The Llama Family Grows

Meta launched Llama 4, but not without some controversy…

In this edition we’ll be covering…

• Meta’s new family of models

• A breakdown on Google’s multimodal search

• A tutorial on how to use 21st.dev for UI inspiration

• 5 more piece of AI news

• 3 AI tools to keep you productive

• And much more…

The Latest in AI

Meet Llama 4

It was a quiet Saturday night until…

Meta dropped Llama 4, and suddenly open-source AI is stealing the spotlight. This new family of models comes in three distinct flavors: Scout, Maverick, and the still-in-training Behemoth.

What’s the buzz? 

• Scout: Boasts an industry-leading 10 million token context window—enough to process “War and Peace” multiple times in a single prompt.

• Maverick: With 400 billion total parameters, it outperforms GPT-4o on coding and reasoning benchmarks. 

• Behemoth: Still in development, this model packs nearly 2 trillion parameters and is showing promise in early STEM evaluations, surpassing GPT-4.5, Claude 3.7, and Gemini 2.0 Pro.

What sets Llama 4 apart is its native multimodality—trained from the ground up on text, images, and video.

Its revolutionary Mixture of Experts (MoE) architecture activates only a fraction of parameters per token, which in English, means greater efficiency.

So What?

Meta’s surprise release of Llama 4 signals a significant shift in the AI landscape, challenging the dominance of closed-source models. However, the launch hasn’t been without controversy.

Some AI experts have expressed skepticism, noting that initial user experiences don’t fully align with Meta’s performance claims.

You can mess around with the new Llama 4 models through Instagram, Whatsapp, Messenger, and Groq.

Together with Morning Brew

The newsletter every professional should be reading

There’s a reason Morning Brew is the gold standard of business news—it’s the easiest and most enjoyable way to stay in the loop on all the headlines impacting your world.

Tech, finance, sales, marketing, and everything in between—we’ve got it all. Just the stuff that matters, served up in a fast, fun read.

Look—over 4 million professionals start their day with Morning Brew’s daily newsletter, and it only takes 5 minutes to read. Sign up for free and see for yourself!

Check it out

Industry Intel

Google’s AI Mode Sees the World

Google just gave Search a serious upgrade—introducing multimodal capabilities to AI Mode that let you snap a pic and ask away.

AI Mode understands the entire scene, including how items relate to each other, their materials, colors, shapes, and layout. Using a technique called “query fan-out,” it goes beyond a single search by asking multiple questions about both the full image and the specific objects within it, which results in more detailed answers compared to simple Google Search.

Here’s the breakdown:

• See It, Ask It: AI Mode now understands images. Snap a photo or upload one, and it can analyze the entire scene—objects, materials, colors, you name it.

• AI Mode understands the entire scene, including how items relate to each other, their materials, colors, shapes, and layout. Using a technique called “query fan-out,” it goes beyond a single search by asking multiple questions about both the full image and the specific objects within it, which results in more detailed answers compared to simple Google Search.

• The feature also lets you ask follow-up questions to narrow down your search, such as “I’m looking for a quick read. Which one of these recommendations is the shortest?”

So What?

Google’s move blurs the lines between visual and text search, making interactions more intuitive.

The company is rolling out AI Mode to millions more Labs users in the U.S., expanding access beyond just Google One AI Premium subscribers.

As AI Mode evolves, expect search to become less about keywords and more about context—whether it’s text or images. I’ve been waiting for multimodal search for a long time, even within ChatGPT, so I’m glad Google is taking a step here.

Tool Spotlight

Copy That! Steal Any UI From 21st.dev

If you’ve ever wanted to “borrow” a beautiful UI component and feed it into an AI tool to customize, this one’s for you.

21st.dev is a repository that lets you visually select and copy components—like buttons, cards, or forms—curated by the community to build React apps. Think copy-paste, but built for developers, designers, and AI workflows.

Here’s how you can integrate it into any software you are building, and this works even better if you’re using an AI-powered development tool like Cursor, bolt, or Replit:

1. Navigate to 21st.dev.

2. Select a component that catches your eye (example here).

3. Open the component and hit Copy Prompt.

4. Feed it into your favorite AI coding assistant and see what happens!

🧠 Pro Tip: This is a killer workflow for frontend developers and indie builders looking to go from idea → working UI fast.

Quick Bites

Stay updated with our favorite highlights, dive in for a full flavor of the coverage!

Microsoft is making changes to Copilot that will allow the artificial intelligence assistant to complete online tasks for you. Using “simple chat prompts,” these “Actions” can be performed in the background while you work on other tasks.

OpenAI released their Academy to help train people in AI for free.

AI image generator Midjourney released its first new model in quite some time, dubbed V7, it's a ground-up rework that is available in alpha to users now.

A new UN Trade and Development (UNCTAD) report projects the global AI market will soar from $189 billion in 2023 to $4.8 trillion by 2033 – a 25-fold increase in just a decade.

GitHub announced GitHub Copilot Pro+, a new individual tier for developers who want to take their coding experience to the next level. It features exclusive access to the latest models like GPT-4.5.

Trending Tools

🗺️ MagicTrips - AI itinerary generator.

🤖 Composio MCP - Instantly connect to 100+ managed MCP Servers with built-In auth.

🧠 Cogmi - tracks your brain's natural peaks and dips—so you can work smarter, waste less energy, and reduce burnout.

The Neural Network

According to our latest poll, most of you will be using GenSpark to book travel!

But wait, there's more: We just scored access to Manus after camping out on that waitlist forever. Get ready for our full breakdown coming soon! 😏

In other news, I went to CrewAI’s summit in NY last Friday, and all I can say is… the stampede of AI agents heading our way is both impressive and slightly terrifying. Prepare your workflows accordingly (and follow along as we help you build them)!

Until we Type Again…

Thank you for reading yet another edition of Digestible AI. Be sure to give us a follow on X, Instagram, and LinkedIn too!

If you have any suggestions or specific feedback, simply reply to this email. Additionally, if you found this insightful, don't hesitate to engage with us on our socials and forward this over to your friends!