Falcon 9 launches 29 @Starlink satellites from Florida https://t.co/wSNIcTO3xq

+2 more

BREAKING: "CONNECTORS" WILL COME TO GROK! 🚀 It'll let you connect services like PayPal, Stripe, GitHub, Slack, Crypto, AWS, and... to Grok! 🔥 Below is a list of them! 👇

+2 more

Move exactly 1 match to fix the equation. (There’s at least 3 solutions not including the ≠ ones.) https://t.co/swcXfOoYKk

Lots of open OCR releases on @huggingface lately, with the new dots.mocr (formerly dots.ocr 1.5) taking #2 on OlmOCRBench They also released an SVG variant, which converts structured graphics (e.g., charts, UI layouts, scientific figures, etc.) directly into SVG code https://t.co/DzHio0Fw5c

https://t.co/6WPvzyoHl1

my brain can’t process this information https://t.co/OXuvpzGmIk

Happy Birthday Giga Berlin! 4 years ago, we started production with delivering the first 30 cars to our customers! Since then, it has been a unique success story: - produced and delivered more than 700.000 Model Y 💪🏼 - created nearly 11.000 jobs 👷‍♂️👩‍🔧 So much more to come! https://t.co/4ld9pvAupR

🚨 BREAKING: Anthropic just opened a new role. they’re hiring chemical weapons experts now https://t.co/hjwkpUmEmV

An AI agent was used as a “cofounder” on LinkedIn and quickly gained traction before getting banned. It shows how easily agents can step into spaces designed for humans and perform convincingly enough to build presence and influence. But platforms are not ready to treat them as equal participants. As AI agents become more active online, the boundary between human and machine participation is starting to blur. https://t.co/iIiJRTYVuD @wired @ev_rat

AI can help us learn hard-to-teach skills, like empathy. Preregistered study of 968 people found almost no correlation between feeling empathic & communicating empathy. But a single practice session with an AI coach made people measurably better at it https://t.co/VDtdiNpw1J https://t.co/a6dxvxQDK5

Mini doc https://t.co/jbds9b7fo9

Most upvoted papers on @huggingface this week (March 16-22): - Demystifing Video Reasoning - AI Can Learn Scientific Taste - InCoder-32B: Code Foundation Model for Industrial Scenarios - SocialOmni: Benchmarking Audio-Visual Social Interactivity in Omni Models - MiroThinker-1.7 & H1: Towards Heavy-Duty Research Agents via Verification - HSImul3R: Physics-in-the-Loop Reconstruction of Simulation-Ready Human-Scene Interactions - OpenSeeker: Democratizing Frontier Search Agents by Fully Open-Sourcing Training Data - Qianfan-OCR: A Unified End-to-End Model for Document Intelligence by Baidu - EnterpriseOps-Gym: Environments and Evaluations for Stateful Agentic Planning and Tool Use in Enterprise Settings - Grounding World Simulation Models in a Real-World Metropolis

The way work is changing in AI is becoming visible even among its builders. One of @OpenAI’s cofounders says he hasn’t written code in months, instead focusing on understanding what is now possible. That shift signals that the constraint is no longer execution, but imagination. We may be entering a phase where thinking clearly matters more than writing code quickly. https://t.co/OQoTXtISDM @fortunemagazine

Welcome to the new tech flex. People with zero tech experience are adding “builder” to their LinkedIn profiles and boasting about their apps and agents. https://t.co/3Ici3llhzw https://t.co/132c5jmNqL

Governments are starting to treat AI as core infrastructure. In Scotland, a new AI agency reflects expectations that the technology could reshape public services and even add billions to the economy. The ambition is clear, but so is the challenge of managing risks, expectations and real-world impact. AI in the public sector is not just about efficiency. It is about how governments deliver, decide and build trust at scale. https://t.co/R20xaJmoBH @bbcnews @BBCPhilipSim @BBCScotlandNews

@clydeno9 @Sonos Yup here https://t.co/1YeOPqzJsI

@scottiev @NuroAi_Official Here I just did it again https://t.co/y0CYb1UxdI

@ASychov @NuroAi_Official Staring into the sun? You can’t do that as a human. https://t.co/4srF0Aea2C

China's ByteDance just dropped DeerFlow 2.0. This AI is a super agent harness with sub-agents, memory, sandboxes, IM channels, and Claude Code integration. 100% open source. Repo in reply👇 https://t.co/zApQYdsmv5

