Instagram | Sponsor this newsletter | Let’s Connect!

My fellow AI explorers

While Google is taking a victory lap with Gemini 3 and their massive TPU architecture, the mood over at OpenAI has shifted from "cruise control" to "Code Red."

I’ve been tracking the "Scaling is Dead" narrative for months—hearing everyone from Ilya Sutskever to Andre Karpathy suggest we hit a wall. Then Google dropped Gemini 3, smashed the wall, and now the race is back on. We are witnessing a pivot back to hard science and massive pre-training runs.

In today’s edition:

• 🚨 OpenAI’s "Code Red" & Secret Project Garlic

• 🏗️ Can AI fix 50 years of stagnation? (The World of Atoms)

• 🔀 How to run every top LLM (Gemini, GPT, Claude) in one place

Must See AI Tools

• 💰 Payman: AI That Pays Humans. Over 10,000+ signed up for the beta

• 💫 SubMagic: An AI tool that edits short-form content for you! (Get 10% off using code “uncoverai” at checkout)

• 🎤 11Labs: #1 AI voice generator (Click Here to get 10,000 free credits upon signing up!)

• 🤖 ManyChat: Automate your responses & conversations on IG, FB and more! (Click Here to get first month for free)

• 🎙️ Syllaby: The only social media marketing tool you’ll ever need - powered by AI! (Get 25% off the first month or any annual plan with code “UNCOVER” at checkout)

Earn a master's in AI for under $2,500

AI skills aren’t optional anymore—they’re a requirement for staying competitive. Now you can earn a Master of Science in Artificial Intelligence, delivered by the Udacity Institute of AI and Technology and awarded by Woolf, an accredited higher education institution.

During Black Friday, you can lock in the savings to earn this fully accredited master’s degree for less than $2,500. Build deep expertise in modern AI, machine learning, generative models, and production deployment—on your own schedule, with real projects that prove your skills.

This offer won’t last, and it’s the most affordable way to get graduate-level training that actually moves your career forward.

Learn More

Open AI

Open AI Declares Code Red: Project 'Garlic' vs. Gemini 3

Sam Altman isn't taking Google's recent dominance lying down; reports confirm OpenAI is scrambling to regain the pre-training crown.

The "scaling is dead" narrative just died, and the labs are back to the grind.

Here’s what’s cooking:

• Project Garlic: A new secret model specifically designed to counter Gemini 3’s gains in pre-training.

• The Pre-training pivot: OpenAI’s Mark Chen admitted they are "building that muscle" again after focusing heavily on post-training (RLHF).

• Delayed flagships: Development on other projects (like ads and shopping) is paused to focus purely on model quality and Chat reliability.

Google's Gemini 3 proved that with enough custom silicon (TPUs), scaling laws still hold. This caught many off guard, including OpenAI, who hadn't released a broadly deployed frontier model since GPT-4o in May 2024.

The Deep Dive:

For the last year, the smartest minds in AI claimed LLMs were hitting a plateau. They argued we needed new paradigms, not just more compute. Google proved them wrong. By leveraging their massive TPU fleet, they brute-forced a leap in intelligence.

Now, OpenAI is playing catch-up. They are incorporating bug fixes from another codenamed model ("Charlotte Pete") into Garlic. The goal? Prove that their infrastructure can match Google's silicon advantage.

🔮 Prediction: The "Application Layer" era is paused. We are entering the "infrastructure war." Expect 2025 to be defined not by new chatbots, but by who has the biggest clusters and the most efficient pre-training runs.

Founders…

🤖 Need an AI Agency to Help Your Business Implement AI Solutions?

Our preferred partner Align AI provides you with an expert AI and Automation implementation team to add 10-40+ hours of increased productivity per employee and achieve your goals faster.

