Interesting approach. Could not find the paper. magicinsert.github.io

Change the programs that someone else put in your mind

High time for an AI fashion show

How do you train, serve, & maintain a multi-task model with multiple clients controlling different data sources all at large scales? At 1:30pm CEST today, Smriti Pramanick, Andrew Gilchrist-Scott & Brad Stocks will be at the #ICML2024 Google booth to discuss our lessons learned.

Perplexity used to have this incredible tool to search Twitter, I absolutely loved it. Unfortunately, they had to shut it down 😭 I am BEGGING X, xAI to please add this functionality directly into Twitter. It's probably pretty straightforward to implement with Grok.

🧠 How to build LLM-powered code interpreters with Groq Inc & E2B Join us next Tuesday for an online technical session and learn how to run AI-generated code. Sign up here: hubs.la/Q02MwYGx0

✨ New features coming to NotebookLM ✨ 🗣️ Customize Audio Overviews and guide the conversation 🤝 Collaborate with teammates in the NotebookLM Business pilot program 🎧 Listen to Audio Overviews while continuing to work See what else is new ↓ goo.gle/3UcO8Na

For many years, Bill Gates used to fly Economy, not because he couldn’t afford First Class, but because he didn’t consider it good value. It’s five or six times more expensive, but only marginally better (and you all get to your destination at the same time). In the same vein, I

Art Analysis with Phi-4, powered by Microsoft’s new advanced Phi-4-multimodal-instruct model transforms how we analyze art, combining deep learning with multimodal capabilities to deliver unparalleled insights. 🎨

Doing a live stream

ChatGPT o3 vs Gemini 2.5 Pro Same prompts. Wildly different results. I’ve included videos to prove it: THREAD ↓

watching a baby grow is such a ridiculously great peak life experience

Ex-Alibaba CTO just made the boldest claim about AI & global power: “China is building the future of AI, not Silicon Valley.” He also revealed why AI by 2030 will look nothing like ChatGPT and how China’s approach is already decades ahead. Here are my top 7 takeaways: 🧵

"If you don't understand the code, your only recourse is to ask AI to fix it for you, which is like paying off credit card debt with another credit card."

Reachy mini + emotion reaction. Fun stuff by Remi Fabre!

WebNN brings hardware-accelerated AI to the browser — using GPUs/NPUs with a standardized W3C API. I wrote a deep dive on how it works, performance, and real code examples: 👉 blog.sidhartharya.com/posts/webnn-ha… #WebNN #WebML #EdgeAI #WebGPU #OnDeviceAI

You're in an ML Engineer interview at Apple. The interviewer asks: "Two models are 88% accurate. - Model A is 89% confident. - Model B is 99% confident. Which one would you pick?" You: "Any would work since both have same accuracy." Interview over. Here's what you missed: