Microsoft has fine-tuned a version of DeepSeek and is considering it as a lower-cost alternative to OpenAI and Anthropic models for Copilot Cowork. Huge development considering the history of the OpenAI/MS relationship.
📣 Introducing the Qwen-Robot Suite — Qwen-RobotNav, Qwen-RobotManip, Qwen-RobotWorld, three foundation models, a full stack for embodied intelligence.
🧭 Qwen-RobotNav — the gateway to mobility.
• Unifies 5 navigation tasks in one model: instruction following, point-goal, object-goal, target tracking, autonomous driving
• Controllable observation protocol
• Tool interface for agentic systems
🤖 Qwen-RobotManip — the foundation of interaction.
• Unified state-action space across heterogeneous robots
• Camera-frame delta poses for coherent cross-embodiment training
• Pretrained on a 38,100+ hour open-source corpus
🌍 Qwen-RobotWorld — infinite worlds for physical agents.
• Single world model, 20+ embodiments
• Natural-language action interface
• Predicts physically grounded futures across manipulation, driving, and navigation
Each model is independently useful, and could be composed as physical-world tools.Together, they form the low-level toolkit for general-purpose agentic systems that don't just see the world, but act in it.
📷 Blog:
https://t.co/ytLcbYET26
📖 Report:
Qwen-RobotNav: https://t.co/uPmSwDYGxg
Qwen-RobotManip: https://t.co/GeyIzJSpU8
Qwen-RobotWorld: https://t.co/SXPH1qzDFy
@teortaxesTex Exactly. The Soviets were dreamers. Their ideology to do big things and tackle civilizational challenges was truly remarkable.
Unlike Western identity-slop Marxist doomers. Dark Age priests.
Napoleon on speed:
“Ask me for anything you like, except time. The loss of time is irreparable in war. Operations only fail through delays. One must be quick, decisive, and always move faster than the enemy expects.”
6. DeepSeek has released details of how they setup their pre-training, SFT, RL pipelines. Those results are even reproduced. Why you more easily trust flying monkeys than reproducible and available results? They showed you they can train large models well.
Another important thing: Chinese models are not strong because they distill US models. Distillation of models via API is *impossible*. If somebody tells you the contrary, they don't understand machine learning:
The Kvochur Bell is one of the most Famous Aerobatic maneuvers ever performed by the sukhoi Su-27 Flanker Family.
The Kvochur Bell demonstrates the incredible low-speed controllability and high angle of attack performance of theSu-27 platform.
Try out the localmaxxing-cli
This should make submissions via agents much easier than through api directly
Just point them to the binaries
Just first run “lmx auth login” to authenticate your account
https://t.co/AgKT54vmcs