Click here to schedule your company’s complimentary AI and Automation Strategy Plan

The Physical World

Peter Thiel: AI vs. The "World of Atoms" Stagnation

We spend 90% of our time obsessing over code, but in a rare new interview, Peter Thiel dropped a heavy reality check. He argues that while the "world of bits" exploded, the "world of atoms" has been comatose since the 1970s. According to Thiel, AI is the only force capable of breaking a 50-year stagnation. Here is Thiel’s thesis:

• The "Mafia-Adjacent" Problem: Thiel points out that digging tunnels in New York costs 10x what it costs in Paris. He explicitly calls the process "mafia-adjacent" and "corrupt labor union adjacent." It’s not a tech problem; it’s a political one.

• The 50-Year Freeze: We haven't seen a new nuclear reactor design approved by the NRC in over 50 years. Thiel argues we are "culturally broken" and risk-averse in the physical world.

• Manufacturing Comeback: Can AI bring factories back from China? Thiel says "yes, at the margins." If AI robotics lowers the cost of labor enough, the arbitrage of shipping jobs to Vietnam disappears.

The Deep Dive: Thiel’s core argument is that we don't need "a dozen miracles" in technology to fix the world; we just need to bypass the blockers. He believes AI can finally optimize the physical layer that has been stuck in the mud. But he offers a sobering warning: Technology alone can't fix "anti-industrial policies." While many of his peers are "exiting" (moving to Singapore or Switzerland), Thiel argues for staying and fighting the "right battles."

He views AI not just as a chatbot, but as the tool that finally makes American manufacturing and infrastructure competitive again—if we can get the regulators out of the way. 🔮 

Takeaway: The next massive AI winners won't just be software companies. They will be companies that use AI to navigate physical bureaucracy—think The Boring Company (tunnels) or next-gen nuclear startups. As Thiel puts it, the opportunity lies where things are "broken not because of science, but because of politics."

AI SaaS Founders

🚨Want Millions of Impressions For Your AI SaaS, Done For You?

At uncovernews.co, we specialize in getting AI SaaS products the attention they deserve through strategic influencer marketing campaigns designed to drive millons of impressions at the fraction of the cost!

Get Your AI Startup’s News or Product In Front of Millions Quickly

30-Second AI Play

Stop Paying for 5 Different AI Subscriptions

Tired of having a subscription for ChatGPT, another for Claude, and wanting to try Gemini 3 without a new bill?

You can route your prompts to the best model automatically. Here’s how:

1. Go to ChatLLM by Abacus.

2. Select the "Route LLM" feature.

3. Type your prompt (e.g., "Write a Python script for data vis" or "Generate a marketing image").

4. The system automatically routes code to the best coder (Claude/GPT) and creative tasks to the best creative model.

   Why it’s a win:

   You get access to GPT-4o, Gemini 3 Pro, Claude 3.5, and image generators like Flux and seedream all in one interface.

   💡 Pro tip: Use their "Humanize" feature to take the AI output and rewrite it to bypass the standard "robotic" tone detection.

   Check it out here

Other Relevant AI News!

🛍️ Black Friday online spending hit a record high as shoppers embraced new AI buying tools to find the best deals.

🎓 It is an "uneasy time" to be a CS student at Stanford, as AI coding capabilities force a rethinking of the curriculum.

🎵 A secretive startup called Klay has jumped into the fray of AI music generation to compete with Suno and Udio.

☁️ Amazon is challenging competitors by launching on-premises Nvidia AI factories to bring compute closer to enterprise data.

⚖️ Meet Soxton, the AI-native law firm shaking up the industry by replacing billable hours with automated legal tech.

Golden Nuggets

• 🧠 OpenAI is pivoting back to pre-training—admitting that scaling isn't dead, they just needed to work out the kinks.

• ⚛️ AI can optimize the physical world, but it can't fix regulatory stagnation without political will.

• 💎 Google's TPU bet has finally paid off, giving them a distinct advantage over Nvidia-dependent labs.

Would love to hear your thoughts on Code Red! Send me your thoughts by replying to this email (yes, I read them all :)

Until our next AI rendezvous,

Anthony | Founder of Uncover AI