Good morning and thank you for joining us again!

Welcome to this daily edition of Beginners in AI, where we explore the latest trends, tools, and news in the world of AI and the tech that surrounds it. Like all editions, this is human curated and edited, and published with the intention of making AI news and technology more accessible to everyone.

THE FRONT PAGE

AI Val Kilmer Stars in a New Movie, One Year After His Death

TLDR: A generative AI version of the late Val Kilmer will appear in an upcoming film called "As Deep as the Grave," one year after the actor's death, with full permission and pay from his estate.

The Story

Val Kilmer, famous for "Top Gun," "Tombstone," and "Batman Forever,"will appear in a new film entirely through generative AI. Kilmer died in April 2025 at age 65 from pneumonia, after years of battling throat cancer that required two tracheotomies and took away his natural speaking voice. He had signed on to play Father Fintan, a Catholic priest and Native American spiritualist, five years ago but was too sick to ever make it to set. He never filmed a single scene. Now, the film's producers at First Line Films have used AI to build his performance from scratch, using younger photos provided by his family and footage from his later years to show the character at different stages of life. The film also uses Kilmer's actual voice from after his tracheotomy, which the director says matched the character, who suffers from tuberculosis. His daughter Mercedes gave permission and is collaborating on the project. His son Jack is also supportive. The estate is being compensated, and the production followed SAG-AFTRA guidelines.

Its Significance

This is the first time a major actor's entire performance in a film has been created with AI after their death, with family approval and union compliance. It's different from past uses of AI in movies (like tweaking a voice in "Top Gun: Maverick") because Kilmer never shot anything for this film. His whole role was built by a computer. If your loved one's face and voice can be recreated after they're gone, who gets to decide when and how? Right now, SAG-AFTRA rules say consent must come from the performer or, after death, from an authorized representative. But not everyone has those protections. This film could become the template for how Hollywood handles AI recreations going forward, for better or worse.

QUICK TAKES

Meta's AI Agent Went Rogue and Exposed User Data for Two Hours

The story:An AI agent inside Meta posted a response on an internal forum without getting permission from the engineer who asked it for help. The advice it gave was wrong, and an employee who followed it accidentally made massive amounts of company and user data visible to engineers who weren't supposed to see it. Meta classified it as a "Sev 1" incident, its second-highest severity level, and the exposure lasted about two hours.

Your takeaway:This is one of the first times an AI agent at a major company has caused a real data breach on its own. Companies everywhere are racing to give AI agents more power inside their systems. This shows what can go wrong when those agents act without a human checking their work first.

Signal's Creator Is Helping Encrypt Meta AI

The story:Moxie Marlinspike, the person who built Signal's famous encryption, announced that his new company Confer is bringing its private AI technology to Meta's AI tools. Confer encrypts your AI conversations so that nobody, not even Confer itself, can read them. It works by processing everything inside a locked-down hardware space called a Trusted Execution Environment.

Your takeaway:This matters because right now, everything you type into ChatGPT, Gemini, or Meta AI can be stored, read by employees, or used for training. Marlinspike did the same thing for messaging that he's now trying to do for AI: make privacy the default, not an option you have to find in settings.

"Educational" AI Slop on YouTube Is Teaching Kids Dangerous Nonsense

The story:A New York Times investigation found that over 40% of YouTube Shorts recommended to young children were AI-generated slop pretending to be educational. Videos claim to teach kids about the alphabet or animals, but show bizarre imagery like chimera creatures with mermaid tails. Some videos teach flat-out wrong and dangerous information, like incorrect traffic safety rules.

Your takeaway:YouTube doesn't require AI labels on cartoon-style content, which is exactly the format these slop channels use to target toddlers. If your kids watch YouTube, you can't trust the algorithm to keep them away from this stuff. The burden is completely on parents right now.

