Qualcomm Takes $6.7T Swing at NVIDIA

And much more…

The Latest in AI

Qualcomm just stormed into NVIDIA’s territory with a bold play most didn't see coming. After years of focusing on smartphones and wireless connectivity, the chipmaker announced new AI accelerator chips for data centers—and sent its stock soaring 15% in a single day.

The company isn't just testing the waters here. They're diving straight into a market projected to consume nearly $6.7 trillion in capital expenditures through 2030, according to McKinsey. While NVIDIA has commanded over 90% market share with its GPUs, Qualcomm is positioning itself as a serious alternative for AI inference—the process of actually running AI models, not training them.

What's making this move interesting:

So What?

The timing couldn't be better for challengers. OpenAI and other AI labs have been actively seeking alternatives after Scale AI's complications with Meta, and companies like AMD are already securing deals to diversify supply chains. Qualcomm entering the ring means more competition, potentially lower prices, and most importantly—options for companies building AI infrastructure.

This move also signals a maturation of the AI hardware market. When a company known for Snapdragon processors starts building chips that can handle the computational demands of ChatGPT-level models, you know the industry is reaching a new phase…

Innovation Showcase

Here's a scenario you'll definitely recognize. 74 unread emails and somewhere in there is the one message you actually need. Sound familiar?

Perplexity just made this problem disappear with Comet, their AI browser that can summarize any content in one click—including your overwhelming inbox.

Here's how to use Comet to summarize emails:

Power User Tip: You can also use Voice Mode next to the chat window to verbally ask Comet questions about your emails. Perfect for when you're juggling multiple tasks and need to quickly understand what's important without reading every word.

The beauty of this workflow is how it compounds. Instead of spending 10-15 minutes scanning emails each morning, you're getting the essentials in seconds. That time savings adds up quickly when you're processing dozens of messages daily!

But reading about AI ≠ actually using AI.

Here's the gap: You consume AI content, but you're not building meaningful products or workflows with it yet.

Inside our learning platform, you’ll have access to:

All created by industry veterans. Start with a free trial 👇

Industry Intel

OpenAI is reportedly working on a tool that could generate music from text and audio prompts. Think ChatGPT, but instead of writing your emails, it's composing guitar tracks and adding soundtracks to your videos.

According to The Information, the company has been collaborating with students from The Juilliard School to help create training data by annotating music scores. The goal? Build something that could add instrumental accompaniment to a vocal track or generate full soundscapes for video content.

This isn't OpenAI's first rodeo with music—they previously released Jukebox and MuseNet—but this new effort comes at a time when AI-generated content is flooding streaming platforms at unprecedented rates.

What we know so far…

So What?

The AI music space is getting crowded fast, with players like Suno, Udio, and ElevenLabs already offering music generation services. But OpenAI entering the arena brings significant weight—and significant controversy. The music industry is already grappling with AI-generated spam cluttering streaming platforms, with Deezer reporting that 28% of tracks uploaded to their service are now fully AI-generated.

Legal battles are inevitable. Suno and Udio are currently being sued by major record labels for alleged copyright infringement, and OpenAI has faced similar lawsuits from publishers like the New York Times. The company's approach to training data and licensing will likely determine whether this becomes a revolutionary tool for creators or another legal quagmire.

For now, there's no release date or confirmation on whether this will be standalone or integrated into ChatGPT. But if OpenAI does nail the execution and licensing, it could democratize music production in ways we haven't seen since affordable home recording became possible.

Quick Bites

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

Meta broke ground on a new AI-optimized data center in El Paso, Texas, capable of scaling to 1GW with 1,800 construction jobs at peak and a commitment to restore 200% of consumed water to local watersheds.

Japan's government formally requested OpenAI refrain from copyright infringement after Sora 2 generated videos featuring characters from anime series like One Piece, Demon Slayer, and Pokémon, calling manga and anime "irreplaceable treasures."

OpenAI revealed that over 1 million people weekly discuss suicide with ChatGPT, representing 0.15% of the platform's 800 million weekly active users, while GPT-5 now responds with "desirable responses" 65% more often than previous versions.

Mercor raised $350 million at a $10 billion valuation—five times its previous round—as the AI-driven recruiting platform pays over $1.5 million daily to more than 30,000 contractors averaging $85+ per hour for AI model training work.

Anthropic and Thinking Machines Lab released research stress-testing model specifications across 12 frontier LLMs, finding that high disagreement in 300,000+ value-tradeoff scenarios predicts specification violations by 5-13 times.

🧬 Antimattr - The world's first truly agentic operating system. Antimattr Node is a hand-held device powered by mattrOS with an AI app store enabling agentic actions across all supported apps and services.

🧠 Letta - Build stateful AI agents with advanced memory that learn and improve over time.

💑 Couple Prompt - Generate realistic AI couple photos from two individual portraits or existing photos.

The Neural Network

In the spirit of Halloween, we dare you to put this prompt into your favorite AI-image generator…

A little sneak peek…

Until we Type Again…

Thank you for reading yet another edition of Digestibly!