🥟 Chao-Down #265 Hugging Face launches open-source version of custom GPTs, FCC wants to criminalize AI robocall spam, Meta adds "AI-generated" label to images created by tools like Midjourney

Could we use AI to better understand the ancient world?

A team of student researchers has won a contest to reveal the content of a 2,000-year-old scroll buried by the eruption of Mount Vesuvius, using machine-learning algorithms. The scroll, part of the Herculaneum scrolls, contains a previously unknown philosophical work that discusses senses and pleasure.

The achievement opens the possibility of deciphering hundreds of other scrolls that could shed light on the ancient world. The contest, called the Vesuvius Challenge, was launched by computer scientist and tech investor Nat Friedman, who organized a $700,000 grand prize for creating an AI system that could read passages from the scroll. The winning team, from Egypt, Switzerland and the United States, used a technique that detected faint cracks in the scans of the scroll that formed the shapes of Greek letters.

It’s an amazing result that has captivated experts in classical texts as they now have more tools at their disposal to investigate what life was like many years ago.

-Alex, your resident Chaos Coordinator.

What happened in AI? 📰

The FCC wants to criminalize AI robocall spam (The Register)

Hugging Face launches open source AI assistant maker to rival OpenAI's custom GPTs (VentureBeat)

Why Quora isn’t useful anymore: A.I. came for the best site on the internet. (Slate)

Meta to add ‘AI generated’ label to images created with OpenAI, Midjourney and other tools (CNN Business)

Inside the Underground Site Where ‘Neural Networks’ Churn Out Fake IDs (404media.co)

Roblox releases real-time AI chat translator (The Verge)

Always be Learnin’ 📕 📖

The State of State of AI Report (airstreet.com)

Bad Actor, Good Advisor: Exploring the Role of Large Language Models in Fake News Detection (arxiv)

We need startups to fight prompt injection, the top LLM security risk | AI Venture Capital (signalfire.com)

Projects to Keep an Eye On 🛠

fetchai/uAgents: A fast and lightweight framework for creating decentralized agents with ease. (Github)

mulab-mir/song-describer-dataset: The Song Describer dataset is an evaluation dataset made of ~1.1k captions for 706 permissively licensed music recordings. (Github)

danswer-ai/danswer: Ask Questions in natural language and get Answers backed by private sources. Connects to tools like Slack, GitHub, Confluence, etc. (Github)

The Latest in AI Research 💡

Discovery of a structural class of antibiotics with explainable deep learning | Nature

Health-LLM: Large Language Models for Health Prediction via Wearable Sensor Data (arxiv)

CheXagent: Towards a Foundation Model for Chest X-Ray Interpretation (arxiv)

The World Outside of AI 🌎

YouTube and Facebook are beating TikTok and Instagram on site visits (qz.com)

Employees are spending the equivalent of a month's grocery bill on the return to the office–and growing more resentful than ever, new survey finds (Fortune)

Dartmouth Reinstates Standardized Testing for College Applicants (Bloomberg)

ESPN, Fox, Warner Bros. to launch joint sports streaming platform (CNBC)

Boomers Will Define the Ozempic Era (The Atlantic)

One Last Bite 😋