Automate Everything!

Friends,

Welcome back! ☕

Today is all about leverage. We are moving past the "wow" phase of AI and into the "work" phase. From NVIDIA’s ruthless efficiency strategies to new tools that learn for you, the theme of this week is clear: automate the routine so you can dominate the complex.

Let’s dive in.

But first EverMemOS is the missing memory layer for AI agents — persistent, reliable, and built for real-world workflows.

1. NVIDIA’s "Insane" New Standard

In a leaked all-hands meeting on November 21, 2025, NVIDIA CEO Jensen Huang didn't mince words: Automate everything!

Following a massive Q3 ( $57B revenue / $4T market cap), Huang declared that not using AI for feasible tasks is "insane." While this sounds aggressive, he reassured staff that this is about growth, not replacement.

NVIDIA grew from 29,600 to 36,000 employees (super few people.) this fiscal year and is looking to add 10,000 more. His advice? Use tools like Cursor even if they are flawed.

Imperfect automation is better than none.

• On X: Tech circles are buzzing (~70% positive sentiment). The consensus? This is the productivity boost the industry needs. But there are also voices against it.

• The advise by Jensen: Start with the "hated tasks"—invoicing, scheduling, basic code.

• The Goal: Scale automation to achieve 30-50% efficiency gains, leaving humans to handle the nuance.

You may want to leverage this AI automation cycle (I put all my thoughts together):

How to Start (Real-world Examples):

I’ve been living this philosophy personally. For months, I’ve refined a Personal Finance Tracker. I simply run a command, and I get a perfect, categorized visualization of expenses, fixed costs, and income. No manual entry.

Here are 5 ways you can start today:

2. Learning at the Speed of Sight

Google Gemini just changed the game with Interactive Images.

We already knew Gemini Pro could summarize complex papers, but now it creates interactive infographics.

Imagine looking at a visual for two minutes and understanding a full scientific paper.

Why this matters:

• Guided Learning: It storylines information. Want to know how a neuron works? It explains it, draws it (or pulls from Shutterstock), makes it interactive, and quizzes you.

• Efficiency: This reduces the "time-to-understanding" drastically.

Note: For specialized logic games like Chess, stick to dedicated engines (Chess.com), but for general concepts, this is a beautiful time to be alive.

3. The New Coding King: Claude Opus 4.5

I know I sound like a broken record, but Claude Opus 4.5 is currently the strongest coding model in existence.

I’ve tested it extensively against Gemini 3. While Gemini is good, Opus 4.5 has a distinct edge: Task Adherence. 

It creates a to-do list based on your prompt and systematically follows up on it.

To help you get the most out of it, I’ve not only scoured the internet (What is the Anthropic team even suggesting?) but tested these methods myself.

Here are the best recommendations for working with Opus 4.5:

Subscribe to Premium to See the Rest