- Machine Learning Times
- Posts
- The Generative A.I. Brief Issue #9
The Generative A.I. Brief Issue #9
Where professionals go to catch up on their A.I. reading of the week and summer.
Michael Kevin Spencer
July 22, 2023
Hey Everyone,
This Newsletter is a new initiative you can sign up for free here. It’s intended for those who like to read articles about A.I. and click on stories. Though perfectly fine if reading headlines is your thing.
What can you expect? You can expect a lot of links categorized for you.
💓Featured Sponsor of this Newsletter today 🚀
TED Tech: An Event for Humans Operating in the Age of Uncertainty
Join TED's first all-tech event bringing together tech, art and culture. On Sept. 18+19 in London you'll dive into the dynamic spaces shaping society with new ideas and insights to help you navigate the uncertainty around AI, Web3 and immersive tech.
💬Introduction for Today🎯
The tone of the week was OpenAI on the decline and Apple and x.AI on the hype incline. Not a very eventful week, but some important funding rounds. Also the future of media does not look good.
"The goal of xAI is to understand the true nature of the universe. We will share more information over the next couple of weeks and months." - Elon Musk
Startup Cerebras Takes on Nvidia With AI Supercomputers
"AI is good for you" by Eric Jang - A book about the last 10 years and the next 10 years of artificial intelligence. [ Buy It - Amazon link ]
📢Top News 📺
Meta - Llama 2: open source, free for research and commercial use (Official)
Microsoft - Furthering our AI ambitions – Announcing Bing Chat Enterprise and Microsoft 365 Copilot pricing (Official)
China finalizes first-of-its-kind rules governing
FTC investigates OpenAI over data leak and ChatGPT’s inaccuracy
🔗Partnership & Announcements🏭
Stability A.I. - Meet FreeWilly, Our Large And Mighty Instruction Fine-Tuned Models (Official)
Wix’s new tool can create entire websites from prompts
Announcing LangSmith, a unified platform for debugging, testing, evaluating, and monitoring your LLM applications
MosaicML launches MPT-7B-8K, a 7B-parameter open-source LLM with 8k context length
Scale AI is releasing its Scale LLM Engine to fine-tune open-source models
Snorkel AI and Together AI empower enterprises to build proprietary LLMs
🎬Videos🎥
Yuval Noah Harari: Human Nature, Intelligence, Power, and Conspiracies | Lex Fridman Podcast [link] #390
⚒️Novel Tools🔥
I omitted all minor A.I. tools in this issue. But if you like prompt-engineering basics, here's a nice beginner's guide.
👾Hacker News📰
OpenAI's Custom Instructions
🗳️ Weekly Poll 📊
Regarding ChatGPT, GPT-3.5 and GPT-4 degradation of quality (controversial).
📝 Papers 🗏
Semantic-SAM: Segment and Recognize Anything at Any Granularity
AnimateDiff: Animate Your Personalized Text-to-Image Diffusion Models without Specific Tuning
Meta-Transformer: A Unified Framework for Multimodal Learning - (Commentary)
Llama 2: Open Foundation and Fine-Tuned Chat Models (Technical Paper)
Model Card and Evaluations for Claude Models (Claude 2 Technical Paper)
Scaling Autoregressive Multi-Modal Models: Pretraining and Instruction Tuning (CM3Leon)
Challenges and Applications of Large Language Models - (Commentary)
Large language models encode clinical knowledge
FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning
How Language Model Hallucinations Can Snowball (May)
Can LLMs Generate Random Numbers? Evaluating LLM Sampling in Controlled Domains
Secrets of RLHF in Large Language Models Part I: PPO
Copy Is All You Need
📰Noted Newsletters📬
Navigating the AI Landscape: Market Trends and Opportunities For Startups
Subscribe to this Newsletter if you like A.I. Papers: Top ML Papers of the Week (LinkedIn)
🤖 Robotics 🦿
Intel Capital backs Figure’s Humanoid robot to the tune of $9 million
Insanely athletic $1,600 robot dog gets GPT-powered conversation option
🥼Healthcare🩺
ChatGPT aces medical exams, forcing a rethink on how we train tomorrow's doctors
MLCommons launches a new platform to benchmark AI medical models