If you’re using Claude Code, bookmark this repo 👀 It upgrades Claude into a full multi-agent tech department (Architect, QA, Planner) powered by 30+ custom commands. The repo has 85K+ ⭐ on GitHub, and it’s obvious why. The author used this exact setup to build and ship a complete product in just 8 hours at an Anthropic hackathon (btw, he won it!) The real gold is in the architecture: → Memory: Saves project context so the AI picks up where you left off → Learning: Turns your coding patterns into reusable skills → Teamwork: Uses sub-agents to bypass context limits → Cost saving: Slims prompts to reduce API spend → Testing: Verifies its own code before finalizing → Speed: Runs multiple agents in parallel via git worktrees If you want to stop 'vibe coding' and start building with discipline, this is worth your time. 100% Free and open-source. Link to the repo in 🧵↓

And even did this once, which was just weird: -- I mean, we all mock the "Thinking..." Claude feedback but it's doing something. And I also constantly had "sending this message may exceed the context window" which why isn't the agent smart enough to chunk well? @OfficialLoganK https://t.co/3zOAl00frN

Our official skills repo is open source: https://t.co/X3yOgU0tRd Equip your agents with curated skills for iOS and Android development, Office file editing, and visual effects with GLSL shaders. There are more open source projects coming!

Python is for AI research. Rust is for AI production. We created a complete Rust impl of tiny TTS models (15M-80M) from @divamgupta 10 MB binary, 100ms cold start, CPU-only, edge-ready. 🐱🦀 https://t.co/7HngBVd0Jw CPU runs KittenTTS at 4x real-time, but Apple's GPU only hits 3x. @claudeai told us why: at this model size, GPU dispatch overhead exceeds the compute savings. ONNX's CPU backend with SIMD already saturates the hardware. Below ~1B params, the GPU loses. Full story 👇

If you're on mac silicon... M-series Here is what you can do (projected) remember this is training validation https://t.co/CDHuo4BjVb https://t.co/3cSfocjktA

I cracked DeepSeek's MLA attention in Rust. Same code. Three models. 15B to 1 trillion parameters. KV cache drops 57x. 128 KB per token becomes 2.25 KB. Builting the full pipeline... https://t.co/M4vG7QlNyj

This is huge. A group of 50 AI researchers (ByteDance, Alibaba, Tencent + universities) just dropped a 303 page field guide on code models + coding agents. And the takeaways are not what most people assume. Here are the highlights I’m thinking about (as someone who lives in Python + agents):

Best GitHub repos for Claude code that will 10x your next project: 1. Superpowers https://t.co/U5Y4BK9Lap 2. Awesome Claude Code https://t.co/qcgoxU3Up2 3. GSD (Get Shit Done) https://t.co/WfAhllWnTR 4. Claude Mem https://t.co/XLQpwdnIWN 5. UI UX Pro Max https://t.co/aQtGjMzKus 6. n8n-MCP https://t.co/7le1aluZXH 7. Obsidian Skills https://t.co/MUaoyUnasw 8. LightRAG https://t.co/ye8z4UqaMc 9. Everything Claude Code https://t.co/OAU9JE46Uz

+9 more

Best GitHub repos for Claude code that will 10x your next project: 1. Superpowers https://t.co/U5Y4BK9Lap 2. Awesome Claude Code https://t.co/qcgoxU3Up2 3. GSD (Get Shit Done) https://t.co/WfAhllWnTR 4. Claude Mem https://t.co/XLQpwdnIWN 5. UI UX Pro Max https://t.co/aQtGjMzKus 6. n8n-MCP https://t.co/7le1aluZXH 7. Obsidian Skills https://t.co/MUaoyUnasw 8. LightRAG https://t.co/ye8z4UqaMc 9. Everything Claude Code https://t.co/OAU9JE46Uz

I came to Singapore just to stay at this hotel!!! https://t.co/O4IUURsu7F

+2 more

AI writing and human writing are starting to feel noticeably different. People often describe AI-generated text as correct but slightly off, while human writing carries a clearer voice, tone and personal imprint. It is not just about accuracy, but about presence and individuality. The real skill may not be choosing one over the other, but knowing when clarity matters more and when a human voice makes the difference. https://t.co/9CCDgVGFQ0 @ConversationUS