Someone built a cool tool with Perplexity Computer to port a Spotify Playlist to Youtube Music automatically by just pasting a playlist URL. Cross service migrations are going to be seamless with tools like Computer. https://t.co/KJJqChfgw7

📢 BREAKING: FT reports that Yann LeCun’s startups AMI Labs raises $1.03 bn to build world models, at a pre-money valuation of $3.5bn. Congratulations @ylecun 🚀 The financing positions the company as a test of LeCun's belief that today's large language models fall ​short of human-level reasoning and autonomy. LeCun ​earlier said AMI aims to build systems capable of reasoning and planning in complex real-world settings. ‌ AMI Labs (Advanced Machine Intelligence Labs) aims to solve the limitations of standard language models by building world models using the Joint Embedding Predictive Architecture to observe spatial data. This visual framework helps the AI internalize how objects behave so it can safely plan complex actions. Relying exclusively on text limits AI to human linguistic output while ignoring the massive bandwidth of unspoken physical laws. Building predictive spatial architectures is the mandatory leap required to achieve reliable autonomous agents. Building predictive spatial architectures is the mandatory leap required to achieve reliable autonomous agents. This fundraising included backing from a global group of investors, including France’s Cathay Innovation, Amazon founder Jeff Bezos’s Bezos Expeditions, Singapore’s Temasek, Seoul-based SBVA and US chip giant Nvidia. The company's near-term target customers are organizations operating complex systems, including manufacturers, automakers, aerospace companies, biomedical firms ​and pharmaceutical groups. Over time, he added, the technology ​could also support consumer applications. "What consumers could be interacting ​with is ⁠a domestic robot. You need a domestic robot to have some level of common sense to really understand the physical world." LeCun said he was also talking ⁠with Meta ​about potentially deploying the technology in its ​Ray-Ban Meta smart glasses. "That's probably one of the shorter term potential applications," he said. --- ft .com/content/e5245ec3-1a58-4eff-ab58-480b6259aaf1

Self-play population-based RL from scratch in StarCraft, one of the papers I had the most fun working on! Growing-Up Together: Structured Exploration for Large Action Spaces. https://t.co/ArCh3V5Zoa

OpenClaw is the most capable open source AI agent out there but getting it running has been the blocker for many. KiloClaw removes that and you can run production-ready OpenClaw instance in under 60 seconds. No SSH, no Docker, no YAML. Key points: - 500+ AI models through Kilo Gateway. No lock-in. - Switch models without rewriting workflows - Compare performance and cost across providers - Build agents that stay reliable as the model landscape changes - One Kilo account. Your existing credits work. - Zero markup on AI tokens. Same transparent pricing. Video Source : KiloClaw Blog : Brendan O'Leary *sponsored*

Whoa! Yann LeCun's world model startup has raised $1.03 billion at $3.5B pre-money valuation https://t.co/cutaD0kEkc

New York is cooked https://t.co/ST9U8Xo6t2

Coding agents got good enough that an entire step in our product became pointless, so we killed it. We used to have builders write requirements, break them into tasks, then hand those tasks to their coding agent. The task breakdown step felt necessary because agents couldn't handle complexity on their own. That's not true anymore. Claude Code, Cursor, and other agents can now take a single requirement and figure out the implementation steps themselves. They break down work as they go, make decisions based on what the code actually looks like, and track what they're doing in real time. They're better at planning their own work than we are at planning it for them. So we removed the Tasks tab. The new flow is simpler: write a solid requirement, hit Build, pick your agent. That's it. No task lists, no manual breakdown, no micromanaging steps the agent should figure out on its own. Your job as a builder is to define what needs to exist and why it matters. The agent handles how. Requirements are the only leverage point that matters now. Tasks still exist, but they're record-keeping. The agent creates them as it works so you can see what it did and resume later if you need to. They're a log of what happened, not a plan you write upfront. This is what building with capable agents actually looks like. Less planning overhead, more time on the part that determines whether your agent builds the right thing.

