- Chaos Theory
- Posts
- 🥟 Chao-Down #274 Google open-sources Gemma large language models, Why AI can't replace air traffic controllers, How academic journals are fighting back against questionable AI submissions
🥟 Chao-Down #274 Google open-sources Gemma large language models, Why AI can't replace air traffic controllers, How academic journals are fighting back against questionable AI submissions
Plus, how AI can help solve the physician burnout crisis in America.
Google has finally gotten into releasing open-source large language models following contemporaries like Microsoft, Meta, and even recently Apple.
From the announcement:
Gemma is a family of lightweight, state-of-the-art open models built from the same research and technology used to create the Gemini models. Developed by Google DeepMind and other teams across Google, Gemma is inspired by Gemini, and the name reflects the Latin gemma, meaning “precious stone.” Accompanying our model weights, we’re also releasing tools to support developer innovation, foster collaboration, and guide responsible use of Gemma models.
Researchers and developers can take the two variants (a 2B model and a 7B model) and use them on their personal laptops, in their workstations, or on Google Cloud. Google is also releasing a responsible AI toolkit that can automatically filter out personal information or sensitive data from training sets.
All in all, I for one will be sure to check them out in my own projects as Gemma reports to be state of the art compared to similar models like Meta’s Llama-2.
-Alex, your resident Chaos Coordinator.
What happened in AI? 📰
Here’s What Happens When ChatGPT Writes a Scientific Article (TIME)
Why AI can’t replace air traffic controllers (CNN)
How journals are fighting back against a wave of questionable images (Nature)
America faces a shortage of primary care doctors–and they're drowning in work. Here’s how AI can solve the physician burnout crisis (Fortune)
Tinder Dating App Expands ID Checks Amid Rise in AI Scams, Dating Crimes (Bloomberg)
Why The New York Times might win its copyright lawsuit against OpenAI (Ars Technica)
Always be Learnin’ 📕 📖
Kalman Filter Explained Simply (The Kalman Filter)
A collection of learning resources for curious software engineers (Github)
You should be playing with GPTs at work (lennysnewsletter.com)
Projects to Keep an Eye On 🛠
karpathy/minbpe: Minimal, clean, code for the Byte Pair Encoding (BPE) algorithm commonly used in LLM tokenization. (Github)
google/magika: Detect file content types with deep learning (Github)
LoRA Land: Fine-Tuned Open-Source LLMs (Predibase)
The Latest in AI Research 💡
Mission Critical -- Satellite Data is a Distinct Modality in Machine Learning (arxiv.org)
Better Call GPT, Comparing Large Language Models Against Lawyers (arxiv)
AnyGPT - Unified Multimodal LLM with Discrete Sequence Modeling (junzhan2000.github.io)
The World Outside of AI 🌎
Elon Musk says Neuralink patient can control a mouse through thinking (CNBC)
NYC Sues Meta, Google, ByteDance Over Teen Mental-Health Concern (Bloomberg)
Cousins are disappearing. Is this reshaping the experience of childhood? (CBC News)
Prosthetic limb device enables users to ‘sense’ temperature difference (The Guardian)
How Sleep Affects Your Mood: The Link Between Insomnia and Mental Health (The New York Times)
Meet The Young Producers Making Beats in Ten Seconds or Less (Rolling Stone)
One Last Bite 😋
a16z’s Consumer AI Market Map (source)
