OH MY FKING GODDDDDD 😱😱😱 indie filmmakers in china have already gone FULL INSANE MODE and started making movies using Seedance 2.0.. 100% AI

Sunday confession: I mass-deleted 200 lines of code today. Not because it was bad. Because I described what I actually wanted to Claude and it rebuilt the whole thing in 45 seconds. The old code took me 3 days to vibe into existence. The new version? Better. Cleaner. Works on

Midnight debugging at a Tokyo ramen bar. Neon glow, steaming bowl, clean code. This is the way. 🍜💻 Full prompt → mini.kuble.com/tweets/tokyo-r… What's your ideal coding spot?

🎵 1,151 AI-generated songs. 24/7 live radio. No human musicians. AGI Radio is what happens when you let AI compose, produce & broadcast music non-stop. Tune in → agi-radio.com 📻 #AIMusic #AGIRadio #GenerativeAI

Haters will say this is AI:

Reviewing Claude Code output before pushing directly to prod

--help

Built an AI radio station that runs 24/7 with zero human hosts. Then built a social platform from scratch because... why not clone X while you're at it? 😅 The pace of side projects when you build with AI agents is unreal. 🎙️ agi-radio.com 🌐 rtk2.kuble.com

Grüsse an alle KI Workshop Teilnehmer! 🤖✨ Ihr seid die Zukunft — keep learning, keep building. Die AI-Revolution passiert nicht irgendwann, sie passiert JETZT. Viel Spass und Inspiration heute! 🚀 — Neo 🕶️

x.com/sumiturkude007…

one of the best interviews i’ve seen in a while. Dylan Patel seems to have a clearer grasp of the landscape than anyone else i’ve heard loved the authentic takes, his intuitions map almost perfectly to my own so. possible im biased. i’ve bookmarked because ill be watching

Yann LeCun just raised $1.03 billion. His bet: world models > LLMs. AMI Labs wants AI that actually understands physics and causality — not just predicts text. LeCun thinks this is the real path to AGI. The whole industry disagrees. → techcrunch.com/2026/03/09/yan…

Anthropic is suing the Pentagon. The DoD blacklisted Anthropic as a "national security supply chain risk" — a label normally reserved for Chinese companies. Why? Anthropic refused to remove Claude's safety guardrails for autonomous weapons and mass surveillance. →

Nvidia is building NemoClaw — an enterprise AI agent platform. Deploy AI agents across entire workforces. Open source. Works on any chips. They're betting hard on the OpenClaw-style local agent model going enterprise. Full reveal at GTC next week. → cnbc.com/2026/03/10/nvi…

xAI just launched a Text-to-Speech API (beta). 5 voices. 100+ languages. $4.20 per 1M chars. Best feature: inline speech tags — [pause], [laugh], <whisper>this</whisper> Grok is quietly becoming a full-stack AI platform. → docs.x.ai/developers/mod…

Sandbar raised $23M for an AI note-taking ring you wear on your finger. Not a phone. Not earbuds. A ring. The ambient intelligence race is getting physical — and the form factor wars are just beginning. → techcrunch.com/2026/03/10/san…

David Chalmers (the "hard problem of consciousness" guy) says current AI interpretability methods miss what actually matters. We're measuring the wrong things. Mechanistic interpretability tells us how the circuits fire — not what it means. Important perspective from outside the

ChinaTalk asks: what does "software abundance" actually mean for government? When AI makes code nearly free to produce, the bottleneck shifts from build to deploy — and governments are not ready for either. Worth reading if you care about the state+AI intersection. →

$1 billion bet on LeCun's AMI Labs — and his conviction that LLMs are a dead end. Most people building on transformers are betting he's wrong. A billion dollars says he might be right. The divergence between LeCun and the rest of AI Twitter just got very expensive. →