🥟 Chao-Down #132 Thousands of authors demand compensation for work used to train AI, DeSantis campaign used an AI-generated Trump voice in ad, OpenAI's ChatGPT now features "custom instructions"

Can we rely on big tech to safeguard AI?

Seven prominent artificial intelligence companies, including Amazon, Alphabet, Meta Platforms, Microsoft, and OpenAI, will unveil voluntary safeguards aimed at reducing abuse and bias in AI technology at an event hosted by President Joe Biden at the White House.

The transparency and security pledge requires companies to subject new AI systems to internal and external testing and invites external teams to probe for security flaws, discrimination, or risks to Americans' rights, health, and safety.

Additionally, the firms commit to sharing information to improve risk mitigation with governments, civil society, and academics, and they will incorporate virtual watermarks into AI-generated content to distinguish real from computer-created media. 

These are welcome moves but with anything, the devil is in the details. How will all this be implemented? We’ll have to see to find out.

-Alex, your resident Chaos Coordinator.

What happened in AI? 📰

Amazon, Meta Among Firms to Unveil AI Safeguards After Biden’s Warning (Bloomberg)

GitHub’s AI-powered Copilot Chat feature launches in public beta (The Verge)

DeSantis PAC uses AI-generated Trump voice in ad attacking ex-president (POLITICO)

Thousands of Authors Demand Compensation for Work That Trains AI (Bloomberg Law)

Why computer-made data is being used to train AI models (Financial Times)

K-pop's biggest music label HYBE looks to lift language barrier with AI (Reuters)

OpenAI’s ChatGPT chatbot now has ‘custom instructions’ to store your preferences (The Verge)

Always be Learnin’ 📕 📖

Your guide to SaaS product metrics (Substack)

Material Security’s Path to Product-Market Fit — Find Your Winning Idea by Selling Products That Don’t Exist Yet (First Round Review)

Rust or Mojo for the future of AI? (Mojo Dojo)

The Building Blocks of Generative AI (Substack)

Projects to Keep an Eye On 🛠

RayVentura/ShortGPT: 🚀🎬 ShortGPT - An experimental AI framework for automated short/video content creation. Enables creators to rapidly produce, manage, and deliver content using AI and automation. (Github)

microsoft/TypeChat: TypeChat is a library that makes it easy to build natural language interfaces using types. (Github)

psychic-api/rag-stack: 🤖 Deploy a private ChatGPT alternative hosted within your VPC. 🔮 Connect it to your organization's knowledge base and use it as a corporate oracle. Supports open-source LLMs like Llama 2, Falcon, and GPT4All. (Github)

The Latest in AI Research 💡

Dao-AILab/flash-attention: Fast and memory-efficient exact attention (Github)

Google DeepMind’s latest research at ICML 2023 (DeepMind)

Matrix Multiplication Using Only Addition (arxiv.org)

The World Outside of AI 🌎

TSMC delays Arizona plant because of US skilled worker shortage (Quartz)

The world’s cheapest Domino’s pizza is in inflation-hit India. It costs $0.60 (Arab News)

Why Apple is threatening to remove iMessage from the UK market (Quartz)

Space Tech Funding Comes Back Down To Earth (Crunchbase)

Homicides Are Down Nearly 10% This Year, but Car Thefts Are Skyrocketing (WSJ)

One Last Bite 😋