Hello and Welcome Back, Neural Squad.

Hold onto your brainwaves, this newsletter’s about to get weird.

Imagine strapping into a $2M sci-fi hair dryer that reads your mind (but panics if you sneeze). Congrats, you’ve just met Meta’s new “brain hat”—a magnet-powered telepathy machine that’s 80% accurate, 100% immobile, and zero percent subtle.

Also , Meta’s AI team went full Pirates of the Caribbean, allegedly torrenting 82TB of books like a digital heist crew.. And OpenAI? They dropped $14 million on a Super Bowl ad comparing ChatGPT to fire and wheels—because nothing says “friendly AI” like comparing yourself to humanity’s greatest hits.

Here’s what you need to know about Todays Briefing:

• Meta’s Mind-Reading Machine — Basically a $2M Brain Hat That Can’t Handle Head-Tilts

• Meta Went Full Pirate Mode: 82TB of Bootleg Books Torrented in Copyright Caper, Lawsuit Reveals

• OpenAI Drops $14M Super Bowl Ad

Meta’s Mind-Reading Machine — Basically a $2M Brain Hat That Can’t Handle Head-Tilts

Key Points:

• Meta built a non-invasive brain-typing system that guesses your keystrokes 80% accurately—no brain implants required!

• Catch? The tech is a clunky, $2M magnetoencephalography (MEG) scanner that looks like an MRI machine tipped sideways. Oh, and you can’t move your head while using it.

• The device only works in shielded rooms (bye, coffee shops) because Earth’s magnetic field drowns out your brain’s tiny signals.

• Meta admits this isn’t a product—it’s brain research to decode how humans process language… and maybe teach AI to do the same.

The Summary: Imagine trying to type with your brain, but you have to sit super still under a giant machine that costs more than a mansion. That’s Meta’s new brain-reading tech! Your brain cells (neurons) send tiny magnetic signals when you think. Meta’s machine reads those signals to guess what letter you’re thinking about. But Earth’s own magnetic field (like the one that makes compasses work) is way stronger than your brain’s signals. So, the machine needs a special room to block Earth’s “noise.” Plus, if you wiggle your head or sip a soda? The machine gets confused. Meta says this isn’t for making gadgets yet—it’s to learn how brains turn thoughts into words.

Why it matters: Brains are still way better than AI at learning languages (sorry, ChatGPT). If Meta cracks how neurons build sentences—first ideas, then words, then letters—it could train AI to think like humans. That’s huge! But don’t expect brain-texting anytime soon. This tech’s like the first clunky computers: cool for science, not your backpack. Still, it’s a peek into a future where AI understands us, not just parrots words. And hey, 80% accuracy? Not bad for a $2M robot mind-reader that can’t handle a sneeze.

Meta Went Full Pirate Mode: 82TB of Bootleg Books Torrented in Copyright Caper, Lawsuit Reveals

Key Points:

• Meta staff allegedly torrented 81.7TB of pirated books from sketchy "shadow libraries" like Anna’s Archive to train its AI, LLaMA.

• Internal messages show employees side-eyeing the ethics: “Torrenting from a corporate laptop doesn’t feel right” (😂 emojis included).

• Zuck himself pushed to “unblock” the project, while Meta tried to hide its tracks to avoid copyright fingerprints.

• Joins OpenAI, Nvidia in the AI Copyright Hall of Shame—everyone’s getting sued for data heists

The Summary: AI models, like Meta’s LLaMA, need to gobble up massive amounts of text to get smart—think of it like a robot reading every book ever. But here’s the twist: using copyrighted books without permission is illegal, like sneaking into a library at night and photocopying everything. Meta’s team allegedly used torrents (aka internet piracy) to download 82 terabytes of books from shady sites. They even joked about hiding their tracks to avoid getting caught—“Oopsie!” Other AI giants, like OpenAI and Nvidia, are also in trouble for similar “borrowing” sprees. The big question: Is AI allowed to break rules if it’s “for science”?

Why it matters: If courts rule against Meta, AI companies might have to pay billions for data or reinvent how they train models—slowing down the AI arms race. Creators (authors, artists) could finally get paid for their work being used by bots. But if Meta wins? Copyright laws might need a serious upgrade to keep up with tech pirates. Either way, it’s a legal drama that’ll shape whether AI grows up as a rule-follower or a chaos gremlin.

TL;DR: Meta’s AI training involved torrenting 82TB of pirated books, internal messages show employees cringing at the ethics, and now the courts get to decide if “AI pirate” is a legit career path.

OpenAI Drops $14M Super Bowl Splash

Key Points:

• OpenAI’s first Super Bowl ad ($$$14M!) hypes AI as humanity’s next big leap, comparing ChatGPT to fire, wheels, and space travel.

• Ad skips scary “superintelligence” talk, focuses on everyday wins (business plans, tutoring) to win over 130M clueless viewers.

• Final animation was 100% human-made, despite using text-to-video tool Sora for early brainstorming.

The Summary: Imagine you’re watching the Super Bowl, and suddenly there’s a cartoon showing how humans invented cool stuff like fire, wheels, and rockets. Now, add AI to that list! OpenAI (the ChatGPT folks) paid $14 million to show their tech is as important as those big inventions. They didn’t talk about robots taking over the world. Instead, they showed AI helping with homework or making a business plan. Even though they used their own AI tool (Sora) to brainstorm ideas, real humans drew the final pictures. Why? Because they want you to think AI is a friendly helper, not a Terminator.

TL;DR: OpenAI spent $14M to convince 130M Super Bowl viewers AI is as cool as fire, wheels, and space—without mentioning robot overlords, because marketing.

Quick Briefing.

• DeepMind AI surpasses math olympiads

• OpenAI CEO Sam Altman admits that AI’s benefits may not be widely distributed

• Scientists Create a Material as Strong as Steel but Light as Styrofoam Using AI

• AMD and Aramco to Explore Potential Collaboration to Help Accelerate Industrial AI Deployment

• Deepseek’s AI model is ‘the best work’ out of China but the hype is ‘exaggerated,’ Google Deepmind CEO says

• AI race must be led by ‘western, liberal, democratic’ countries, says UK minister

AI meme of the day.

Your opinion matters!

Your feedback helps me create better emails for you!

Your feedback helps us create better emails for you!

please feel free to reply to this mail with anything that’s on your mind :)

Thanks for reading.

Until next time!

Erfan and The Neural Brief team :)