Stanford Unveils 3D Chip
On Christmas Eve, researchers unveiled a chip that builds upward instead of outward. It’s being called “the Manhattan of computing,” and it might be the breakthrough that makes AI’s next leap possible.
📰 The Rundown
🏗️ Stanford Unveils 3D Chip That Could Break AI’s Biggest Bottleneck
➡️ The move: Engineers from Stanford, Carnegie Mellon, MIT, and UPenn have built the first commercial 3D computer chip that stacks memory and computing vertically, like floors in a skyscraper. Unlike traditional flat chips where data crawls across crowded horizontal paths, this design uses vertical wiring as “high-speed elevators” that move massive amounts of data quickly. In hardware tests, the prototype already beats 2D chips by 4x. Simulations show potential for 12x gains on AI workloads.
⚡ Why it matters: AI models like ChatGPT and Claude constantly shuttle data between memory and processing units. On flat chips, this creates a “memory wall” where processors wait for data. The new design bypasses that bottleneck entirely. The researchers project a path to 100x to 1,000x improvements in energy efficiency. Most significantly, the chip was manufactured at a U.S. commercial foundry, proving this isn’t just lab curiosity.
🎯 Your takeaway: The infrastructure that powers AI is about to change fundamentally. Faster, cheaper, more efficient AI isn’t just a software problem anymore.
📉 AI Blamed for 55,000 U.S. Job Cuts in 2025
➡️ The move: According to consulting firm Challenger, Gray & Christmas, artificial intelligence was cited as the reason for nearly 55,000 U.S. layoffs this year. The biggest cuts came from Amazon (14,000 corporate roles), Microsoft (15,000 across multiple rounds), and Salesforce (4,000 customer support positions). Total job cuts hit 1.17 million in 2025, the highest since the pandemic.
⚡ Why it matters: Some experts argue companies are “AI-washing” their layoffs, using automation as a convenient explanation for corrections after pandemic-era overhiring. But the pattern is undeniable: firms are restructuring around AI capabilities. Salesforce CEO Marc Benioff said AI now handles 50% of the company’s customer service workload.
🎯 Your takeaway: The 55,000 figure represents only jobs where companies explicitly blamed AI. The real number affected by automation is higher. The professionals who survive these cuts aren’t avoiding AI. They’re the ones who learned to work with it.
👑 Google Ends 2025 on Top of the AI Race
➡️ The move: Google entered 2025 perceived as playing catch-up to OpenAI. It’s ending the year on top. Gemini’s monthly active users grew 30% between August and December, while ChatGPT grew just 15%. OpenAI CEO Sam Altman declared a “code red” as the company scrambled to match Google’s Gemini 3 models. Google Cloud revenue hit $15.1 billion in Q3, up 34% year over year.
⚡ Why it matters: The AI race isn’t a one-company show. Google’s Gemini integration across Search, Workspace, and Android gives it distribution that OpenAI can only dream of. Analyst Gene Munster predicts Google will be the best performing Mag 7 stock in 2026.
🎯 Your takeaway: Competition means better tools for everyone. Both platforms have free tiers. If you’ve only tried one, the holidays are a good time to experiment with the other.
🔧 Tool Spotlight: Read AI
Read AI analyzes your meetings and gives you personalized feedback on how you communicate. It goes beyond basic transcription to track sentiment, engagement levels, and speaking patterns across all your calls.
What makes it different: The “Speaker Coach” feature identifies trends across your meetings and flags issues like speaking too fast, using filler words, or dominating conversations. It creates a 2-minute highlight reel of key discussion points and tracks action items over time.
Best for: Anyone who spends significant time in meetings and wants data-driven feedback on their communication effectiveness.
Pricing: Free plan includes basic reporting and 5 meeting reports per month. Paid plans start at $19.75/month for unlimited meetings and advanced analytics.
👉 Try it: Connect Read AI to your next Zoom or Teams call and review your speaker analytics afterward.
✨ Try This Today: The First Draft, Not Final Draft Mindset
Most people fall into one of two traps with AI output: they either accept it verbatim (risky) or reject AI entirely because the first output wasn’t perfect.
The insight: AI is your drafting partner, not your ghostwriter. The goal isn’t to remove yourself from the process. It’s to start at 60% instead of 0%.
The technique: Use AI to generate a rough draft, then apply your expertise with this four-step refinement:
- Fact Check — Verify statistics, quotes, and specific claims
- Voice Check — Does this sound like you or like generic AI?
- Value Check — Cut filler, add your unique perspective
- Action Check — Is the ask crystal clear?
The mantra: “AI gets me to the starting line faster. I still run the race.”
📚 Go Deeper: Expanded Lesson with Exercises
✨ The Wire
🔗 Fujitsu launched Physical AI 1.0 in collaboration with NVIDIA, integrating agentic AI with enterprise procurement automation. Fujitsu
🔗 Shadow AI continues to grow, with over 50% of enterprise AI app adoption now unsanctioned. 22% of files uploaded to GenAI tools contain sensitive data. Help Net Security
🔗 Ex-Facebook privacy chief predicts 2026’s AI focus will shift from capability to efficiency. “We run our brains on 20 watts. We don’t need gigawatt power centers to reason.” CNBC
Neural Notes — AI that amplifies your value, not replaces it.
