Rami's Readings #82 - ✨ Llama 3.2 Released! 🚀🔥
Welcome to Rami’s Readings #82 - a weekly digest of interesting articles, papers, videos, and X threads from my various sources across the Internet. Expect a list of reads covering AI, technology, business, culture, fashion, travel, and more. Learn about what I do at ramisayar.com/about.
🤖 AI Reads
Llama 3.2: Revolutionizing Edge AI and Vision with Open, Customizable Models
Notes: Big release from Meta! The family of models now includes two vision LLM models and lightweight text-only models (1B and 3B) designed for #EdgeAI. Also, Qualcomm announced they are partnering with Meta to speed up inference on Snapdragon 8 Gen 3 and Snapdragon X Series. I have the lightweight models running on my M1. In my opinion, the text generation is on-par with the Llama 3.1 8B model. I covered the Llama 3.1 release in #73.
Allen Institute for AI Released Molmo
Notes: Hometown Seattle in the house! AI2 (Short-hand for the non-profit) released a family of open-source multimodal AI models performing on benchmarks as well as GPT-4o.
OpenAI: Fundraising, Restructuring, People
Notes: Three articles to catch up on everything to do with OpenAI this week.
Comments on bfloat16
Notes: A reminder that floating-point arithmetic is foundational for machine learning and computer science. I remember going to a Meta C++ Conference in 2012 and being blown away by the significant performance gains achieved through optimizing the initial buffer length of std::string to match English language patterns.
💼 Business Reads
Stanford Prof. Strebulaev’s Top Colleges for Unicorn Founders List
Notes: MIT ❤️
Intel Is Releasing Another Microcode Update to Protect Crashing Raptor Lake CPUs
Notes: My Intel saga continues… Another round of BIOS updates. 😮💨
Xi Has Finally Realized What’s Ailing China
Notes: This article is an opinion piece, so take it with a grain of salt. Similarly, I find therelaxation of property ownership rules in Shanghai equally interesting. More on the Hukou.
Signing off from Woods Coffee.