Missed the AI Super Bowl?

Welcome back apprentices! 👋

If you thought AI hype had peaked, think again. 

Nvidia’s GTC 2025 just dropped, and Jensen Huang took the stage like a tech messiah in a leather jacket, delivering a two-hour sermon on the future of AI. From next-gen chips to humanoid robots and self-driving cars, the event was packed with enough announcements to make Silicon Valley collectively short-circuit.

Wall Street, however? Not as easily impressed. 

Despite the spectacle, Nvidia’s stock dipped during the event, leaving investors wondering if they had missed something. But in the days since, the stock has been climbing back up, signaling that investors may be warming up to Nvidia’s AI-driven vision after all.

So, what did Nvidia actually announce? 

And why does it feel like we’re one keynote away from robot coworkers and AI-powered everything? 

Let’s dive in. 

In today's email

• First Fully AI-Generated News Edition

• AI’s Moore’s Law

• AI Hunts Wildfires from Space

• AI Stock Watch

• Nvidia’s GTC 2025

• TxAgent Redefines Precision Medicine

_{Read Time: 4 minutes}

Quick News 

📰 Italian newspaper Il Foglio just made history by publishing an entire edition written exclusively by artificial intelligence — from the headlines to the opinion columns (yes, even the sarcasm). This four-page AI experiment, neatly tucked into the paper’s usual edition, proves that AI can string words together with journalistic flair — just without pesky things like interviews or human emotion. While the grammar was flawless, readers noted a distinct lack of bad coffee-fueled rants, which some argue is the soul of real journalism.

💼 AI isn’t just getting smarter — it’s pulling longer shifts. Researchers at METR found that AI’s ability to complete complex tasks doubles every 7 months, with top models like 3.7 Sonnet now handling 59-minute tasks with 50% reliability (which, let’s be honest, is better than some human interns). If this trend continues, AI will be running entire months-long projects solo by 2030, meaning your future coworker might never need a coffee break — or a salary.

🔥 Google just launched FireSat, a literal space snitch for wildfires. This AI-powered satellite can spot fires as small as a hot barbecue (5x5 meters) and report them every 20 minutes—way faster than traditional satellites that still think dial-up is fast. Backed by $13 million and some serious sci-fi vibes, the goal is to launch over 50 of these orbiting smoke detectors to outpace climate change (and maybe shame us into fewer gender reveal disasters).

Together with 1440

The Daily Newsletter for Intellectually Curious Readers

Join over 4 million Americans who start their day with 1440 – your daily digest for unbiased, fact-centric news. From politics to sports, we cover it all by analyzing over 100 sources. Our concise, 5-minute read lands in your inbox each morning at no cost. Experience news without the noise; let 1440 help you make up your own mind. Sign up now and invite your friends and family to be part of the informed.

Join for free today!

NVIDIA
AI, Robots, and Jensen Huang’s Master Plan

This week, Jensen Huang transformed GTC 2025 into AI’s Super Bowl, unveiling next-gen GPUs, humanoid robots, self-driving cars, and AI that practically trains itself.

Here’s exactly what went down:

• New GPUs Galore: Blackwell Ultra (2025), Vera Rubin (2026), Feynman (2028) — each promising performance so ridiculous, we’ll soon need AI to measure it.

• AI Factories: Huang’s new vision for data centers, built to train AI models at a scale that makes sci-fi look outdated.

• Robots Take the Stage: Isaac Gr00T N1, Nvidia’s first humanoid robot AI, plus Newton, a physics engine built with DeepMind and Disney (yes, Mickey Mouse is now in robotics).

• AI Workstations for Your Desk: DGX Spark and DGX Station bring data center-grade computing to personal workstations — because Huang thinks your office needs more raw computing power than NASA.

• GM’s AI-Driven Future: Nvidia is partnering with GM to roll out self-driving cars and AI-powered factories. Somewhere, Elon Musk is probably watching.

Huang also dropped a wild stat: AI computing demand is 100x higher than last year, proving that every tech CEO is either underestimating AI — or Nvidia’s just selling GPUs faster than ever.

So what’s the deal? Huang is all-in on AI, predicting that robots will fill 50 million job vacancies by 2030 — which sounds great, unless your job involves standing still and looking busy. Meanwhile, enterprise AI is moving in-house, with companies ditching the cloud and setting up their own AI models instead of renting a Google-sized data center.