AI can’t replicate this key part of practicing medicine
New AI from Harvard predicts who is most at risk of pancreatic cancer
Preparation over panic: How a Boston hospital is priming medical residents for an era of AI medicine
Meet CancerGPT: An AI That Predicts the Results of Cancer Treatment Research
🧪 Science 🔬
Meta and Azure: The AI wars might have an armistice deal sooner than expected
💰Stock Market & Investing🛒
This 'gamechanger' technology is harder than rocket science and will drive the next AI wave
Apple just added $71 billion in market value on news it's developing an 'Apple GPT' AI
Microsoft will feast on $14 billion in new revenue every year if just 10% of Office users sign on for its A.I.-powered CoPilot—Macquarie report
So, that's the end of OpenAI's ChatGPT moat? Open-Source Llama-2 might be it
Critical chip firm ASML, caught in China export restrictions, posts 38% rise in profit
🚀 Startups 🧠
Cognaize raises $18M to build a better LLM for the finance sector, one that keeps humans in the loop
Tractable snaps up $65M led by SoftBank for car and property damage appraisals using AI
Kitchen robot Botinkit raises $13M from DJI angel and others
FedML raises $11.5M to combine MLOps tools with a decentralized AI compute network
CarePredict Receives $29M in Series A-3 Funding. Combining artificial intelligence, predictive analytics, and wearable devices.
💀AGI and ASI 👽🌋
AI just wrote a bill to regulate itself
🎓 Education & Learning📚
Teaching Programming in the Age of ChatGPT
Weeknotes: Self-hosted language models with LLM plugins, a new Datasette tutorial, a dozen package releases, a dozen TILs
🎥Videos & Podcasts🎙️
🎖️National Defense & Military🕊️
Failed to upload terminators.
🔮Future of Work & Automation🦾
Will workers be paid differently in the age of AI?
AI will replace outsourced coders in India in 2 years, Stability AI CEO says
Google's Genesis project seeks to be journalists' AI-powered personal assistant
💸Venture Capital🦄
Conviction Embed - An accelerator for AI-native startups with a $150k grant and $400k+ in credits
Nvidia Accelerates AI Startup Investments, Nears Deal With Cloud Provider Lambda Labs
🐲 China A.I.🏯
China’s OpenAI challenger Zhipu AI gets Meituan funding
How Does China’s Approach To AI Regulation Differ From The US And EU?
Tech Leaders Warn the U.S. Military Is Falling Behind China on AI
🦹🏻♂️Dystopian 🗽
How AI could perpetuate racism, sexism and other biases in society
Elon Musk says he told China's leaders that if a 'digital superintelligence' is created, it could take charge of the country
US judge finds flaws in artists' lawsuit against AI companies
🏛Governance and Regulation⚖️
Biden will push for more AI authority
OpenAI loses its trust and safety leader
Exec tells first UN council meeting that big tech can’t be trusted to guarantee AI safety (Anthropic's Jack Clark)
AI Leaders Set to Accede to White House Demand for Safeguards
Moving AI governance forward (OpenAI)
AI Safety and Corporate Power – remarks given at the United Nations Security Council
OpenAI doesn't want GPT-4 to be used for facial recognition
EU data protection poses major challenges for OpenAI and ChatGPT
The $1 billion gamble to ensure AI doesn’t destroy humanity (Anthropic related)
💬Opinion Editorial🙄
How to Use AI to Do Stuff: An Opinionated Guide
I have zero respect for news and media organizations accepting money from OpenAI
🎭Entertainment & the Arts🎨
Actors vs. AI: Strike brings focus to emerging use of advanced tech
People are making money pretending to be AI on TikTok. It's both confusing and fascinating. (laughs)
📊Infographics 📏
🌏Manifestations of Utopia🏙️
ChatGPT boosts productivity and quality for educated professionals, MIT study finds
🌌Quantum Computing⚗️
Preparing For a Quantum Leap: Researchers Chart Future For Use of Quantum Computing in Particle Physics
📣Public Relations💞
Microsoft brings Bing Chat to the enterprise
Expedia adds new AI features to improve your travel planning
🙉Last Memes🎭
NYC subway using AI to track fare evasion
The end is near.
