🥟 Chao-Down #211 Humane debuts a $699 AI Pin as a digital assistant without a screen, OpenAI seeks partnerships to generate AI training data, NVIDIA aims to build AI chips for China to avoid US ban

Self-driving cars and large language models? OpenAI is betting that combo will work.

After the recall of the GM Cruise fleet and suspension of operations on California roads, you would think that the investment pace for autonomous vehicles would stop. Apparently, that’s not the case and some startups are looking to tackle the problem in new ways.

Ghost Autonomy, a company building autonomous driving software for automaker partners, announced that it would start using multimodal large language models — AI models that can understand text as well as images — in self-driving.

To achieve this, Ghost has partnered with OpenAI through the OpenAI Startup Fund to gain early access to OpenAI models and compute from Microsoft Azure alongside a $5 million investment.

Will this technique help solve self-driving’s reliability and trust problems? We’ll have to see to find out.

-Alex, your resident Chaos Coordinator.

What happened in AI? 📰

Nvidia develops AI chips for China in latest bid to avoid US restrictions (ft.com)

OpenAI seeks partnerships to generate AI training data (Reuters)

Humane Debuts The $699 AI Pin, A Pocket-Sized Digital Assistant Without A Screen (Forbes)

Meta taps Hugging Face for startup accelerator to spur adoption of open source AI models (TechCrunch)

Ghost, now OpenAI-backed, claims LLMs will overcome self-driving setbacks — but experts are skeptical (TechCrunch)

China AI Startup Stockpiled 18 Months of Nvidia Chips Before Ban (Bloomberg)

Always be Learnin’ 📕 📖

An intuitive introduction to text embeddings (Stack Overflow)

10 hard-to-swallow truths they won't tell you about software engineer job (mensurdurakovic.com)

From Image Adjustments to AI: Photoshop Through the Years | WDD (webdesignerdepot.com)

Projects to Keep an Eye On 🛠

laiyer-ai/llm-guard: The Security Toolkit for LLM Interactions (Github)

Introducing Adept Experiments - Using AI workflows to delegate repetitive tasks (Adept)

yangjianxin1/LongQLoRA: LongQLoRA: Extent Context Length of LLMs Efficiently (Github)

The Latest in AI Research 💡

Cheating Depth: Enhancing 3D Surface Anomaly Detection via Depth Simulation (arxiv)

Multimodal ChatGPT for Medical Applications: an Experimental Study of GPT-4V (arxiv)

Battle of the Backbones: A Large-Scale Comparison of Pretrained Models across Computer Vision Tasks (arxiv)

The World Outside of AI 🌎

How big is science’s fake-paper problem? (nature.com)

It’s U.S. vs. China in an Increasingly Divided World Economy (WSJ)

How Does the World’s Largest Hedge Fund Really Make Its Money? (The New York Times)

Rise in Infant Mortality Shows US Needs to Take Better Care of Moms (Bloomberg)

Google's Waze will now warn you about roads prone to car accidents (CNBC)

One Last Bite 😋

I gave this demo during the last Semantic Kernel Office Hours on how to build AI Agents using SK and AutoGen, a popular open-source framework from Microsoft Research!