Practically AI
Posts
🧠 AI Gets Real: Gmail, Grok, Deep-Fakes

🧠 AI Gets Real: Gmail, Grok, Deep-Fakes

Today in AI: Gmail gets Gemini, xAI teases its first coding agent, and deep-fakes cross the line.

January 09, 2026

👋 Hello hello,

Hope this Friday comes with a few emails and even fewer meetings.

This week’s AI updates have range.

xAI is teasing its first “vibe-coding” agent. Gmail just graduated into the Gemini era. And a viral deep-fake video has once again reminded everyone that generative media still needs a moral compass.

Let’s dig in.

🔥🔥🔥 Three big updates

1. Grok Build — xAI’s upcoming “vibe-coding” agent

Elon Musk’s xAI is building something called Grok Build — an AI agent that lives inside X and lets you build apps through natural conversation. Early testers describe it as “vibe coding,” where you sketch ideas in chat and the agent writes the logic, layout, and code in real time.

It’s a small peek at what a truly native AI development experience could look like on a social platform — where building software feels less like typing and more like thinking out loud.

2. Gmail joins the Gemini era

Google is bringing Gemini 3 into Gmail — and it’s more than just a search upgrade.

Gemini now acts as your personal inbox assistant: it can summarize threads, draft replies in your tone, pull deadlines from emails, and even build task lists automatically. The idea is to turn Gmail from a place where you “check messages” into a place where you “get things done.”

This update is rolling out gradually across Gemini apps and Google Workspace.

3. Deep-fakes aren’t a joke — and this viral video proved it

A “prank” video on X used AI-generated Black faces to depict people stealing from a store — and it spread fast.

Beyond being offensive, it’s a case study in how deep-fake tools can be weaponized for harm and misinformation. As synthetic video becomes easier to create, the conversation around AI ethics is no longer theoretical — it’s urgent.

🔥🔥 Two tools worth trying

1) 🎬 Sora 2 System Prompt for Pro-level Videos

This creative-director prompt (see attached image) lets you generate cinematic videos using ChatGPT and Sora 2. It guides you through story concept, tone, camera movement, lighting, and more — then outputs a fully-formatted JSON prompt ready for rendering. If you’re a filmmaker or designer, this is basically AI storyboarding on steroids.

2) 🌐 Awesome Gemini Prompts

An open-source library of ready-to-use Gemini prompts for research, coding, and writing tasks. Think of it as a community-built Prompt Hub for Gemini users — organized by use case and updated daily.

🔥 Things You Didn’t Know You Could Do With AI

Build interactive learning apps with NotebookLM + Gemini Canvas

Turn your NotebookLM projects into an interactive LLM learning system.

Here’s how:

1️⃣ Open Gemini Canvas and add your NotebookLM project as a source.
2️⃣ Ask Gemini to generate a learning loop with daily lessons, quizzes, and resources.
3️⃣ Each new chapter unlocks only when you pass the test.

You can use this setup for course creation, team training, or personal skill tracking — and it feels like an AI that grows with you.

Here are my results: https://gemini.google.com/share/b3f0f5953bf4

Here’s the prompt I used: https://prompt-genie.com/share/canvas-x-notebooklm-interactive-learning-app-pro/2dffa122-2bb0-4521-ba68-4d49bd43dfa1