🥟 Chao-Down #174 Intel's new GenAI CPU for laptops, CZI builds massive GPU cluster for medical research, Google's Bard gets "extensions", GE creates AI-powered ultrasounds

Will personal computing soon have its own dedicated generative AI chip?

Intel just announced that its upcoming “Meteor Lake” chip will be able to run a handful of generative AI services directly on a person’s laptop, which would eliminate the need for using external, third-party cloud services. This new CPU will be part of Intel’s updated Core Ultra branding and released on December 14th.

This will be Intel’s first ever Neural Processing Unit (NPU) made specifically for AI acceleration and will enable programs like AI chatbots to operate directly on someone’s own hardware. This is especially relevant for people who want to keep their sensitive data on their own devices instead of sending it to the cloud.

Will Intel be able to deliver? We’ll have to see to find out.

-Alex, your resident Chaos Coordinator.

What happened in AI? 📰

Chan Zuckerberg Initiative is building an expensive new AI GPU cluster for medical research (The Verge)

Google’s Bard chatbot can now find answers in your Gmail, Docs, Drive (The Verge)

GE HealthCare is working on AI-powered ultrasounds to combat pediatric and maternal mortality rates (Engadget)

Get Ready for ChatGPT-Style AI Chatbots That Do Your Boring Chores (WIRED)

Intel pitches the 'AI PC' at software developer event (Reuters)

Always be Learnin’ 📕 📖

AI Regulation - There have been multiple call to regulate AI. It is too early to do so. (Elad Gil’s blog)

Why are CNN, FOX, and US websites visually so aggressive? (UX Collective)

What We Learned Building A Call Summary App with GPT-4 (Medium)

Projects to Keep an Eye On 🛠

IBM/ModuleFormer: ModuleFormer is a MoE-based architecture that includes two different types of experts: stick-breaking attention heads and feedforward experts. (Github)

15 times Faster than Llama 2: Introducing DeciLM - NAS-Generated LLM with Variable GQA (Deci.ai)

ise-uiuc/Repilot: Repilot, a patch generation tool introduced in the ESEC/FSE'23 paper "Copiloting the Copilots: Fusing Large Language Models with Completion Engines for Automated Program Repair" (Github)

The Latest in AI Research 💡

Pushing Mixture of Experts to the Limit: Extremely Parameter Efficient MoE for Instruction Tuning (arxiv)

HAMUR: Hyper Adapter for Multi-Domain Recommendation (arxiv)

Differentiable JPEG: The Devil is in the Details (christophreich1996.github.io)

The World Outside of AI 🌎

Elon Musk: X/Twitter to charge all users with monthly subscription fees (axios.com)

Student loans: Women bear the brunt of the student debt crisis (Yahoo)

US Workers Strikes Reach a Record High Since 2000 (Business Insider)

How Covid-19 vaccine misinformation stayed one step ahead of Facebook (Vox)

5:01 and Done: No One Wants to Schmooze After Work (WSJ)

Fintech faces its reckoning: It’s only a matter of time until the house of cards collapses (TechCrunch)

One Last Bite 😋