🥟 Chao-Down #206 Elon Musk to unveil new startup xAI's tech, ChatGPT traffic returns to near-peak levels, Microsoft breaks boundaries of small AI models, UK builds a $273 million AI supercomputer

Apparently this weekend, we’ll get to see what Elon Musk’s new AI startup xAI has been working on over the past several months. According to Musk, “In some important respects, [xAI] is the best that currently exists.”

Announced in early July, the company seeks to “understand the true nature of the universe” and has amassed a team of AI researchers and a war chest of GPU resources to accomplish that. Another big benefit being under the Musk umbrella is unfettered access to the Twitter (X) treasure trove of data.

What do you expect to come out from xAI?

-Alex, your resident Chaos Coordinator.

What happened in AI? 📰

ChatGPT Recovers to Near-Peak Levels (Similarweb)

UK to invest $273 million in Turing AI supercomputer (CNBC)

CIOs Assess Whether Microsoft’s AI Copilot Justifies Premium Price (WSJ)

AI ‘breakthrough’: neural net has human-like ability to generalize language (Nature)

Scarlett Johansson takes legal action against use of image for AI (The Guardian)

AI hype is built on high test scores. Those tests are flawed. (MIT Technology Review)

Microsoft pushes the boundaries of small AI models (Semafor)

Always be Learnin’ 📕 📖

Designing for safe and trustworthy AI (UX Collective)

How Uber design products (and why) (Substack)

What the executive order means for openness in AI (aisnakeoil.com)

Projects to Keep an Eye On 🛠

mailcorahul/auto_labeler: auto_labeler - An all-in-one library to automatically label vision data (Github)

seanoliver/audioflare: An all-in-one AI audio playground using Cloudflare AI Workers to transcribe, analyze, summarize, and translate any audio file. (Github)

baaivision/JudgeLM: An open-sourced LLM judge for evaluating LLM-generated answers. (Github)

The Latest in AI Research 💡

Exploring OCR Capabilities of GPT-4V(ision) : A Quantitative and In-depth Evaluation (arxiv)

Latent Consistency Models: Synthesizing High-Resolution Images with Few-step Inference (latent-consistency-models.github.io)

Eliciting Human Preferences with Language Models (arxiv)

The World Outside of AI 🌎

Parkinson's disease symptoms can now be tracked via Apple Watch apps (statnews.com)

Artificial Sweeteners: Are They Bad For Your Health? (Bloomberg)

How Aid in Dying Became Medical, Not Moral (The New York Times)

A Spotify playlist with 715 songs known to give people chills (qz.com)

Every trick Microsoft pulled to make you browse Edge instead of Chrome (The Verge)

How will TV’s streaming wars end? (ft.com)

One Last Bite 😋

The Beatles are releasing one final song thanks to AI!