👋 Hey, Fellow Humans! Welcome back to Practical AI News, where we sift through the AI madness so you don’t have to.
This week? We’ve got a new contender for the AI throne—Anthropic’s Claude 3.7 Sonnet. It’s claiming best-in-class coding skills, but Grok-3 is out here quietly dunking on it in real-world tests. Meanwhile, Google and Alibaba are fighting over who has the best AI-generated videos, and DeepSeek is under pressure to drop its next-gen model sooner than expected.
So, who’s really the new AI GOAT1? Let’s dive in.
Chapters
00:00 The Rise of AI Coding Contenders Claude 3.7 vs Grok-3
03:55 The Challenges of AI in Programming
10:01 Emerging AI Technologies and Trends
13:22 Veo 2 and Wan2.1 Video AI Models
15:34 AI Innovations in Everyday Life
18:49 Outro
🤖 Claude 3.7 Sonnet: Smartest AI Yet?
Anthropic dropped Claude 3.7 Sonnet, boasting:
✔️ Hybrid reasoning mode (quick thinking or "extended" mode for deeper analysis)
✔️ Best-in-class coding performance (according to Anthropic’s own charts 👀)
✔️ Available in Cursor IDE for programmers
But here’s the plot twist…
⚡ Grok-3 actually outperformed it in real-world coding tests.
🔍 Extended thinking mode? Great in theory, but overcomplicated Jan’s code so much he didn’t even want to understand it.
Verdict? If you're a dev, Claude 3.7 is powerful but pricey—DeepSeek R1 + Claude 3.5 might actually give you better value.
🚀 Grok-3’s Wild Ride: From Controversy to #1
Elon Musk’s Grok-3 is officially the most downloaded AI app after making it free.
🔥 Better than Claude for some coding tasks.
🔥 Less censorship than competitors.
🔥 Deep Search is a game-changer—pulls fresh data from X (formerly Twitter).
But…
❌ Still unavailable via API.
❌ Some features locked behind "Big Brain" mode.
❌ Servers getting overloaded—cue the "Try again later" errors.
🎥 AI Video Wars: Google Veo 2 vs. Alibaba Wan2.1
First, Google Veo 2 claimed to be the best AI video generator.
Then, Alibaba said, “Hold my GPU,” and dropped Wan2.1, which beat everything in the VBench benchmark (but—oops!—Veo 2 wasn’t included 🤔).
📢 Which is actually better?
🎬 Google Veo 2—not open-source, but stunning sample clips. Available on VideoFX (might need VPN to try it) and Vertex AI.
🎬 Alibaba Wan2.1—open-source, top of the VBench leaderboard, but some creepy-looking samples.
Try them out and judge for yourself!
💰 AI’s Billion-Dollar Race: Who’s In & Who’s Out?
🚀 Apple is going all-in on AI, announcing a $500B investment in the U.S. over the next four years.
🤯 Meta is rumored to be building a $200B AI data center (because, clearly, one Metaverse wasn’t enough).
💸 Meanwhile, Microsoft is backing off, canceling some U.S. data center leases. Change of heart, or just playing it safe?
🔥 Quick Hits: AI Roundup
🔬 An AI can now diagnose multiple diseases from a single blood test. Future doctors, take notes.
🌊 China built an underwater AI Computing Center, cooled by seawater. (Someone watched too much sci-fi.)
💻 DeepSeek R2 is dropping earlier than expected—competition is heating up!
📱 Grok-3 hit #1 on the App Store after xAI made it free. Elon’s fanbase strikes again.
🚀 What Do You Think?
Which AI is actually the best right now—Claude 3.7, Grok-3, DeepSeek, or something else?
Let us know in the comments, and don’t forget to subscribe to Practical AI News for your weekly AI reality check! 🎙️✨
GOAT - Greatest Of All Time in their field
Share this post