Meta's Llama 4 family debuts

Together with

Welcome Humans🤖,

Here is what we have Today:

• 🎤 Amazon's Nova Sonic makes AI sound like Barry White

• ⏰ GPT-5 delayed: OpenAI pivots to release o3 first

• 🧠 DeepSeek cracks AI's logic problem

• 🦙 Meta unleashes Llama 4 & 2T parameter "Behemoth"

• 💼 New Job Opportunities

AMAZON

💥 Amazon’s AI Just Got Louder, Smarter, and Cheaper

Amazon just dropped two big updates in the AI world — a voice model called Nova Sonic and a boosted version of its video model, Nova Reels 1.1. Together, they’re making waves in speech and video generation.

What’s new?
🗣️ Nova Sonic turns voice input into natural-sounding speech in just 1.09 seconds. That’s faster — and smoother — than OpenAI’s voice models.

It’s also better in chaos:

• 4.2% word error rate across multiple languages

• 46.7% more accurate than GPT-4o in noisy, multi-speaker settings

🎞️ Nova Reels 1.1 takes video content further:

• Generates 2-minute videos, way up from previous limits

• Supports manual shot-by-shot creation or quick auto mode

• High-quality output with more creative control

Where can I try this?
Both models are now on Amazon Bedrock. The kicker?
👉 Nova Sonic is 80% cheaper than OpenAI’s offerings.

Why it matters
Amazon is no longer just dabbling in GenAI — it’s pushing hard. Between Nova Sonic, Reels 1.1, Alexa+, and its Act browser agent, Amazon’s building a full-stack AI alternative that’s cheaper, faster, and developer-ready.

Together with Artisan

Hire an AI BDR to Automate Your LinkedIn Outreach

Sales reps are wasting time on manual LinkedIn outreach. Our AI BDR Ava fully automates personalized LinkedIn outreach using your team’s profiles—getting you leads on autopilot.

She operates within the Artisan platform, which consolidates every tool you need for outbound:

• 300M+ High-Quality B2B Prospects

• Automated Lead Enrichment With 10+ Data Sources Included

• Full Email Deliverability Management

• Personalization Waterfall using LinkedIn, Twitter, Web Scraping & More

Book a demo to see what Ava can do.

META

🧠 Scout, Maverick & Behemoth: Meta’s New Llama 4 AI Team

Meta just unleashed its Llama 4 model family — and it’s loaded with brainpower, multimodal tricks, and jaw-dropping context length.

Let’s break it down:

Scout, the entry-level model, punches way above its weight. With 109B parameters and a 10 million token context window, it outperforms Google’s Gemma 3 and Mistral’s latest — and it runs on a single H100 GPU.

Maverick is where things get wild. This 400B parameter beast handles 1 million tokens, outshines GPT-4o and Gemini 2.0 Flash, and still manages to be more cost-efficient. Yes, you read that right.

But Meta’s not done flexing.

They’re previewing a giant-in-the-making: Behemoth, a 2 trillion parameter teacher model still in training. Early whispers say it outperforms GPT-4.5, Claude 3.7, and Gemini 2.0 Pro. The heavyweight title might be up for grabs.

What makes these models so fast?

It’s all thanks to MoE (mixture-of-experts) architecture — only a few experts activate per token, cutting down on computation while keeping the quality high.

Where can you use them?

Scout and Maverick are live now — ready for download or accessible via Meta AI in WhatsApp, Messenger, and Instagram.

Why should you care?

After DeepSeek R1 shook up the open-source scene earlier this year, Meta needed a comeback. Llama 4 is that comeback — with huge leaps in efficiency, context handling, and multimodal muscle.

But one question still lingers:
Will these models feel next-gen… or just benchmark better?

Trending Today

• Onyx Solar has unveiled its solar panels that integrates photovoltaic cells within the glass panels allowing buildings to generate renewable energy while maintaining the structural and aesthetic properties of conventional glass. These panels can be customized to match various architectural designs offering a range of transparency levels, colors and finishes.

• OpenAI is now releasing both its o3 reasoning model and a next-gen o4-mini in the coming weeks, after initially scrapping o3’s public debut. The shift comes as GPT-5 faces a slight delay, now expected in a few months, to allow smoother integration of advanced reasoning capabilities and ensure infrastructure readiness.

• DeepSeek, in collaboration with Tsinghua University, has introduced a new approach combining two powerful reasoning techniques to help LLMS tackle complex queries with greater accuracy and speed. The resulting DeepSeek-GRM models outperform existing methods by aligning outputs with human preferences and enhancing logical reasoning.

• Ex-OpenAI CTO Mira Murati’s secretive Thinking Machines Lab has quietly brought on two OpenAI alums as advisors, including Bob McGrew, the ChatGPT-maker’s former chief research officer. All told, Murati’s roster now includes nearly 20 ex-OpenAI staffers, or about half of the startup’s total lineup. It has yet to reveal exactly what it’s working on

Recommended Reading
If WE had to recommend other newsletters

AI Tool Report is an Newsletter, we enjoy reading daily. AI Tool Report delivers top techniques on how AI can transform your business filled with practical tips, real-world examples.

A Smart Bear is for people who love thinking about strategy, startups, product, marketing, decision-making, and founder psychology.

Job Opportunities

• Mistral AI - Staff Software Engineer, JS - Apply

• Synthesia - Business Analyst, GTM - Apply

• DeepMind - Forward Deployment Engineer - Apply

• Union- Software Engineer - Apply

• Tempus - Principal Quality Engineer - Apply

AI News

Ideas? Comments? Complaints?

We read your emails, comments and poll replies daily.

Get the most important AI, tech, and science news in a free daily email.
New Here? Subscribe!
Sponsorship Slots Open for April and Reach over 800+ active readers. (Now 40% off) 🤯

Until next time, Stay Informed!