• Chaos Theory
  • Posts
  • 🥟 Chao-Down #256 Tech giants partner with National Science Foundation to create federal AI hub, Apple steps up plans to bring AI to the iPhone, China fears drive an AI industrial complex

🥟 Chao-Down #256 Tech giants partner with National Science Foundation to create federal AI hub, Apple steps up plans to bring AI to the iPhone, China fears drive an AI industrial complex

Plus, Microsoft forms new GenAI team to work on small language models.

If 2023 was all about Large Language Models, Microsoft is betting that this year will be the year of SLMs or Small language Models. 

In an effort to create AI models that are cheaper to run yet just as capable on some reasoning tasks as more powerful models like GPT-4, Microsoft has created a new multi-disciplinary GenAI team featuring people from Azure, Microsoft Research, and more. The team will ultimately report to CTO and EVP of AI Kevin Scott.

To combat costs and minimize latency, Microsoft sees small language models like their open-sourced Phi models as key to realizing a Copilot-driven future. By developing their own models, Microsoft can also lessen its reliance on OpenAI which until now has been its primary source for new AI.

Especially as more AI-powered Windows PCs start coming to market, expect Phi and other SLMs to make a big splash in 2024.

Satya Nadella on stage at Microsoft Ignite 2023 announcing Phi-2.

-Alex, your resident Chaos Coordinator.

What happened in AI? 📰

China fears drive AI industrial complex (axios.com)

Before AI comes for your job, it wants your job applications (qz.com)

Apple boosts plans to bring generative AI to iPhones (ft.com)

Microsoft’s Journey to $3 Trillion Aided by Cloud Growth, AI Optimism, and Big Deals (WSJ)

OpenAI, Nvidia, NSF, NASA And More Partner On Federal Program To Increase Access To AI Resources (Forbes)

Pope Francis, victim of AI, warns against its 'perverse' dangers (Reuters)

Always be Learnin’ 📕 📖

Hiring ONLY seniors is the worst policy in the software industry (Substack)

Comparison of AI Models across Quality, Performance, Price | ArtificialAnalysis.ai

Evaluations are all we need - by Rohit Krishnan (strangeloopcanon.com)

Projects to Keep an Eye On 🛠

Python & JavaScript Libraries (Ollama Blog)

Vchitect/Vlogger: Make Your Dream A Vlog (Github)

FlagOpen/TACO: Topics in Algorithmic COde generation dataset (Github)

The Latest in AI Research 💡

Neural MMO 2.0: A Massively Multi-task Addition to Massively Multi-agent Learning (arxiv)

Hidden Flaws Behind Expert-Level Accuracy of GPT-4 Vision in Medicine (arxiv)

Supervised Fine-tuning in turn Improves Visual Foundation Models (arxiv)

The World Outside of AI 🌎

Baby boomers hoard homes, leaving fixes for younger buyers (qz.com)

Netflix is turning into cable TV (The Verge)

Moody's: Shell companies have millions of 'red flags' including directors aged 123 and older (Fortune)

Unhappy at Work? Quit Quitting Costs US $1.9 Trillion in Productivity (Bloomberg)

Instant is in, fancy is out. Nothing stays the same for long in the world of home-brewed coffee (qz.com)

A fresh look at risks for developing young-onset dementia (Harvard Health)

One Last Bite 😋

How are real estate agents using AI?