, and others to answer user questions about current events.

🏭 SoftBank Floats ‘Trump Industrial Parks’ for AI Hardware
→ SoftBank’s Masayoshi Son is proposing Trump-branded industrial parks on U.S. federal land to host chip fabs and AI infrastructure, though the ambitious plan remains uncertain.

_{Big Screen AI}
🎥 Hollywood’s Newest It Girl Is… A JPEG

The Bite:
Hollywood has a new up-and-comer: Tilly Norwood, an actress who does not exist, does not have a trailer, and will never ask for green M&Ms.

She’s a fully AI-generated “star” built by producer Eline van der Velden, who’s openly trying to make “the Scarlett Johansson of the AI genre.”

Actors are understandably a little jumpy. SAG-AFTRA president Sean Astin calls AI a “tsunami” but says he wants performers to “surf the wave,” not get wiped out.

Meanwhile, a Beverly Hills startup, Kartel.ai, is out here cloning reporter Jo Ling Kent’s face for a full coffee ad.
No set, no crew, just her Instagram and a laptop.

Snacks

• Meet Tilly: It took thousands of tweaks to get her look and emotions “camera ready.”

• AI-only lane: Her creator wants Tilly in her own AI genre, not acting opposite real humans.

• Union energy: SAG-AFTRA is fine with avatars, not fine with AI replacing paid performers.

• Ad world flex: Kartel.ai made a whole “Cup of Jo” campaign from Jo Ling Kent’s photos and a quick capture session.

• Time and money: Brands get fast, cheap, infinite takes; humans get to wonder where they still fit.

Why It Bites
We’re creeping into a world where your favorite “celebrity” might be a character that no one has ever seen in real life, but who still has a brand deal, a fanbase, and a meticulously managed personality.

That’s kinda fun (endless weird stories, no scandalous paparazzi pics)… until you remember these faces are built from real humans’ data, real styles, and real performances.

The real question isn’t “Will AI replace all actors?”

It’s “How many of the faces on screen will quietly be synthetic and will anyone watching care as long as the show is good?”