Despite the hype, Nvidia’s stock dipped 3.5% on Tuesday before slowly rebounding, suggesting investors were either underwhelmed or just waiting for Huang to unveil a fully sentient GPU. 

But if Nvidia’s vision plays out, your next coworker might be a humanoid robot — with an AI-generated employee badge.

What Do You Want More Of?

Jensen Huang just dropped a robot-packed, AI-infused, GPU-loaded keynote — because why stop at graphics cards when you can build self-driving cars, humanoid robots, and AI workstations? 

But what deserves a deeper dive?

• Blackwell Ultra & Future GPUs – Are these chips truly next-level, or just another excuse to upgrade?

• AI Factories & Data Centers – Is Huang secretly building Skynet?

• Humanoid Robots – Are they coming for jobs… or just our coffee orders?

• GM’s AI-Powered Future – Self-driving cars + Nvidia… what could possibly go wrong?

🚀 Vote now and we’ll put together an in-depth breakdown of the topic YOU want to explore! 

Halfway through today’s edition — how are we doing?

Let us know what you love, what could be better, or any brilliant ideas you’ve got!

Now, let’s dive back in — what’s AI been up to this week?

Harvard & MIT
The AI Doctor That Won’t Forget Your Meds (or Take a Lunch Break)

Harvard’s Zitnik Lab has unleashed TxAgent, an AI model that could revolutionize how we prescribe medications — apparently, even doctors need a little AI-powered backup. 

Unlike existing models that simply suggest drugs, TxAgent is designed to reason through multi-step treatment decisions while pulling real-time biomedical data from 211 specialized tools (no, it won’t Google your symptoms and tell you it’s cancer).

Early results? TxAgent outperforms GPT-4o by 25.8% and DeepSeek-R1 in structured multi-step reasoning, scoring an impressive 92.1% accuracy in complex drug interaction and contraindication tasks. It doesn’t just match drugs to symptoms — it predicts how medications interact, flags potential risks, and refines treatment plans based on real-world clinical data. 

In short, it's the AI-powered pharmacist you wish your insurance covered.

What’s the deal? The pharmaceutical industry loses billions a year due to medication errors and ineffective treatments. TxAgent could cut that number drastically by ensuring prescriptions are not only correct but also optimized for each patient’s unique medical history. And with AI now handling multi-step reasoning for personalized medicine, doctors might finally spend less time buried in paperwork and more time actually treating patients

If AI keeps getting better at medicine, your next second opinion might come from an algorithm — and unlike your doctor, it won’t judge you for Googling your symptoms at 3 AM.

Even Quicker News 

🎬 Hollywood’s biggest names just told AI companies to keep their digital hands off copyrighted content — because robots don’t get royalties. Meanwhile, OpenAI and Google are arguing that fair use should cover AI training, proving once again that lawyers will be the real winners here.

🚀 Adobe just dropped an entire AI agent army to handle marketing, customer experiences, and more — because one chatbot just wasn’t cutting it. With traffic from AI-driven shopping up 1,200%, these bots might sell you something before you even know you want it. 

⚖️ Elon tried to hit the brakes on OpenAI, but the court basically said, "Not today, Musk." Now, he’s left watching OpenAI speed ahead while his lawsuit stalls out like a Cybertruck on launch day.

Today’s Toolbox

🎮 Roblox introduced Cube 3D, an AI that turns text prompts into fully functional 3D objects—because typing "/generate dragon" is way easier than learning Blender. With 85M+ daily users, AI-powered game dev is about to get so easy that even your grandma might start making battle arenas. 

🧠 Nvidia just dropped its Llama Nemotron models, proving that even AI needs a brain upgrade. With sizes from "tiny genius" (8B) to "overqualified for everything" (249B), these models reason faster, smarter, and might just outthink your boss. 

🎙️ Anthropic is giving Claude a voice — because typing is so last year — teaming up with Amazon and ElevenLabs to make it happen. Soon, it might summarize your meetings before you even realize you were daydreaming.

Let's Chat About AI

This is all for today!

Replay to this email with your thoughts.

AI is more than just a buzzword. It’s a shift in how we live and work. And understanding it a bit better means you can make smarter choices about the tech you use every day.

_FEEDBACK