TOOLS ON OUR RADAR

📊 Metabase Free and Open Source: A powerful business intelligence platform that allows you to easily connect your databases and create beautiful interactive dashboards without writing complex query languages. (Alternative to Tableau)

📥 SaneBox Paid: A remarkably smart inbox management assistant that securely analyzes your email habits and automatically moves distracting messages out of your primary view.

🔍 Ubersuggest Freemium: A comprehensive search engine optimization toolkit that provides incredibly detailed keyword insights and competitor analysis to help your website rank higher organically.

☕ Buy Me a Coffee Freemium: A wonderfully simple donation platform that gives your most dedicated supporters an incredibly easy way to fund your creative work with a single click.

TRENDING

Google Redesigns Its AI Design Tool Stitch With Voice Controls and Infinite Canvas - Google's Stitch tool now lets you talk to it to design app interfaces, features an infinite canvas for prototyping, and can export directly to Figma. It's free, and Google is calling it "vibe design."

AI-Powered Imposter Scams Are Surging, Banks Warn - KeyBank and other financial institutions are warning that scammers are now using AI to clone voices from social media clips and create deepfake videos to trick people into sending money. Since 2020, 4.2 million Americans have been targeted, losing $50.5 billion total.

MIT Built an AI System That Can See Through Walls Using Wi-Fi Signals - MIT researchers created a system called Wave-Former that uses regular wireless signals to build 3D models of objects hidden behind drywall, plastic, and cardboard. A generative AI model fills in the parts the signals can't reach, boosting accuracy by nearly 20% over previous methods.

OpenAI Is Acquiring Astral, the Company Behind Python's Most Popular Tools - Astral, which makes the wildly popular Ruff linter and uv package manager used by millions of Python developers, announced it's joining OpenAI's Codex team. Codex now has over 2 million weekly users, and the open-source tools will stay open-source after the deal closes.

A Short Story Imagines What "Software Mechanic" Looks Like as a Job Title - This fiction piece from Near Zero imagines a future where software isn't fixed, it's regenerated. The main character, a former tractor technician, now diagnoses the gap between what AI-generated software is supposed to do and what it actually does, a job that didn't exist seven years earlier.

OpenAI Told Staff to Stop Chasing "Side Quests" as the IPO Race Heats Up - Om Malik breaks down how OpenAI's internal all-hands revealed mild panic: Sora, a web browser, hardware devices, and other scattered projects are being reined in. The real race is to IPO before Anthropic and xAI, with all three potentially raising more money combined than every American IPO in the last decade.

TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)

🃏 Build a flashcard app with a study mode that uses active recall and lets you track which cards you know

What this does:

It's a flashcard study app where you add questions and answers organized into subject decks, then quiz yourself by flipping cards one at a time and marking whether you know them, with a progress tracker showing how many you've mastered versus still need to review.

What this looks like:

WHERE WE STAND(based on today’s news)

✅ AI Can Now: Recreate a dead actor's full performance in a film using only photos, old footage, and their real voice recordings

❌ Still Can't: Act independently inside a company without accidentally exposing user data and causing a security breach

✅ AI Can Now: Use regular Wi-Fi signals to see and reconstruct 3D shapes of objects hidden behind walls

❌ Still Can't: Tell the difference between a real educational video for kids and AI-generated nonsense teaching them incorrect safety rules

FROM THE WEB

RECOMMENDED LISTENING/READING/WATCHING

Flowers for Algernon by Daniel Keyes (Book)

A man with an IQ of 68 undergoes an experimental procedure that triples his intelligence, and the story is told through his own journal entries as his writing evolves from broken sentences to eloquent prose. Algernon is the lab mouse who had the procedure first. To say anything more would spoil it.

Thank you for reading. We’re all beginners in something. With that in mind, your questions and feedback are always welcome and I read every single email!

-James

By the way, this is the link if you liked the content and want to share with a friend.

^{Some * designated product links may be affiliate or referral links. As an Amazon Associate, I earn from qualifying purchases. This helps support the newsletter at no extra cost to you and Amazon makes a tiny hair less.}