How the workers in the 100 most common occupations in America donated.

The keelingcurve.ucsd.edu web site has graphs of CO2 concentrations over different time intervals. Here are three, all scary: ~1960 - present 1700 - present Last 800k years

The Economist just published a deeply-researched story about car bloat, and it's very, very damning: "For every life that the heaviest 1% of SUVs and trucks save, there are more than a dozen lives lost in other vehicles." Well worth your time: economist.com/interactive/un…

25 yr old pizza delivery man runs into burning house, saves 4 kids who tell him another might be in the house. He goes back in, finds the girl, jumps out a window with her, and carries her to a cop who captures the moment on his bodycam. Unbelievable!

This is so dumb

Thread of satisfying photos for perfectionists 1. Moon bridge

It was a pleasure to give a talk at this year’s Deep Learning Indaba, presenting our work on Globalizing fairness, aiming to bridge the gap in fairness research in non-western contexts. Thank you to Deep Learning Indaba organizers for the invitation and attendees for the great questions!

🪲Excited to present the publication for InkubaLM! -> 🤗 : huggingface.co/lelapa/InkubaL… -> Paper: arxiv.org/abs/2408.17024 🧵 1/n

We are collaborating with Zindi to Tuning AI at Meta LLMs for African Language Machine Translation. Data was collected partly with funds from Google.org Ghana Natural Language Processing (NLP)

Our Head of Google Research, Yossi Matias, sits down with Dr. R. Kim, Chief Medical Officer of Aravind Eye Hospital for an inspiring fireside chat! Discover how #AI is transforming vision care, helping prevent blindness, and improving lives worldwide: goo.gle/4d7BLst

We just shipped Function Calling in Google AI Studio, you can now easily test the models capability quickly without leaving the UI! 📣 🚢

New Android updates: TalkBack, our screen reader for people who are blind or have low vision, now gives detailed audio descriptions of images, powered by Gemini. Plus, Circle to Search can identify songs + earthquake alerts are expanding to 50 US states. blog.google/products/andro…

#MLSys2025 call for papers is out! The conference will be led by the general chair Matei Zaharia , PC chairs Yingyan (Celine) Lin, and Gauri Joshi. Consider submitting and bringing your latest works in AI and systems—more details at mlsys.org.

We’re delighted that Google.org is providing $10M in funding to Raspberry Pi Foundation for Experience AI, a free educational program co-created with Google DeepMind. It provides teachers with training and resources to educate young people about AI. 💡 dpmd.ai/4dPHaWb

Read about a new Heat Resilience tool that applies AI to satellite and aerial imagery, helping cities quantify how they can reduce surface temperatures with cooling interventions, such as planting trees and installing highly reflective surfaces on roofs. blog.google/outreach-initi…

Somehow seeing athletes compete in not their regular events is even more impressive (10.37s for 100m is pretty darn fast).

I still think multimodality is highly under leveraged. Flash can accept video and audio today which no other model API can do. With its context window Gemini Flash can analyze an hour of video for $0.07 🤯

Very nice site with safety data from 22M miles of rider-only driving data from Waymo. I'll draw your attention to this graph of data from Phoenix+San Francisco (where 21M of those miles happened). Injury-reported-crash rates are 3.65X lower than vs human drivers! (0.8 vs. 2.92

Waymos driving in SF are safer than humans driving in Phoenix