- Chaos Theory
- Posts
- 🥟 Chao-Down #95 Pew Research finds many people haven't tried ChatGPT themselves, Lawyer apologizes for fake court citations from ChatGPT, Artist Grimes reviews AI-generated songs of Grimes
🥟 Chao-Down #95 Pew Research finds many people haven't tried ChatGPT themselves, Lawyer apologizes for fake court citations from ChatGPT, Artist Grimes reviews AI-generated songs of Grimes
Plus, the story of NVIDIA's H100 launch.
Despite the headlines from mainstream media, it seems that most people haven’t tried out ChatGPT.
Pew Research says that while a majority of Americans (58%) have heard of OpenAI’s product, few (14%) have actually tried it themselves. As imagined, those with higher household incomes and formal education are more likely to have heard and used the service.
It just goes to show that even with a super popular app like ChatGPT, companies still need to overcome the last hurdle of actual usage for widespread adoption of AI.
-Alex, your resident Chaos Coordinator.
What happened in AI? 📰
Grimes Reviews A.I. Grimes Songs (The New York Times)
How Nvidia created the chip powering the generative AI boom (Financial Times)
All the Nvidia news announced by Jensen Huang at Computex (TechCrunch)
Democratic Inputs to AI (Open AI)
Lawyer apologizes for fake court citations from ChatGPT (CNN Business)
Always be Learnin’ 📕 📖
How To Finetune GPT Like Large Language Models on a Custom Dataset (Lightning AI)
All the Hard Stuff Nobody Talks About when Building Products with LLMs (Honeycomb)
How to Build AI Products People Want (Reforge)
Projects to Keep an Eye On 🛠
wenhuchen/TheoremQA: The dataset and code for paper: TheoremQA: A Theorem-driven Question Answering dataset (Github)
anse-app/anse: Supercharged experience for ChatGPT, DALL-E and Stable Diffusion. (Github)
Voyager | An Open-Ended Embodied Agent with Large Language Models (Minedojo)
The Latest in AI Research 💡
Textually Pretrained Speech Language Models (huji.ac.il)
RWKV: Reinventing RNNs for the Transformer Era (arxiv)
Training Diffusion Models with Reinforcement Learning (arxiv)
Continually Improving Extractive QA via Human Feedback (arxiv)
The World Outside of AI 🌎
DoorDash hit with $1B lawsuit for allegedly upcharging iPhone users (NYPost)
Downtown real estate and commercial buildings are struggling. Why won’t landlords lower the rent? (Slate)
Meta, Amazon, Twitter layoffs hit teams fighting hate speech, bullying (CNBC)
Realtors leaving after signing up when home sales were hot (Axios)
Inside Google Founder Sergey Brin’s Secret Plan to Build Airships (Bloomberg)