NotebookLM: Do a deep research report and make a video where a consultant gives Sauron a strategy for actually winning the War of the Ring: "All you need to do is sign off to put a simple door on your volcano" The new video generation feature for NotebookLM is very impressive. https://t.co/hpMVMiiDon

https://t.co/1P2AUgUmyv

https://t.co/1P2AUgUmyv

Liftoff! https://t.co/xdAryTrayy

Let me get this straight > the silent filibuster isn't in the Constitution, it's literally just made up procedure > it allows a 40-senator minority to block one of the most important and overwhelmingly popular bills of all time > luckily, the majority could easily end the silent filibuster and do the will of 84% the people. Except… they continue to apply loser logic This isn’t business as usual. Anyone trying to support loser logic isn’t just a loser, they’re literally destroying the future of secure elections. Which is traitorous

テスラ家庭用蓄電池 Powerwall 3 の日本展開を決定。 https://t.co/uzqalN1xrF

I scraped tech company cafeteria menus. Today, Nvidia served truffle mushroom pizza while Tesla had soyrizo nachos. Choose your employer accordingly https://t.co/0quIV07Fje

Online anonymity may soon be over: LLMs perform large-scale de-anonymization on unstructured text, matching HackerNews users to LinkedIn profiles. https://t.co/JiNVJLdfMu

https://t.co/yEU8GFNRYk

geez have you seen how good Grok Imagine is getting one shotted this video with a 6 word prompt https://t.co/g48q5KX8ka

Agent Hooks in VS Code let you enforce policies, run checks, and guide Copilot at key moments during a session. Instead of repeating prompts, you can program how agents behave in your workflow. Learn how hooks work and when to use them. https://t.co/dHQ9z9pJw3 https://t.co/ycggIJAVbn

interesting chart (US only, normalized as % of software engineer) https://t.co/3ZYERsWHuq

BREAKING: Meta has denied all rumors of Wang’s demise https://t.co/ZgAOl6yDMs

BREAKING: Meta has denied all rumors of Wang’s demise https://t.co/ZgAOl6yDMs

Today, Speechify launches the ability to join your podcast. You can create a podcast from any document or prompt, in any style, and then step in. Ask questions, talk to the hosts, get answers. Your news article already talks back in Speechify. Your textbook responds to your questions. Your work content becomes a dialogue. Now too, all the podcasts you create on Speechify. In the very near future, it will seem odd to consume information without interactivity. Every single reading you do for work and for learning will be interactive. Speechify is pulling that future forward. One way to think about Speechify is an AI Agent for every professional and student that supports you across everything you need to read, consume & learn for work and for school. Launching first on our web app. Information is no longer a one-way street. Welcome to the conversation! – Written with Speechify Voice Typing

🚨 The @a16z consumer AI Top 100 is back! For the sixth time, we ranked consumer AI websites and mobile apps by usage (monthly unique visits and MAUs). This edition, we changed the rules. Here's why - and what the new list says about where consumer AI is heading 👇 https://t.co/4XSHw3bnGD

Meet MetaClaw 🦞— Just talk to your agent, it learns and evolves. 💬 Conversations become training trajectories ⚡ Models update live with hot-swapped weights 🧠 Failures generate new reusable skills 💻 No GPU cluster required Under the hood: 🔄 Online SkillRL training Training runs asynchronously while the model continues serving. 🐚 Skill evolution When the agent fails, an LLM analyzes the trajectory and generates new reusable skills. 📌 Skill injection Relevant skills are retrieved and injected into the system prompt at each step to guide behavior in real time. Built on Kimi-2.5 via Tinker cloud LoRA → fine-tuning costs about $10 and requires no GPU cluster 💻 Fully open-sourced https://t.co/RR24ZJvZ7T Built with @openclaw and @thinkymachines Kudos to the team @richardxp888, Jianwen Chen, @Xinyu2ML, @lillianwei423, @StephenQS0710, Zeyu Zheng, @cihangxie!

Announcing The Future Vision XPRIZE. A global competition with $3.5M+ in prize funding challenging creators anywhere on Earth to imagine hopeful, technology-forward futures worth building toward. Not warnings. Blueprints. Futures that inspire us to go boldly. Someone in your timeline is sitting on a vision that could change the world and doesn't know this exists yet. Share this. Be the reason they find it.

Have you thought of vibecoding with voice? Wanna win a pair of Meta Ray-Ban glasses? We’re hosting Voice HackSprint 2.0 in San Francisco: a curated one-day sprint for builders exploring what voice interfaces can become. This time we’re teaming up with @emergentlabs, @EntelligenceAI and Vallo to bring together researchers, engineers, students, and founders who are actively building with voice. The format stays simple: give great builders @smallest_AI credits + partner APIs, create the right constraints, and see what they can ship in a single focused day. Voice is quickly becoming one of the most natural interfaces for AI. Come build it! Link 👇

03.10.26 https://t.co/n9s02hoqjw

03.10.26 https://t.co/n9s02hoqjw

Hey I know you're a purely mercenary businessman only really concerned with your bottom line but I'm going to need you to go ahead and stand directly in the path of an anti-ship cruise missile and get obliterated into nothing to achieve a political goal so ambiguous it's impossible for me to even articulate what it is