🥟 Chao-Down #198 Researchers find ChatGPT is better at following depression treatment guidelines than doctors, Stanford introduces the Foundation Model Transparency Index, Doctors give ChatGPT tips

As we’re all learning how to talk with AI, people are realizing that being polite and emphasizing your words with all caps, is the best way to get the output you want.

Below we can see a prompt coming out from DALL-E 3, the newest image generation model from OpenAI. Based off the use of the words “please” and the all caps phrase “DO NOT UNDER ANY CIRCUMSTANCES”, we can see a glimpse at how experts at OpenAI try to wrangle the right behavior from these generative AI models.

It just goes to show that prompt engineering is still a very much in-demand skill and will be central to how we interact with the future of AI

-Alex, your resident Chaos Coordinator.

What happened in AI? 📰

ChatGPT may be better than a GP at following depression guidelines – study (The Guardian)

How to Use ChatGPT for Health: Doctors, Professionals Give Tips (Bloomberg)

AI Is Becoming a Band-Aid over Bad, Broken Tech Industry Design Choices (Scientific American)

Scale AI wants to be America’s AI arms dealer to compete with China (The Washington Post)

Health providers say AI chatbots could improve care. But research says some are perpetuating racism (qz.com)

Protecting customers with generative AI indemnification (Google Cloud Blog)

Always be Learnin’ 📕 📖

Embeddings: What they are and why they matter (simonwillison.net)

Parallel Bets, Microsoft, and AI Strategies (MatthewBall.co)

Projects to Keep an Eye On 🛠

Introducing The Foundation Model Transparency Index (stanford.edu)

jiangjiechen/auction-arena: Source code for our paper: "Put Your Money Where Your Mouth Is: Evaluating Strategic Planning and Execution of LLM Agents in an Auction Arena" (Github)

YiVal/YiVal: Your Auto-tuning Assistant for GenAI Applications (Github)

The Latest in AI Research 💡

DiffuSeq-v2: Bridging Discrete and Continuous Text Spaces for Accelerated Seq2Seq Diffusion Models (arxiv)

AgentTuning: Enabling Generalized Agent Abilities for LLMs (arxiv)

A taxonomy and review of generalization research in NLP (Nature Machine Intelligence)

The World Outside of AI 🌎

Why hasn't the new conflict in the Middle East increased US gas prices at the pump? (qz.com)

Exclusive: Nvidia to make Arm-based PC chips in major new challenge to Intel | (Reuters)

‘Reddit can survive without search’: company reportedly threatens to block Google - (The Verge)

48% of Gen Z will holiday shop using social media (CNBC)

“Netflix effect” is back as studios license old shows to competitors again (Ars Technica)

Thiel’s Unicorn Success Is Awkward for Colleges (Bloomberg)

One Last Bite 😋