Power On: Apple’s new Siri AI is just good enough to pull it out of its AI crisis. My hands-on impressions of the new assistant and AI features. bloomberg.com/news/newslette…

Alex Wang (Meta AI Chief) says: "The era of <open source everything> is over - and Meta is now spending hundreds of billions $ to build an AI that manages your personal life" After a year of rebuilding Meta’s AI labs, - Wang is signaling a massive shift in strategy to win the "Superintelligence" race against OpenAI and Anthropic Technological advances define the course of history... and we are ensuring the United States leads - said Wang Bookmark it & Watch it right now

Yann LeCun (LeBased) weighs in on the @AnthropicAI debacle. I have to say I agree with 100% with Yann here. "One reaps what one sows." 👏

“Luckily we are not a species that reasons only. Our ways of being will always be more than our ways of knowing.”

History repeats itself. Every century has it's great tools. Ours is digital, it's AI, and it's a tool of unprecedented capability. A digital tool for a digital society. The power to mold the world, yet again, in the hands of a few.

The obvious downstream consequence of this is that even Americans will need security clearance to access the most advanced AI. Advanced AI will be gate-kept.

WATCH: We're building the most dictator-ready tech in history. Why? Meet the frontier AI trilemma. You can: 1️⃣Operate globally 2️⃣Comply with restrictive 'safety' laws 3️⃣Serve unrestricted cognition But you can only pick 2! Given US directive suspending Fable 5 & Mythos, I think state control is coming even faster than I predicted last week.

Fable 5 / Mythos von Anthropic wurde durch die US Regierung abgeschaltet. Bis auf den Cybersecurity Benchmark ist Fable 5 ca 10% besser als Opus. Das war vermutlich das große Problem.

Does LLM really need to be a helpful assistant all the time? No. If you want to simulate people, “perfectly helpful” could be the wrong objective. Meet OdysSim, a journey toward LLMs beyond assistants, as behavioral foundation models (10B tokens of real human behavior; 23 sim benchmarks, finally in one place. new open models: outperform or on par with GPT-5.5, Gemini 3.1, or Claude Opus 4.7 in many behavior-sim dimensions). Human behavior simulation is becoming essential. Agent evaluation needs realistic users before real users show up. Medical and classroom training need realistic patients and students. Social science needs synthetic participants at scale. But real people are not ideal assistants. Real patients panic or ignore good advice. Real students misunderstand. Real customers are vague, picky, impatient, or simply leave. Human behavior is messy, diverse, and often imperfect. Frontier LLMs are getting better at math, code, and long-horizon tasks. They are NOT getting better at simulating human behavior. If anything, they drift the other way: more assistant-ish, more homogeneous, fewer of the errors and quirks real humans show. This is no accident. The whole pipeline is built for helpfulness and task success, not behavioral realism. And you can't prompt your way out of that. So we rethink the recipe from scratch and release: 🧠 The OdysSim corpus: 21.4M real human interactions (~10B tokens) from 62 sources, every conversation retrofitted with social grounding (who is talking, and why) 📏 SOUL-Index: 23 human-behavior benchmarks unified into one suite across 5 axes 🤖 OSim-8B: open weights; tops more SOUL-Index benchmarks than any frontier model, acts more like a real user than any of them on τ-bench (nearly matching real humans in the reaction dimension), and writes far more human-like text along the way.

in bringing back memes older than anthropic to make sure i’m not banned

New open-source LoRA built in-house: LTX2.3 Audio Reactive LoRA. Built for music-driven video generation: stronger beat sync, graphic texture, color separation, light pulses, particles, and shapes that move with the track. Model: huggingface.co/fal/ltx2.3-au…

We know what AI is capable of…Now the challenge is ensuring it advances safely 🫆 At World Summit AI Amsterdam, experts from research, governance, and industry will come together to explore what it takes to build AI that is resilient to misuse while continuing to drive innovation forward 🤖 This panel discussion will explore: ➡️ Strategies for mitigating AI risk at scale ➡️ Secure-by-design approaches to AI development ➡️  Governance frameworks for emerging technologies ➡️  The balance between innovation, oversight, and accountability ➡️  Building robust AI systems that society can trust Join Zico Kolter, Alice Xiang, and Helen King as they discuss the technical, ethical, and operational foundations needed to safeguard AI for society 🌍 The future of AI depends not only on what these systems can do, but on how responsibly we choose to build and govern them. Can you afford to miss the conversation?  📍 World Summit AI Amsterdam | 7–8 October 2026 🔗 Check out the first look at the agenda here (hubs.li/Q04kK-w50)

Someone cutting onions? Or apples? 🥹🍎 #timcook #apple #wwdc #wwdc26 #iphone

Me in @CityAM - sneering at Dubai as we stumble ever closer to the economic abyss is loser behaviour. Instead, Brexit Britain should: 🐾Learn why people vote with their feet & move there 🤝Do more trade deals like this 📈 Compete, don't accept defeat cityam.com/gulf-trade-deal-b…