Rhett

Humanoid skills are about to make Claude Skills look like a joke Unitree just launched UNISTORE, an app store for humanoid robots. You can create, publish, buy, and deploy robot skills with one click across all their models. Just the motion library is live right now: basic moves, dances, martial arts, all available as free downloads. It's only out in China now (international coming soon). But the platform move is obvious: Unitree is building the iOS of physical robots. Now imagine what this looks like once it matures: Right now when you download a skill for Claude or Codex, it changes how your AI writes and works inside your computer. It's all just digital output, stuff that happens on a screen. But when you download a skill for a robot, it changes what a physical machine can do in your house, your warehouse, your office. One click and suddenly your robot knows how to make coffee. Another and it's running nightly inventory in a warehouse, sorting and logging without supervision. Another and it's harvesting ripe produce row by row from your garden, adjusting grip pressure for each fruit so nothing gets bruised. We're going from downloading apps for phones to downloading skills for AI to downloading physical capabilities for robots that operate in the real world. Very cool to think about. I made a visual of what this might look like in the future:

It’s happening…

You guys aren't going to believe this (I had to double check it was real). Unitree has made an actual mecha like Gundam, the GD01.

Opus 4.7 is amazing

This video is going viral on Instagram. The first fight ever between an Engine and Unitree robot at our new store space in SF.

🚨 OPEN SOURCE AI IS LITERALLY UNSTOPPABLE 🚨 The legendary founder of Redis (Antirez) just dropped ds4 - a custom native inference engine built specifically for DeepSeek v4 Flash This is earth shattering! Here is why: DeepSeek v4 Flash is a quasi-frontier model with a massive 1M context window You can now run it LOCALLY on a 128GB Mac using specialized 2-bit quantization The architecture is reimagined—he moved the KV cache from RAM directly to the SSD disk! 🤯 We already know DeepSeek v4 Flash is insanely good for agentic loops - Now you don't even need the cloud to run it Closed-source labs are burning tens of billions on massive GPU clusters while single brilliant developers are running frontier-level AI on laptops! They told us open-source would be worthless against trillion-dollar monopolies Instead, pure hacker culture + incredible open-weight models are completely rewriting the rules Open Source will ALWAYS win 💕

deepseek is raising a monster $7 billion round at $50B val making it china's largest ever AI raise but what shocks me the most is the founder, liang wenfeng: > he's personally contributing 40% of the round himself. $3 billion. > he owns 90 PERCENT of the company (unheard of at this valuation) > deepseek was founded inside his hedge fund, one of China's most successful funds. guys a fucking beast. this round is meant to achieve 2 things: 1. acquire as much compute to push out new deepseek models more often 2. turn deepseek revenue-positive by pushing new enterprise products (same tactic as OAI and anthropic) deepseek v4.1 is expected to release soon.

Neural networks might speak English, but they think in shapes. Understanding their rich *neural geometry* is key to understanding how they work – and to debugging and controlling them with precision. Starting today, we’re releasing a series of posts on this research agenda. 🧵

Are you kidding me??? It grasps multiple objects with different ways, all at once with… a single hand??? No pauses. 1x speed. GENE-26.5 is @gs_ai_’s robotics-native multimodal foundation model. It’s trained on 200,000+ hours of real human hand data (motion, force, touch) and runs on a 54-DoF bimanual system: Scaling that human data 4x lifted real-robot success rates from 16.6% to 65.6% on long-horizon dexterous tasks!! Same model weights, zero fine-tuning for this exact sequence. You know those tiny coordinated movements you do without thinking…? Robots couldn’t reliably do that before. Now they can. Today. This is the video you’ll send to my friends, outside of our bubble, when they say “robots are still just demos.” Congrats to the entire team around @zhou_xian_! Credit: Seen at Zu Wang (@zuwang95) Genesis official announcement for the full story + longer demo: genesis.ai/blog/gene-26-5…) ——— Weekly robotics and AI insights. Subscribe free: 22astronauts.com

Usage limits are up, effective today we're: 1) Doubling Claude Code's 5-hour limits for Pro, Max, Team and seat-based Enterprise plans 2) Removing peak hours limit reduction on Claude Code for Pro and Max plans 3) Substantially raising our API rate limits for Opus models

Introducing SubQ - a major breakthrough in LLM intelligence. It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA), And the first frontier model with a 12 million token context window which is: - 52x faster than FlashAttention at 1MM tokens - Less than 5% the cost of Opus Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention). Only a small fraction actually matter. @subquadratic finds and focuses only on the ones that do. That's nearly 1,000x less compute and a new way for LLMs to scale.

Balancing commercial goals and robotics research can be tricky, but with Atlas we're making it work.

This is the world changing Rosheim joint. It fundamentally changed robotic motion. You will be hearing a lot about these joint over this decade.

Opencode x Deepseek v4 pro x High 800 million tokens = $10 The Chinese can have all my data

Claude can do CAD now. Damn.

hicieron un trailer con ia de como seria la pelicula del gta san andreas, ya esta en otro nivel esto

It is really incredible what these robots can do now from omnixtreme - new @RoboPapers episode tomorrow

DeepSeek just permanently cut cached input prices by 10x across the entire API 139× cheaper than GPT-5.5 and 83× cheaper than Claude Sonnet 4.6 btw

Yeah okay, Lego bros, brodettes and brotheys are cooked with this one. GPT 2 Image can create full Lego sets! With actual Bricklink IDs so you can order the parts and build it. Whole new business opportunity here for the taking.

MICROSOFT DROPPED A 4B PARAMETER MODEL THAT TURNS ONE IMAGE INTO A 3D ASSET IN 3 SECONDS and it's open source TRELLIS.2 fully textured, physically accurate 3D models with PBR textures out of the box not a rough mesh..not a placeholder roughness, metallic, opacity the kind of detail that makes things look real under any lighting and it handles the weird stuff too..open surfaces, hollow interiors, geometry that breaks every other tool the model doesn't know the word "limitation" apparently microsoft.github.io/TRELLIS.2 demo is live on hugging face right now